Collection and Analysis of Data for the Review required under Article 30(9) of Directive 2010/75/EU on Industrial Emissions (IED)

Transcription

1 Final Report Task 1 Collection and Analysis of Data for the Review required under Article 30(9) of Directive 2010/75/EU on Industrial Emissions (IED) 05 July 2013 Issue 2 Submitted to: European Commission Attention: DG ENV C.3, Industrial Emissions Team Avenue du Beaulieu, Brussels BELGIUM Submitted by: ICF International 3 rd Floor, Kean House 6 Kean Street London WC2B 4AS U.K.

2 blank page

3 Final Report Task 1 Collection and Analysis of Data for the Review required under Article 30(9) of Directive 2010/75/EU on Industrial Emissions (IED) 05 July 2013 Issue 2 Submitted to: European Commission Attention: DG ENV C.3, Industrial Emissions Team Avenue du Beaulieu, Brussels BELGIUM Submitted by: ICF International 3 rd Floor, Kean House 6 Kean Street London WC2B 4AS U.K.

4 blank page

5 Table of Contents 1. INTRODUCTION CONTEXT OBJECTIVES THIS REPORT CATEGORY (A) DIESEL ENGINES INTRODUCTION DESCRIPTION OF DIESEL ENGINES Sector Overview Industrial emissions directive Types of (liquid-fuelled) stationary engines and their cycles Fuel types Use and operation Speed and stroke of large stationary diesel engines Size / capacity Emission abatement measures and BAT EXISTING EMISSION LIMIT VALUES FOR DIESEL ENGINES International European Union Other countries DATA COLLECTION EXERCISE FOR THIS STUDY Overview Methodology OVERVIEW OF DIESEL ENGINES IN THE EU Number Capacity Geographical location Operational mode Current emissions Emission reduction potential FUTURE TRENDS REFERENCES CATEGORY (B) RECOVERY BOILERS WITHIN INSTALLATIONS FOR THE PRODUCTION OF PULP INTRODUCTION DESCRIPTION AND BACKGROUND Sector overview (summarised from JRC, 2012) Overview of the chemical pulp process Combustion plants in pulp mills Fuels used in recovery boilers Emissions to air Air emission abatement measures EXISTING LEGAL PROVISIONS Industrial Emissions Directive Member States Non-EU countries DATA COLLECTION EXERCISE FOR THIS STUDY Overview Methodology for developing an inventory of recovery boilers OVERVIEW OF RECOVERY BOILERS IN THE EU Number of recovery boilers Capacities of the recovery boilers Current emissions July 2013 i

6 3.5.4 Emission reduction potential FUTURE TRENDS Methods of pulp production Black liquor gasification Partial Borate Autocaustising REFERENCES CATEGORY (C) COMBUSTION PLANTS WITHIN REFINERIES FIRING THE DISTILLATION AND CONVERSION RESIDUES FROM THE REFINING OF CRUDE-OIL FOR OWN CONSUMPTION, ALONE OR WITH OTHER FUELS INTRODUCTION DESCRIPTION AND BACKGROUND Sector background References in the IED Defining the fuels in scope of distillation and conversion residues from the refining of crude oil Petrochemical complexes Specificity of the energy systems Size / capacity Combustion techniques METHODOLOGY FOR ASSESSMENT OF PLANTS IN THE EU Data sources Methodology OVERVIEW OF COMBUSTION PLANTS WITHIN REFINERIES FIRING THE DISTILLATION AND CONVERSION RESIDUES FROM THE REFINING OF CRUDE-OIL FOR OWN CONSUMPTION, ALONE OR WITH OTHER FUELS Number of combustion plants Capacity of combustion plants Fuel consumption Current emissions Abatement options FUTURE TRENDS REFERENCES CATEGORY (D) COMBUSTION PLANTS FIRING GASES OTHER THAN NATURAL GAS INTRODUCTION DESCRIPTION AND BACKGROUND References in the IED Refineries Steel industry Chemical industry Other sectors firing gases other than natural gas Abatement measures for combustion plants firing gases other than natural gas EXISTING LEGAL PROVISIONS Industrial Emissions Directive Member States DATA COLLECTION EXERCISE AND METHODOLOGY Overview of methodology Data sources and methodology refineries Data sources steelworks gases RESULTS Overview of all LCPs firing gases other than natural gas Overview of combustion plants in refineries firing RFG Overview of combustion plants firing steel industry gases Overview of combustion plants firing other gases in the chemical industry Other sectors with combustion plants firing other gases than natural gas REFERENCES July 2013 ii

7 6. CATEGORY (E) COMBUSTION PLANTS IN CHEMICAL INSTALLATIONS USING LIQUID PRODUCTION RESIDUES AS NON-COMMERCIAL FUEL FOR OWN CONSUMPTION INTRODUCTION DESCRIPTION AND BACKGROUND EXISTING EMISSION LIMIT VALUES European Union Other jurisdictions emission limit values DATA COLLECTION AND METHODOLOGY OVERVIEW Numbers Rated thermal input of the plants FUELS USED EMISSIONS ABATEMENT TECHNIQUES NO x Emissions FUTURE TRENDS REFERENCES July 2013 iii

8 Table of Figures FIGURE 2.1 ELECTRICAL EFFICIENCY OF COMPRESSION IGNITION ENGINES (DATA SOURCE: EIPPCB) 27 FIGURE 2.2 OPERATING HOURS BASED ON CUMULATIVE RATED THERMAL INPUT (DATA SOURCE: EIPPCB) 28 FIGURE 2.3 SOX EMISSION LEVELS OF COMPRESSION IGNITION ENGINES (MG/NM 3, 15% O 2, DRY) (DATA SOURCE: EIPPCB) 29 FIGURE 2.4 NOX EMISSION LEVELS OF COMPRESSION IGNITION ENGINES (MG/NM 3, 15% O 2, DRY) (DATA SOURCE: EIPPCB) 31 FIGURE 2.5 DUST EMISSION LEVELS OF COMPRESSION IGNITION ENGINES (MG/NM 3, 15% O 2, DRY) (DATA SOURCE: EIPPCB) 32 FIGURE 2.6 ESTIMATED NUMBERS OF DIESEL ENGINE PLANTS IN THE EU (SOURCE: THIS STUDY) 33 FIGURE 2.7 ESTIMATED NUMBERS OF DIESEL ENGINE PLANTS IN THE EU SPLIT BY AGE (SOURCE: THIS STUDY) 34 FIGURE 2.8 ESTIMATED RATED THERMAL INPUT (MW TH ) OF DIESEL ENGINE PLANTS IN THE EU (SOURCE: THIS STUDY) 34 FIGURE 2.9 PROPORTIONS OF DIESEL ENGINE PLANTS, CAPACITIES, ENERGY INPUT AND EMISSIONS ACROSS LOCATION CATEGORIES (SOURCE: THIS STUDY) 35 FIGURE 2.10 OPERATIONAL MODE OF DIESEL ENGINE PLANTS RATED THERMAL INPUT IN THE EU (SOURCE: THIS STUDY) 36 FIGURE 3.1 CHEMICAL PULP PRODUCTION IN EUROPE (DATA SOURCE: JRC, 2012) 44 FIGURE 3.2 OVERVIEW OF THE MAIN PROCESSES IN A KRAFT PULP MILL (SOURCE: JRC, 2012); RECOVERY BOILER HIGHLIGHTED 44 FIGURE 3.3 OVERVIEW OF THE KRAFT CHEMICAL RECOVERY PROCESS (SOURCE: TRAN & VAKKILAINNEN, 2008) 45 FIGURE 3.4 COMPLETE VIEW OF A BOILER AND SCHEMATIC VIEW OF THE LOWER AND MIDDLE PART OF THE FURNACE OF A RECOVERY BOILER OF A KRAFT PULP MILL (SOURCE: JRC, 2012) 46 FEEDSTOCKS IN EU RECOVERY BOILERS OTHER THAN BLACK LIQUOR AND SUPPLEMENTARY CONVENTIONAL FUELS (DATA SOURCE: EIPPCB) 47 FIGURE 3.6 NO X EMISSION CONCENTRATION AS A FUNCTION OF DRY SOLID CONTENT IN BLACK LIQUOR (DATA SOURCE: EIPPCB) 49 FIGURE 3.7 NO X EMISSION LEVELS PLOTTED AGAINST CO EMISSION LEVELS, SEPARATED BY CATEGORY OF REAL EXCESS OXYGEN CONCENTRATION (DATA SOURCE: EIPPCB) 50 FIGURE 3.8 SO 2 EMISSION CONCENTRATIONS OF RECOVERY BOILERS IN THE EU AS A FUNCTION OF DRY SOLID CONTENT OF BLACK LIQUOR (DATA SOURCE: EIPPCB) 52 FIGURE 3.9 FLUE-GAS SCRUBBER FOR RECOVERY BOILERS (SOURCE: JRC, 2012) 53 FIGURE 3.10 DUST EMISSION CONCENTRATIONS OF RECOVERY BOILERS (DATA SOURCE: EIPPCB) 57 FIGURE 3.11 NUMBER OF RECOVERY BOILERS IN EACH MEMBER STATE (SOURCE: THIS STUDY) 65 FIGURE 3.12 ESTIMATED RATED THERMAL INPUT OF RECOVERY BOILERS IN THE EU (SOURCE: THIS STUDY) 66 FIGURE 4.1 REFINERY FUEL OIL FIRING AS A FRACTION OF OVERALL REFINERY FUEL COMBUSTION (SOURCE: CONCAWE, 2010) 77 FIGURE 4.2 REFINERY LCP FUELS IN 2009 (DATA SOURCE: 2009 LCP INVENTORY) 77 FIGURE 4.3 ESTIMATED DISTRIBUTION OF ANNUAL AVERAGE REFINERY LCP SO 2 CONCENTRATIONS IN 1998, 2002 AND 2006 (SOURCE: CONCAWE, 2010) 81 FIGURE 4.4 REFINERY FUEL OIL CONSUMPTION AS A PROPORTION OF TOTAL ENERGY INPUT (DATA SOURCE: 2009 LCP INVENTORY)85 FIGURE 4.5 ESTIMATED DISTRIBUTION (ACROSS CUMULATIVE TOTAL ENERGY CONTENT) OF SO 2 EMISSION CONCENTRATION FROM REFINERY PLANTS FIRING REFINERY FUEL OIL IN 2009 (SOURCE: THIS STUDY, BASED ON 2009 LCP INVENTORY) 86 FIGURE 4.6 ESTIMATED DISTRIBUTION (ACROSS CUMULATIVE TOTAL ENERGY CONTENT) OF NO X EMISSION CONCENTRATION FROM REFINERY PLANTS FIRING REFINERY FUEL OIL IN 2009 (SOURCE: THIS STUDY, BASED ON 2009 LCP INVENTORY) 87 FIGURE 4.7 ESTIMATED DISTRIBUTION (ACROSS CUMULATIVE TOTAL ENERGY CONTENT) OF DUST EMISSION CONCENTRATION FROM REFINERY PLANTS FIRING REFINERY FUEL OIL IN 2009 (SOURCE: THIS STUDY, BASED ON 2009 LCP INVENTORY) 87 FIGURE 5.8 THE PROPORTIONS OF EU27 LCP TOTALS THAT LCPS FIRING VARYING PERCENTAGES OF OTHERS GASES MAKE UP FOR NUMBER, CAPACITY AND EMISSIONS (DATA SOURCE: 2009 LCP INVENTORY) 100 FIGURE 5.9 RFG CONSUMPTION AS A PROPORTION OF TOTAL ENERGY INPUT IN REFINERY COMBUSTION PLANTS 50MW TH OR MORE (DATA SOURCE: 2009 LCP INVENTORY) 102 FIGURE 3.5 FIGURE 5.10 ESTIMATED DISTRIBUTION (ACROSS CUMULATIVE TOTAL ENERGY CONTENT) OF SO 2 EMISSION CONCENTRATION FROM REFINERY PLANTS FIRING RFG IN 2009 (SOURCE: THIS STUDY, BASED ON 2009 LCP INVENTORY) 104 FIGURE 5.11 ESTIMATED DISTRIBUTION (ACROSS CUMULATIVE TOTAL ENERGY CONTENT) OF NO X EMISSION CONCENTRATION FROM REFINERY PLANTS FIRING RFG IN 2009 (SOURCE: THIS STUDY, BASED ON 2009 LCP INVENTORY) 104 FIGURE 5.12 ESTIMATED DISTRIBUTION (ACROSS CUMULATIVE TOTAL ENERGY CONTENT) OF DUST EMISSION CONCENTRATION FROM REFINERY PLANTS FIRING RFG IN 2009 (SOURCE: THIS STUDY, BASED ON 2009 LCP INVENTORY) 105 FIGURE 5.13 ANNUAL AVERAGE EMISSION CONCENTRATIONS OF COMBUSTION PLANTS FIRING STEEL INDUSTRY PROCESS OFF-GASES (DATA SOURCE: EIPPCB DATASET) 107 FIGURE 5.14 PROPORTION OF OTHER GASES OF TOTAL ENERGY INPUT IN CHEMICAL INDUSTRY COMBUSTION PLANTS 50MW TH OR MORE (DATA SOURCE: 2009 LCP INVENTORY) 108 July 2013 iv

9 FIGURE 6.1 ESTIMATED RATED THERMAL INPUT OF CHEMICAL INSTALLATIONS BURNING NON-COMMERCIAL LIQUID FUELS IN GERMANY (SOURCE: THIS STUDY, BASED ON DATA FROM AUTHORITIES IN GERMANY) 117 FIGURE 6.2 PROPORTIONS OF TYPES OF LIQUID PRODUCTION RESIDUES BEING FIRED AT CHEMICAL INSTALLATIONS (SOURCE: CEFIC SURVEY) 118 FIGURE 6.3 DAILY AVERAGE NO X EMISSIONS FROM A UTILITY BOILER RECORDED OVER ONE YEAR (SOURCE: CEFIC 2012) 121 Table of tables TABLE 2.1 TYPES OF STATIONARY ENGINES DISTINGUISHED IN THE GOTHENBURG PROTOCOL THAT ARE DIESEL ENGINES, AND RELEVANT CATEGORISATION UNDER IED DEFINITIONS 15 TABLE 2.2 NOX EMISSION LIMIT VALUES FOR MARINE DIESEL ENGINES IN MARPOL ANNEX VI TECHNICAL NO X CODE (N = ENGINE SPEED, RPM) 16 TABLE 2.3 EMISSION LIMIT VALUES FOR STATIONARY DIESEL ENGINE DRIVEN POWER PLANTS OF RATED THERMAL INPUT 3 TO 50 MW TH (WORLD BANK) 16 TABLE 2.4 PLANTS IN DEGRADED AND NON-DEGRADED AIRSHEDS - WORLD BANK 17 TABLE 2.5 EMISSION LIMITS (5% OXYGEN) FOR DIESEL ENGINES IN BELGIUM (FLEMISH REGION) 18 TABLE 2.6 ELVS FOR EXISTING DIESEL ENGINE INSTALLATIONS WHOSE CONSTRUCTION HAS BEEN STARTED BEFORE 17 MAY 2006 IN CZECH REPUBLIC (5% OXYGEN) 19 TABLE 2.7 ELVS FOR NEW DIESEL ENGINE INSTALLATIONS WHOSE CONSTRUCTION HAS BEEN STARTED AFTER 17 MAY 2006 IN CZECH REPUBLIC (5% OXYGEN) 19 TABLE 2.8 EMISSION LIMITS FOR DIESEL ENGINES IN FRANCE (5% OXYGEN) 20 TABLE 2.9 EMISSION LIMITS FOR DIESEL ENGINES IN GERMANY (5% OXYGEN) 20 TABLE 2.10 ELVS FOR ENGINE PLANTS 50 MWTH IN ITALY (3% OXYGEN) 21 TABLE 2.11 ELVS (3% OXYGEN) FOR DIESEL ENGINE PLANTS <50MW TH IN THE NETHERLANDS (BEMS) 21 TABLE 2.12 ELVS (3% OXYGEN) FOR ENGINE PLANTS IN THE NETHERLANDS (BEES-A) 21 TABLE 2.13 ELVS FOR DIESEL ENGINES IN INDIA (15% OXYGEN) 22 TABLE 2.14 ELVS FOR DIESEL ENGINES IN JAPAN (SOURCE: VDMA, 2011) 23 TABLE 2.15 STATUS OF CONSULTATION WITH MEMBER STATES ON DIESEL ENGINES 26 TABLE 2.16 ASSUMPTIONS FOR ANNUAL AVERAGE SO 2 EMISSION CONCENTRATIONS (15% O 2, DRY BASIS) OF DIESEL ENGINE PLANTS (DATA SOURCE: AVERAGE OF UNVALIDATED DATA FROM EIPPCB) 30 TABLE 2.17 ASSUMPTIONS FOR ANNUAL AVERAGE NO X EMISSION CONCENTRATIONS (15% O 2, DRY BASIS) OF DIESEL ENGINE PLANTS (DATA SOURCE: WEIGHTED AVERAGE OF UNVALIDATED DATA FROM EIPPCB) 31 TABLE 2.18 ASSUMPTIONS FOR ANNUAL AVERAGE DUST EMISSION CONCENTRATIONS (15% O 2, DRY BASIS) OF DIESEL ENGINE PLANTS (DATA SOURCE: WEIGHTED AVERAGE OF UNVALIDATED DATA FROM EIPPCB) 32 TABLE 2.20 ESTIMATED SO 2, NO X, DUST AND CO EMISSIONS (KT PER ANNUM) FROM DIESEL ENGINE PLANTS IN THE EU, COMPARED TO TOTAL LCP EMISSIONS AND TOTAL INDUSTRIAL COMBUSTION EMISSIONS 37 TABLE 2.21 ESTIMATED POTENTIAL NO X EMISSION REDUCTIONS WITH SCR 38 TABLE 2.22 ESTIMATED POTENTIAL DUST EMISSION REDUCTIONS USING PRIMARY MEASURES AND WITH BAG FILTERS 39 TABLE 2.23 ESTIMATED POTENTIAL SO 2 EMISSION REDUCTIONS FROM LOW SULPHUR HFO OR FGD 40 TABLE 3.1 OPERATING VARIABLES AFFECTING SULPHUR EMISSIONS FROM RECOVERY BOILERS 48 TABLE 3.2 OPERATING VARIABLES AFFECTING NO X EMISSIONS FROM RECOVERY BOILERS 50 TABLE 3.3 STATUS OF CONSULTATION WITH MEMBER STATES ON OPERATIONAL RECOVERY BOILERS 62 TABLE 3.4 ESTIMATED SO 2, NO X, DUST AND TRS EMISSION LEVELS (MG/NM 3 ) AND EMISSIONS (KT PER ANNUM) FROM RECOVERY BOILERS IN THE EU, COMPARED TO REPORTED ANNUAL TOTAL LCP EMISSIONS, NECD TOTAL AND TOTAL INDUSTRIAL COMBUSTION EMISSIONS 67 TABLE 3.5 ESTIMATED SO 2, NO X, DUST AND TRS EMISSIONS (KT PER ANNUM) FROM RECOVERY BOILERS PER MEMBER STATE (SOURCE: THIS STUDY) 67 TABLE 3.6 ESTIMATED EFFECTIVENESS AND COSTS OF SO 2 ABATEMENT SCENARIOS 68 TABLE 3.7 ESTIMATED EFFECTIVENESS AND COSTS OF NOX ABATEMENT SCENARIOS 69 July 2013 v

10 TABLE 3.8 ESTIMATED EFFECTIVENESS AND COSTS OF DUST ABATEMENT SCENARIOS 69 TABLE 4.1 DETERMINATIVE FUELS PER POLLUTANT AND PER CAPACITY CLASS 74 TABLE 4.2 SULPHUR, NITROGEN AND METAL CONTENT OF FRACTIONS SUITABLE FOR LIQUID REFINERY FUELS (SOURCE: JRC, 2012)76 TABLE 4.3 SINGLE FUELLED REFINERY LCPS: NUMBER, ENERGY INPUT AND EMISSION FACTORS (DATA SOURCE: 2009 LCP INVENTORY) 80 TABLE 4.4 FUEL CHARACTERISTICS ASSUMPTIONS MADE (SOURCE: CONCAWE, 2010) 82 TABLE 4.5 NUMBER OF REFINERY COMBUSTION PLANTS SPLIT BY CAPACITY CLASS 83 TABLE 4.6 RATED THERMAL INPUT (MW TH ) OF REFINERY COMBUSTION PLANTS SPLIT BY CAPACITY CLASS 84 TABLE 4.7 FUEL CONSUMPTION OF REFINERY COMBUSTION PLANTS 84 TABLE 4.8 ESTIMATED SO 2, NO X AND DUST EMISSION LEVELS (MG/NM 3 ) AND EMISSIONS (KT PER ANNUM) FROM THE COMBUSTION OF REFINERY FUEL OIL IN REFINERY COMBUSTION PLANTS IN THE EU27 AND CROATIA, COMPARED TO REPORTED ANNUAL TOTAL LCP EMISSIONS, NECD TOTAL AND TOTAL INDUSTRIAL COMBUSTION EMISSIONS 85 TABLE 4.9 METAL CONTENT OF RESIDUAL FUEL OIL TYPICALLY USED IN REFINERIES (SOURCE: JRC, 2012) 89 TABLE 5.1 REFERENCES IN THE IED TO GASES OTHER THAN NATURAL GAS 93 TABLE 5.2 THE NUMBER OF LCPS, THEIR RATED THERMAL INPUT AND ANNUAL SO 2, NO X AND DUST EMISSIONS FOR THE EU27 FOR WHICH OTHER GASES MAKE CONSTITUTES DIFFERENT PROPORTIONS OF TOTAL ENERGY INPUT (SOURCE: 2009 LCP INVENTORY) 100 TABLE 5.3 NUMBER OF REFINERY COMBUSTION PLANTS SPLIT BY CAPACITY CLASS 101 TABLE 5.4 RATED THERMAL INPUT (GW TH ) OF REFINERY COMBUSTION PLANTS SPLIT BY CAPACITY CLASS 101 TABLE 5.5 FUEL CONSUMPTION OF REFINERY COMBUSTION PLANTS 50MW TH OR MORE 102 TABLE 5.6 ESTIMATED SO 2, NO X AND DUST EMISSION LEVELS (MG/NM 3 ) AND EMISSIONS (KT PER ANNUM) FROM THE TABLE 5.7 TABLE 5.8 TABLE 5.9 TABLE 6.1 COMBUSTION OF RFG IN REFINERY COMBUSTION PLANTS IN THE EU27 AND CROATIA, COMPARED TO REPORTED ANNUAL TOTAL LCP EMISSIONS, NECD TOTAL AND TOTAL INDUSTRIAL COMBUSTION EMISSIONS 103 ACHIEVED EMISSION VALUES FOR GAS-FIRED BOILERS AND TURBINES WHEN USING PROCESS GASES FROM IRON AND STEEL WORKS (SOURCE: JRC, 2012) 106 ESTIMATED SO 2, NO X AND DUST EMISSIONS (KT PER ANNUM) FROM COMBUSTION PLANTS IN THE CHEMICAL INDUSTRY FIRING OTHER GASES, COMPARED TO REPORTED ANNUAL TOTAL LCP EMISSIONS, NECD TOTAL AND TOTAL INDUSTRIAL COMBUSTION EMISSIONS 109 ESTIMATED SO 2, NO X AND DUST EMISSIONS (KT PER ANNUM) FROM COMBUSTION PLANTS FIRING OTHER GASES, NOT INCLUDED IN ABOVE ANALYSES, COMPARED TO REPORTED ANNUAL TOTAL LCP EMISSIONS, NECD TOTAL AND TOTAL INDUSTRIAL COMBUSTION EMISSIONS 110 ESTIMATED SO 2, NO X AND DUST EMISSIONS FROM CHEMICAL INSTALLATIONS USING LIQUID PRODUCTION RESIDUES AS NON-COMMERCIAL FUEL FOR OWN CONSUMPTION 119 July 2013 vi

11 1. Introduction 1.1 Context Contract /ENV/2012/627812/C3 Collection and Analysis of Data for the Review required Central to the regulation of emissions from industrial installations in the EU is the Industrial Emissions Directive (Directive 2010/75/EU, IED), the successor of the IPPC Directive (2008/1/EC) and a number of related sectoral directives. The IED entered into force on 6 January 2011 and had to be transposed into national legislation by Member States by 7 January The IED replaces the IPPC Directive and the sectoral directives as of 7 January 2014, with the exception of the LCP Directive, which will be repealed with effect from 1 January The main changes compared to the individual directives it replaces include the strengthening of Best Available Techniques (BAT) implementation (stronger role of BAT Reference Documents and BAT associated emission levels (BAT AELs) in them), clearer provisions on permit reconsideration and inspections and stricter EU wide emission limit values for large combustion plants (LCPs). Minimum standard ELVs for LCPs (rated thermal input of at least 50MW) were re-examined during the detailed review of the IPPC Directive, which culminated in the IED. The review noted that in many cases the ELVs set in the LCP Directive were much higher than the range of emission levels associated with the application of BAT (BAT-AELs) as defined in the BAT Reference Document on Large Combustion Plants (LCP BREF) adopted in Under the IED the minimum standard ELVs were brought into line with BAT, mainly based on the upper end ranges of BAT-AELs defined in the LCP BREF. However for some types of combustion plants there was insufficient information on the BAT and/or on the impacts of setting or revising the ELVs at the time of negotiations on the IED. Therefore, for certain LCPs either no EU-wide minimum ELVs have been defined in Annex V or the ELVs set out in the 2001 LCP Directive were simply maintained. In these cases, the Commission was called upon to review, on the basis of the BAT, the need to establish Union-wide emission limit values and/or to amend the emission limit values set out in Annex V. The relevant provisions in the IED include: Article 30(8) Article 30(9) The emission limit values set out in Parts 1 and 2 of Annex V shall not apply to the following combustion plants: (a) diesel engines; (b) recovery boilers within installations for the production of pulp. For the following combustion plants, on the basis of the best available techniques, the Commission shall review the need to establish Union-wide emission limit values and to amend the emission limit values set out in Annex V: (a) the combustion plants referred to in paragraph 8; (b) combustion plants within refineries firing the distillation and conversion residues from the refining of crude-oil for own consumption, alone or with other fuels, taking into account the specificity of the energy systems of refineries; (c) combustion plants firing gases other than natural gas; August

12 (d) combustion plants in chemical installations using liquid production residues as non-commercial fuel for own consumption. The Commission shall, by 31 December 2013, report the results of this review to the European Parliament and to the Council accompanied, if appropriate, by a legislative proposal. 1.2 Objectives The objective set out in the service request is to provide support to the Commission on the review mentioned in Article 30(9) of the IED on the need to establish and/or amend the emission limit values set out in Annex V of the IED for certain types of combustion plants. This will require: a) Defining the different categories of combustion plants concerned in more detail b) Collecting data / information to describe the current situation of these plants; future trends; emission limit values applied to these plants outside the EU; techniques for preventing or reducing emissions taking into account BREFs and BAT conclusions c) Analysing data including the emission reduction potential from these plants Additional clarification was received from the Commission during the inception meeting on the expectations for the study. The objective is to technically describe the 5 plant categories that are the subject of the service request and to gather information on these plants in terms of their number, capacity, fuels, type of plant, emissions (levels and total), current abatement techniques, potential additional abatement techniques, location, uses etc. Such technical information needs to be sought to better understand why for these plants different ELVs would be required compared to regular LCPs, and to justify technically any such distinction. This study is driven by the review requirement in Article 30(9) of the IED, and its main aim is to complement the information being gathered through the on-going BREF revision process to better understand the relevance and emission reduction potential of these plants. Complementing the BREF revision process means that the study should align with the BAT conclusions (where already available) where possible, while making use of the information gathered in developing these conclusions. 1.3 This report This report forms the final output of Task 1 of the study Collection and analysis of data for the review required under Art. 30(9) of Directive 2010/75/EU on industrial emissions (led), specific contract No /ENV/2012/627812/C3 implementing Framework Contract No ENV.C.3/FRA/2011/0030. Task 2 final report is included in a separate document. This final report aims to set out definitional aspects related to five categories of combustion plant, and to report on the estimates developed from information gathered on these plants number, capacity, emissions and abatement techniques and emission reduction potential. The report is structured with one section for each category of combustion plant being assessed, i.e.: Chapter 2 covers category (a) Diesel engines August

13 Chapter 3 covers category (b) Recovery boilers within installations for the production of pulp Chapter 4 covers category (c) combustion plants within refineries firing the distillation and conversion residues from the refining of crude-oil for own consumption, alone or with other fuels, taking into account the specificity of the energy systems of refineries; Chapter 5 covers category (d) combustion plants firing gases other than natural gas; and Chapter 6 covers category (e) combustion plants in chemical installations using liquid production residues as non-commercial fuel for own consumption. The report has made use of the BREFs and draft BREFs that were available at the time of writing. The following versions of the relevant BREFs and related documents have been used: Large Combustion Plants (LCP): Adopted BREF July 2006 Production of Pulp, Paper and Board (PP): Draft May 2012 Refining of Mineral Oil and Gas (REF): Draft March 2012 Iron and Steel Production: Adopted BREF March 2012 and BAT Conclusions March Large Volume Organic Chemical (LVOC) Industry: Adopted BREF February 2003 It is important to note that different information and/or conclusions may come out of revised IED BREFs and their BAT conclusions (LCP, LVOC, PP, REF). August

14 August

15 2. Category (a) diesel engines 2.1 Introduction The IED does not currently include emission limit values for diesel engines. This chapter is focussed on looking at emission levels of currently operating stationary engines in the EU that equal or exceed 50MW th on a common stack basis and how these may change in the future. This chapter is set out with the following sections: Section 2.2 describes diesel engines by way of engine types, fuel types, use and operation and what abatement measures are applicable. It also notes which engines are not diesel engines. Section 2.3 notes existing legal definitions of diesel engines and how they are covered in legislation Section 2.4 describes the data collection process in this study for diesel engines; and Section 2.5 summarises estimates of diesel engines in the EU based on data collected and extrapolated as described in Section Description of diesel engines Sector Overview The adopted 2006 LCP BREF (EC, 2006) indicates that, globally, reciprocating engines are widely used, especially for continuously running power generation applications, commonly set up for use as larger base load or smaller decentralised CHP (combined heat and power) plants. However, within Europe the LCP BREF highlights that in Europe, few such plants exist in interconnected systems for power generation with liquid fuels and that applications are limited to more isolated systems (e.g. operated on islands) where no other fuel supply possibility exists. The trend in applications in the last 50 years from emergency back-up to continuous power generation is considered to be mainly due to the improvements to the efficiencies (and thus fuel consumption) of medium and low speed engines. Other cited advantages of reciprocating engines are optimum set-up for matching different load demands, short construction time, easy maintenance and robust design (EC, 2006). The LCP BREF notes that CHP plants using engines are well suited for, e.g. industry applications, local utility companies, residential and commercial buildings. Heat can be recovered as steam, hot water, hot air, etc. Possible utilisation options for the recovered heat are in district heating/cooling, desalination processes, and to preheat the air for some processes, etc Industrial emissions directive The IED includes the following definitions in Article 3: (34) gas engine means an internal combustion engine which operates according to the Otto cycle and uses spark ignition or, in case of dual fuel engines, compression ignition to burn fuel; (35) diesel engine means an internal combustion engine which operates according to the diesel cycle and uses compression ignition to burn fuel; According to the definition in the Article 3 (34) and (35), what sets out a diesel engine from a gas engine is the operational cycle in terms of it running on Diesel or Otto cycle, i.e. the July

16 cycle under which the engine is currently operating, together with the mode of ignition that the engine employs. The definitions in Articles 3(34) and 3(35) explicitly define engines as follows: Engine following the Otto cycle and using spark ignition gas engine Engine following the Otto cycle and using compression ignition gas engine Engine following the Diesel cycle and using compression ignition diesel engine. I.e. any engine operating on the diesel cycle, using compression ignition to burn fuel, will be classified as diesel engine, regardless of whether or not the fuel used by the engine is a liquid fuel (diesel, HFO, etc.). For gas engines the IED only currently includes ELVs when gas engines are firing gaseous fuels. Based on the above definition, the following engine types are not diesel engines and so are not the subject of this chapter: Any engine using spark-ignition, as this would not be classified as diesel engines under the IED. Diesel engines that have been modified to run in gas mode (Otto cycle) and are operated in gas mode, because the operation of these engines would not be classified as diesel engines under the IED. Dual fuel engines operating in gas mode because these would not be classified as diesel engines under the IED Types of (liquid-fuelled) stationary engines and their cycles The IED definition of diesel engines refers to the cycle of the engine and the mode of ignition used to initiate combustion of the fuel. The background on types of stationary engines firing liquid fuels and their cycles is set out here. There are two main types of stationary internal combustion engine used for continuous power generation application that can combust liquid fuels: Single-fuelled diesel engines; and Dual fuel engines; Gas-diesel engines have the ability to switch between high pressure gaseous fuel and 100% liquid fuel, and as such may at first be considered to be a subset of dual fuel engines. However, dual fuel engines as described below rely on low pressure gas when in gas mode which results in differences in the cycle that is used when in gas mode compared to gasdiesel engines. Gas-diesel engines whilst firing high pressure gas would be classed as a diesel engine if the engine is running according to the diesel cycle and using compression ignition (i.e. a liquid 'self-ignited' pilot fuel as the ignition source for the high pressure gaseous fuel). Gas-diesel engines as described are considered in this report as diesel engines. OTTO AND DIESEL CYCLES The Otto cycle is a spark-ignition cycle. The Otto cycle begins with the intake stroke in which the piston moves down the cylinder, causing a drop in pressure which draws in the air/fuel mixture into the cylinder. The Compression stroke follows in which the piston compresses a mixture of air and fuel (i.e. increasing the pressure with decreasing volume) which adds heat July

17 to the mixture. At the end of the compression stroke a spark ignites the mixture. The combustion forces the piston down the cylinder, driving a crank this is the power stroke increasing the volume and decreasing the pressure. In the exhaust stroke the exhaust valve is opened and, with the piston moving up the cylinder, the exhaust gases are ejected. In contrast to the Otto cycle, the diesel cycle is a compression ignition cycle. This means that no spark is utilised to ignite a mixture of air and fuel. This difference is manifest in the engine strokes in that in the diesel cycle only air is fed into the cylinder rather than a mixture of air and fuel. The fuel is instead injected after the air is compressed i.e. after the compression stroke at which point the air is hot enough to cause the fuel to combust without a spark (i.e. compression ignition ). The remaining two strokes are identical in the two cycles. The 2006 LCP BREF notes that in the diesel cycle combustion is realised partially at constant volume with an increase in the pressure, with the main combustion process occurring at constant pressure. Combustion is not continuous but occurs only during one part of the cycle. End-ofcompression pressure and temperature are important parameters to ensure good combustion. Maximum pressure must be limited to prevent damage. The engine materials can bear temperatures of about 1200 C, which allows a maximum cycle temperature of 2500 C. Thus the efficiency of this kind of engine is around %. In the Otto cycle, the air-fuel mixture does not reach a sufficiently high temperature through compression to combust without a spark ( auto-ignition ). Auto-ignition is a process to be avoided in the Otto cycle because spontaneous combustion at the incorrect point in the cycle ( knocking ) reduces the power output and efficiency. As such, avoiding auto-ignition effectively limits the compression ratio of the Otto cycle. The diesel cycle does not have this limitation such that compression ratios for the diesel cycle are higher than for the Otto cycle, with consequent higher temperatures inside diesel cycle cylinders. DUAL FUEL ENGINES In recent years, dual fuel engines have become available to the market. These engines typically have the possibility to run at full load on either liquid or (low pressure) gaseous fuel, and when running with these fuels operation is described as liquid mode or gas mode accordingly. The LCP BREF describes the dual fuel engine as follows: ( )The engine type is fuel versatile, it can be run on low pressure natural gas or liquid fuels such as diesel oil (back-up fuel, etc.), heavy oils, etc. and it can operate at full load in both fuel modes. In the gas mode, the engine is operated according to the lean-burn principle, i.e. there is about twice as much air in the cylinder compared to the minimum needed for complete combustion of the gas. This allows a controlled combustion and a high specific cylinder output without immediate risk of knocking or self-ignition. In gas engines, the compression of the air/gas mixture with the piston does not heat the gas enough to start the combustion process, therefore some additional energy needs to be added and this is carried out by injecting a small pilot fuel stream (for instance diesel oil). As a liquid fuel such as diesel oil has a lower self-ignition temperature than gas, the heat in the cylinder close to the top position is enough to ignite the liquid fuel which, in turn creates enough heat to cause the air/gas mixture to burn. The amount of pilot fuel is typically below one to two per cent of the total fuel consumption at full load. This engine works according to the diesel process when liquid fuel is used and according to the Otto-cycle when gas fuel is used. July

18 Therefore, according to this description, a dual fuel engine could be classified under the IED definitions as either a Diesel Engine or Gas Engine, as it is able to work under both Diesel and Otto cycles; and uses compression ignition. Whilst dual fuel engines are running in liquid mode they operate according to the diesel cycle (EC, 2006; CIMAC, 2005) and therefore dual fuel engines when running in 100% liquid mode are diesel engines according to the IED. in view of the definition in Article 3(35) of the Directive because when a dual fuel engine is operating in 100% liquid mode it uses the diesel cycle (in contrast to when it is operating in gas mode ) and uses compression ignition for lighting the air/fuel mixture. When dual fuel engines are running in gas mode, they operate according to the Otto cycle but without the spark ignition they use compression ignition (EC, 2006; CIMAC, 2005; Wärtsilä, undated), and hence would be classified as gas engines under Article 3(34) of the Directive. These engines use a small amount of liquid fuel as a pilot fuel stream in order to achieve ignition without a spark, but the use of the liquid fuel pilot stream does not influence the categorisation of the engine according to the IED definitions. Dual fuel engines also utilise liquid fuels during start-up and shutdown (SUSD) phases. The use of 100% liquid fuels during SUSD phases does not define the engine as a diesel engine under the IED; as set out earlier this definition is done through the use of the operational cycle and the mode of ignition. Dual fuel engines are few in number at present, but may increase in total installed capacity over time. Dual fuel engines primarily operate in gas mode for economic reasons and whenever gaseous fuel is available and in general only operate in liquid mode during non-routine or emergency conditions (e.g. when gaseous fuel is not available). Dual fuel engines are optimised for gas mode. The NO X emission levels of a dual fuel engine in liquid mode are higher than a regular diesel engine due to the lower compression ratio. EUROMOT asserts that under current conditions it is highly unlikely to be cost-efficient to run a dual fuel engine on HFO rather than natural gas such that this mode would only be instigated under very exceptional circumstances (e.g. gas supply interruption) if at all 1. Indeed, the LCP BREF indicates in its Table 6.48 that no monitoring of dual fuel engines in back-up mode (i.e. liquid firing) is expected. As such, little data are expected to be available on the use of liquid mode of dual fuel engines, and this may limit the analysis on this type of diesel engine. It should however be possible to verify which mode liquid or gas that a dual fuel engine is running in due to the different fuel supply lines. Since the classification of a dual fuel engine as a diesel engine or gas engine within the context of the IED depends on its operational mode, a permit for a dual fuel engine would need to draw on the ELVs for both liquid mode and gas mode operation Fuel types Diesel engines are fuel flexible (EC, 2006). The liquid fuel options available to diesel engines include crude oil (CIMAC, 2005; MAN, undated), heavy fuel oils, refinery vacuum residuals, light fuel oils including gas oil (diesel) and liquid biofuel. The LCP BREF also noted that natural gas can be used as a fuel in diesel engines; this is consistent with the 1 Personal Communication, EUROMOT, 4 January July

19 gas-diesel engine type listed above, with CIMAC (2005) and with the Gothenburg Protocol which lists natural gas as a fuel for diesel engines. Since it is necessary for liquid fuel to have low enough viscosity to flow along the fuel lines, heavy fuel oil must first be heated (to lower its viscosity) before it can be used in a diesel engine. The 2006 LCP BREF describes this process as follows: In heavy fuel oil operation, the fuel is first pre-cleaned and heated in the fuel treatment system before injection into the engine. The filters and separators in the fuel treatment system remove impurities and water in the fuel. The heavy fuel oil is preheated to the required viscosity, for good fuel atomisation at the nozzle. The liquid fuel pressure is boosted to about bar (dependent on engine type) to achieve a droplet distribution small enough for fast and complete combustion. When operating with light fuel oil no preheating or separation of the fuel is usually needed. The nozzle design for the fuel inlet is one of the key factors for the combustion process. It is our understanding that all liquid fuels are within the scope of this assessment, including crude oil and heavy fuel oil. As identified earlier, gas-diesel engines are also classed as diesel engines, such that (high pressure) natural gas is also within the scope of this assessment. Liquefied petroleum gas (LPG), which can be used as a fuel in engines, is not considered within the scope of this study as LPG-fired engines operate under the Otto cycle and would hence be considered as "gas engines" under the IED Use and operation The operational characteristics of (large) stationary diesel engines vary according to their uses. Some example uses are: Use in remote, off-grid locations. In these instances, the engine is likely to provide base load electricity generation and as such it will be running >4,000hrs per year. Back-up. If diesel engines are used to provide back-up (or black-start) capability then the engines are unlikely to be used for a large number of hours per year. Multi-unit plant: larger stationary engine plants may consist of several engines connected together to provide flexibility to match varying loads. At lower loads on the plant a small number of engines will be run (efficiently) at near to maximum load, whilst at higher loads, additional units will also be in operation. Article 34 of the IED concerns small isolated systems. This article provides an exemption until 2020 for such designated plants from the ELVs (or desulphurisation rates) for existing plants. In the case of remotely located diesel engines, Article 34 of the IED is not currently relevant as there are no ELVs set for diesel engines in the IED Annex V. Article 34 could only become applicable to diesel engines in the case in which such emission limits are set by competent authorities. No restriction has been made in the data gathering underpinning this report in terms of the operational characteristics of diesel engines. I.e. all engines have been included, whether they are run for 1 or 8760 hours per year. Clearly, however, the latter of these would contribute most to the total emissions burden. July

20 Large stationary diesel engines are similar in technology, size and fuels to large diesel engines propelling ocean going vessels. However, the use and operation of large stationary diesel engines are not in most cases similar to marine diesel engines Speed and stroke of large stationary diesel engines Large stationary diesel engines can be classified according to the rotation speed of the engine. Typically, three categories of engine speed are established: Slow speed engines have rotations per minute (rpm) of less than 300rpm; medium speed engines are commonly rpm and high speed engines have >1200rpm. These are the speed categories used in the Gothenburg Protocol, which covers individual engines from 1MW th in capacity, i.e. smaller than the minimum unit size referred to in the aggregation rules of the IED Chapter III (Article 29). High speed engines are typically the smallest engines and are commonly used to provide emergency back-up power generation in e.g. hospitals. The engine speed impacts on the NO X emissions from the engine because NO X formation is a function of residence time in the combustion chamber (i.e. engines with lower engine speeds tend to have higher NO X emissions than higher speed engines). There is no limitation of scope of diesel engines in terms of the speed of the engine, but as high speed engines are typically smaller, these may not be falling under the definition of a combustion plant under Chapter III of the IED (see also next section). The speed classification of the engine is related to the number of strokes of the engine. Two stroke diesel engines are typically slow speed engines. Four stroke engines are either medium or high speed engines, but taking into account high speed engines not being in scope as noted above, for the purposes of this report four stroke engines can be considered to be medium speed engines Size / capacity AGGREGATION RULES AND COMMON STACK The threshold for inclusion of engines within the scope of Chapter III of the IED is a rated thermal input, at a (common) stack level, of at least 50MW th. The rated thermal input is calculated by an aggregation rule set out in Article 29(3) that sets a minimum rated thermal input for individual combustion units (i.e. engines) of 15MW th. Whilst this does not necessarily remove engines less than 15MW th from the scope of Chapter III of the IED, for the purposes of this study engines less than 15MW th will not be considered. 2 In relation to Article 29(3) of the IED, this assumption is considered to be appropriate because whilst <15 MW th units would be part of a plant if the plant s total rated thermal input counting only >15 MW th units exceeds or equals 50 MW th, engine plants commonly consist of banks of engines of the same unit size and so it was assumed to be unlikely for there to be any engine plants of at least 50 MW th counted from >15 MW th units only AND which have additional <15 MW th engines. 2 n_finalpdf/_en_1.0_&a=d July

21 However, the data gathering for this study has been in some cases limited by a lack of clear data on correct aggregation of engine units to a 50MW th stack level. DIRECTLY ASSOCIATED ACTIVITIES Engines that would be covered by the provisions of Chapter II of the IED as directly associated activities to IED activities (Annex I) only are not included in the scope of this work, since these engines would not be required to meet the provisions of Chapter III. AVAILABILITY OF ENGINES The LCP BREF notes that individual medium speed diesel engines are available with capacities up to 50MW th or more, and medium speed dual-fuel engines are available with rated thermal inputs of up to 40MW th. Low speed engines are larger: diesel engines of 130MW th or more are available, whilst low speed dual fuel engines are available with rated thermal input up to 85MW th Emission abatement measures and BAT The pollutants of significance for diesel engines are NO X, SO 2 (depending on the fuel sulphur content) and particulates. Well operated and maintained large diesel engines have low CO and hydrocarbon (HC) emissions (the low emission levels are also facilitated by high combustion temperatures) although CO emissions can be negatively affected by NO X control measures (UNECE, 2012). BAT The LCP BREF (EC, 2006) identifies the following BAT for diesel engines: Electrical efficiencies of 40-45% Total efficiency in CHP applications (low pressure steam) of 60 to 70% Total efficiency in hot water production of 85% For reduction of dust (and heavy metal) emissions, in-engine measures in combination with the use of low ash and low sulphur fuel are BAT. BAT for dedusting exhausts from four stroke engines running on HFO is primary measures with a dust BAT-AEL of <50mg/Nm 3 (15% oxygen). 3 For reduction of SO 2 emissions, low sulphur fuel oil (or gas) is considered to be BAT (no BAT-AEL). If low sulphur fuel oil is not available then flue gas desulphurisation (FGD) is considered BAT. For reduction of NOx emissions, secondary end-of-pipe treatment selective catalytic reduction (SCR) is considered to be part of BAT in combination with primary measures (see limitations of SCR below) such as the Miller concept, delayed fuel injection, direct water injection (DWI) and humid air injection (HAM). No NOx BAT- AEL is set. PRIMARY MEASURES FOR NO X CONTROL The most important parameter governing the rate of NOx formation in a diesel engine is the combustion temperature: the higher the temperature the higher the NOx and lower the unburned emissions of carbon monoxide (CO) and hydrocarbons (HC). Furthermore, 3 A split view on the BAT-AEL was recorded in the BREF. July

22 reductions in NO X emissions can increase emissions of CO and HC which are otherwise low (UNECE, 2012). The Gothenburg Protocol sector guidance documents (UNECE, 2012) describe the primary NO X abatement measures available for the control of emissions from new diesel stationary engines. These are listed as: Optimising the engine for low-no X operation, for example using the Millar concept. 4 Diesel engines in production at the time of publication of the LCP BREF in 2006 optimised for NO X emissions (i.e. with engine modifications) should achieve <2000 mg/nm 3 when fired with HFO at 15% oxygen. Fuel injection retards achieve 10% to 20% NO X reduction, with an increase in fuel consumption of up to 3%. The addition of water, as injection into the combustion chamber, water-in-fuel emulsion or humidification of the combustion air. These wet methods are available for engines with access to suitable water suppliers. Ambient relative humidity has a significant impact on the resulting NOx emissions from a diesel engine. These methods are more commonly used on marine diesel engines. There is a fuel consumption penalty. Conversion to a low pressure gas dual fuel engine if natural gas is available (and which would then categorise the engine when running in gas mode as a gas engine). The development of diesel engines over the past 20 years has led to improvements in the NO X performance of the engines such that for a given electrical efficiency, modern engine NO X emission levels are up to 40% lower than those from the earlier 1990s by optimising for low NO X and adopting the Millar concept (UNECE, 2012). Further reductions in NO X emission levels from engines without secondary measures would be possible through improvements in the turbochargers and utilising higher pressure ratios. SECONDARY MEASURES FOR NO X CONTROL: SCR The only applicable secondary measure for NO X control of diesel engines is Selective Catalytic Reduction (SCR) (UNECE, 2012). The LCP BREF (EC, 2006) notes that from 1995 to 2005, NO X emissions from HFO fired diesel engines have been reduced considerably by primary measures on the engine, sometimes in combination with SCR. However the BREF notes that SCR has limited applicability for small diesel and two stroke engines that operate with varying loads, i.e. operated frequently on isolated systems, for a reduced number of hours only, and starting up or shutting down several times a day according to the electricity demand. SCR is therefore not recognised as being BAT for engines with frequent load variation. This is because an SCR unit would not function effectively when the operating conditions and the consequent catalyst temperature are fluctuating frequently outside the necessary effective temperature window (EC, 2006). UNECE (2012) notes additional conditions for the correct operation of SCR equipment as: Trace metals found in the fuels (such as Na, K, Ca, Mg, As, Se, P) should be minimised through fuel quality standards so as to avoid poisoning of the catalyst. 4 The Millar concept is addressing thermal NOx formation by early closing timing of the air inlet valves to suppress in-cylinder temperatures. July

23 For certain fuels (HFO or other residual fuels), a soot blowing system needs to be installed in the SCR reactor to keep the elements clean and avoid pressure drop increases. Examples of achievable NO X emission levels (at 15% oxygen, firing HFO) associated with diesel engine with SCR operation are 150 mg/nm 3 when running on diesel, and 325 mg/nm 3 when running on oil with 0.45% S content (UNECE, 2012). Costs of fitting SCR are given in UNECE (2012) and, for a NOx abatement efficiency of 80% on an HFO-fired medium speed diesel engine plant are: capital costs between 32/kWe and 42/kWe (lower costs for larger engines) and operating and maintenance costs between 5/MWh and 6.5/MWh (lower costs for use of ammonia). These costs are based on urea 40% solution costs of 200/t, urea granulate 400/t, 25% ammonia solution 225/t. MEASURES TO REDUCE DUST The LCP BREF (EC, 2006) noted that filter systems for dust control were under development for larger engines, but not (yet) established techniques, so particle filter systems did not form part of BAT. For large capacity plants consisting of a number of small capacity engines, each engine could be fitted with its own filter for particle / soot removal. Bag filters applied to diesel engines have lower dust filtration efficiencies than the dust filtration efficiencies achieved by electrostatic precipitators (ESPs) fitted to boiler based combustion plants. Due to the different temperature and oxygen content of the diesel flue-gas, the electrical properties of the diesel particulates (e.g. resistivity, etc.) are different compared to particulates from a boiler flue-gas, such that testing of ESPs is needed prior to commercial release. It is unclear at this stage in the development of the revised LCP BREF whether filter systems will form part of the revised BAT conclusion. It is noted however that five of the diesel engines/plants in the list of 28 engines/plants from the EIPPCB data set (described in Section 2.4) had bag filters for dust control fitted. The 2006 LCP BREF indicates that the dust filtration options are only applicable on new plants and not retrofittable. 2.3 Existing emission limit values for diesel engines International GOTHENBURG PROTOCOL The Gothenburg Protocol includes limit values for new stationary engines. 5 The Protocol sets limit values for various types of stationary engines disaggregated to a higher level than under the IED but does not explicitly define the terms used (definitions are to a certain extent implicit). In the recent amendment to the Protocol, limit values for NO X for new stationary engines are set separately for gas engines, dual fuel engines and diesel engines (compression ignition). Among the limit values set for diesel engines, additional distinction is made according to speed of the engine (in terms of rpm) and, for slow and medium speed engines, additional distinction is made according to the rated thermal input and fuel type of the engines. 5 New stationary source means any stationary source of which the construction or substantial modification is commenced after the expiry of one year from the date of entry into force for a Party of the Protocol. July

24 Derogations from the default emission limits, in the form of significantly higher limit values in a transitional period of 10 years, are provided for instances in which SCR cannot be applied for technical or logistical reasons of cases where the availability of sufficient amounts of high quality fuel cannot be guaranteed. These higher limit values are: 1,850 mg/m3 for dual fuel engines in liquid mode, 1,300mg/m 3 for medium/slow speed diesel engines between 5MW th and 20MW th and 1,850mg/m 3 for medium/slow diesel engines greater than 20MW th. Table 2.1 compares the categories of stationary engines for which NO X limit values are provided in the Gothenburg Protocol against whether these engines would come within the scope of diesel engines under the IED. The limit values are based on 15% reference oxygen content. The limit values do not apply to engines running less than 500 hours per year. For engines running between 500 and 1500 hours per year, engines may be exempted from compliance with the ELVs in the table in case they are applying primary NO X measures and meet the ELVs set out in the paragraph above. July

25 Table 2.1 Types of stationary engines distinguished in the Gothenburg Protocol that are diesel engines, and relevant categorisation under IED definitions Gothenburg Protocol Annex V point 9 types of stationary engines distinguished Categorisation under IED Category Capacity Fuel(s) Other NO X ELV (mg/m 3 ) for new engines Dual fuel engines 1-20 MW th all liquid fuels in liquid mode 225 Art 3(35) diesel engine Note 2 (if meets rated thermal input criteria) Dual fuel engines > 20 MW th all liquid fuels in liquid mode 225 Art 3(35) diesel engine Note 2 (if meets rated thermal input criteria) Diesel engines 5-20 MW th Heavy Fuel Oil (HFO) and biooils Compression ignition Slow (< 300 rpm)/ Medium (300-1,200 rpm)/ speed 225 Art 3(35) (if meets rated thermal input criteria) Diesel engines 5-20 MW th Light Fuel Oil and Natural Gas Compression ignition Slow (< 300 rpm)/ Medium (300-1,200 rpm)/ speed 190 Art 3(35) diesel engine Note 3 (if meets rated thermal input criteria) Diesel engines >20 MW th HFO and bio-oils Compression ignition Slow (< 300 rpm)/ Medium (300-1,200 rpm)/ speed 190 Art 3(35) diesel engine assuming Diesel cycle (if meets rated thermal input criteria) Diesel engines >20 MW th Light Fuel Oil and Natural Gas Compression ignition Slow (< 300 rpm)/ Medium (300-1,200 rpm)/ speed 190 Art 3(35) diesel engine assuming Diesel cycle (if meets rated thermal input criteria) Diesel engines >5 MW th [not specified] Compression ignition High speed (>1,200 rpm) 190 Art 3(35) diesel engine Note 3 (if meets rated thermal input criteria), although as noted in Section 2.2.7, high speed engines are likely out of scope. Note 1 Dual fuel engines in gas mode are normally using the Otto cycle and compression ignition and hence would be classified as gas engines under Article 3(34). If a dual fuel engine in gas mode used the Diesel cycle and compression ignition then it would be a diesel engine under Article 3(35). Note 2 Dual fuel engines in liquid mode are normally using the Diesel cycle and compression ignition and hence would be classified as diesel engines under IED Article 3(35). If a dual fuel engine in liquid mode used the Otto cycle and compression ignition then it would be a gas engine under IED Article 3(34). Note 3 Assuming the diesel engines are using the Diesel cycle (compression ignition is mentioned explicitly), even if firing (high pressure) natural gas, these would be classified as diesel engines under IED Article 3(35). If the engine was using the Otto cycle and compression ignition then it would be a gas engine under IED Article 3(34). July

26 MARPOL ANNEX VI (MARINE DIESEL ENGINES) The NOx Technical Code of MARPOL Annex VI 6, which was most recently updated in 2008, sets NO X control requirements on the emissions from new installed marine diesel engines over 130kW in power. Different emission limits are set depending on the age of the engines: those newly installed from 2011 ( Tier II ) have tighter limits applied compared to those newly installed from year 2000 ( Tier I ). A third Tier III applies in cases of designated Emission Control Areas (e.g. North Sea and English Channel; Baltic Sea). The emission limits are in units per power output (g/kwh) and depend on the engine speed (although different speed categories are used than in the Gothenburg Protocol for new stationary engines). The table below summarises the emission limits. Table 2.2 NOx emission limit values for marine diesel engines in MARPOL Annex VI Technical NO X code (n = engine speed, rpm) Tier Ship construction date on or after NO X emission limit (g/kwh) n< n<2000 n 2000 I 1 January n II 1 January n III 1 January 2016* n * subject to a technical review to be concluded 2013 this date could be delayed, regulation WORLD BANK The World Bank General EHS Guidelines 7 set out emission limit values for stationary engine driven power plants (gas and diesel engines) of 3 to 50 MW th for stationary engines that are financed by World Bank in countries where no national emission limit values for those installations exist or where limits are available but less strict than those stipulated by the World Bank. Emission bonuses for NOx are granted for power plants with high efficiency (this is to be defined on a project by project basis). The limit values apply to engines operating more than 500 hours per year. For NO X, different limit values are indicated depending on the bore (cylinder diameter) of the engine. Table 2.3 Bore (mm) Emission limit values for stationary diesel engine driven power plants of rated thermal input 3 to 50 MW th (World Bank) Emission limit (mg/nm 3, 15% oxygen) < 400 PM SO 2 NO X 50 or % S < 400 and high efficiency 1600 (Note 1) (Note 2) Note 1 - if justified by project specific considerations (e. g. Economic feasibility of using lower ash content fuel, or adding secondary treatment to meet 50mg/Nm 3, and available environmental capacity of the site) Note 2 - if justified by project specific considerations (e. g. Economic feasibility of using lower S content fuel, or adding secondary treatment to meet levels of using 1.5%S, and available environmental capacity of the site) International Convention for the Prevention of Pollution From Ships 7 July

27 The EHS Guidelines also specify the limit values for power plants with a thermal capacity of more than 50 MW and more than 500 operating hours per year and countries financed by World Bank with national regulations different from the limits set out by EHS Guidelines are expected to comply with whichever is more stringent. If less stringent levels than those provided in these EHS Guidelines are appropriate, a detailed justification is needed as part of the site-specific environmental assessment. Moreover, the guidelines differ in installations in Degraded Airsheds (DA) (where airshed should be considered as being degraded if national legislated standards are exceeded or in their absence) and Non-Degraded Airsheds (NDA). Table 2.4 Type of plants Plants in Degraded Airsheds Plants in Degraded and Non-Degraded Airsheds - World Bank Fuel type Operation mode B / Capacity (MW th) / Bore size (mm) PM limit (mg/m n 3 ) A Fuel sulphur content limit (%) NO x limit (mg/m n 3 ) A Natural gas CI, DF / >50MW th Liquid fuels MW th MW th 0.2 Natural gas CI, DF / >50 MW th C Plants in Non- Degraded Airsheds Liquid fuels CI / MW th / < 400mm CI / MW th / 400mm 50 < 2% or 1170 SO 2 mg/m n DF / MW th 2000 Liquid fuels 300 MW th 50 < 1% or 585 SO 2 mg/m n A Dry gas excess 15% O 2 content B CI=Compression Ignition, DF=Dual Fuel C Compression Ignition (CI) engines may require different emissions values which should be evaluated on a case-by-case basis through the environmental assessment process European Union The IED does not include any emission limit values for diesel engines. For gas engines, the IED only includes ELVs for gaseous fuel firing. The Directive on sulphur content in liquid fuels 8 directly limits the SO 2 emissions from diesel engines that are fired with fuels within the scope of this Directive. This is through limitation of the S content of HFO to 1% S and of gasoil to 0.1%. National regulations in specific EU member countries were also examined in terms of emission limit values and the results are summarised below. In many cases the Member States legislation also covered smaller diesel engines than those presented in this section (e.g. unit sizes below 3 MW th ); these ELVs have not been reproduced in this section. 8 Directive 1999/32/EC relating to a reduction in the sulphur content of certain liquid fuels, as amended by Directive 2009/30/EC July

28 BELGIUM (FLEMISH REGION) Contract /ENV/2012/627812/C3 Collection and Analysis of Data for the Review required Legislative requirements for stationary engines in Flanders are described in chapter 5.31 of VLAREM II (Order of the Flemish Government of 1 June 1995 concerning General and Sectoral provisions related to environmental safety. 9 The emission limit values depend on annual operating hours. Note: this study has not identified any diesel engine combustion plants >50MW th operating in Belgium. Table 2.5 Emission limits (5% oxygen) for diesel engines in Belgium (Flemish Region) First license for operation granted Engine capacity (MW th) Dust (mg/nm 3 ) Max sulphur content LFO (% mass) Max sulphur content HFO (% mass) NO X (mg/nm 3 ) CO (mg/nm 3 ) Organic subst. Before 1 January % 1.0% On or after 1 January 1993 and before 1 January % 1.0% On or after 1 January 2000 and before 1 January % 0.1% 1.0% 1.0% On or after 1 January 2005 and before 1 January % 0.1% 1.0% 1.0% 1000 Note First license to operate is granted on or after 1 January % 0.1% 1.0% 1.0% Note: In the case of diesel engines powered by liquid biomass products, with the exception of biomass waste, a higher NO X ELV of up to 2000 mg/nm 3 could have been set for the period up to 31 December CZECH REPUBLIC VDMA (2011) indicates that the regulation of emissions from diesel engines in the Czech Republic is defined through Order 146/2007. The new limit values came into force on 1 January The regulation sets limits for installations constructed and developed before 17 May 2006 and new installations whose development and construction started after 17 May SO 2 emission limits are limited through maximum sulphur contents in fuels laid down in separate legislation. 9 July

29 Table 2.6 ELVs for existing diesel engine installations whose construction has been started before 17 May 2006 in Czech Republic (5% oxygen) Engine type/ power category Fuel type NO x mg/m n 3 SP mg/m n 3 ΣC mg/m n 3 CO mg/m n 3 Compression ignition > 5 MW th heavy fuel oil, gas oil A 650 natural gas, degasifying C A 650 A Total concentration of all organic substances except methane with a mass flow over 3 kg/h. C With injection ignition. Table 2.7 ELVs for new diesel engine installations whose construction has been started after 17 May 2006 in Czech Republic (5% oxygen) Engine type/ power category Fuel type NO x mg/m n 3 SP mg/m n 3 ΣC mg/m n 3 CO mg/m n 3 Compression ignition > 5 MW th heavy fuel oil 600 A A 650 gas oil 500 A A 650 natural gas, degasifying C 500 A A 650 A The emission limits for NO x applies from 1 January The emission limits shall not apply to engines operated for less than 500 hours per annum. C With injection ignition. FINLAND The Finnish environmental protection agency defines limits on SO 2 -, NO x - and particle emissions of small combustion plants (i.e. with MW th <50) and was not further examined as the plants covered are out of the scope of this report. The Finnish authorities confirmed that there are no diesel engine plants >50MW th in Finland. FRANCE The French legislation Arrêté du 11/08/99 relatif à la réduction des émissions polluantes des moteurs et turbines à combustion ainsi que des chaudières utilisées en postcombustion soumis à autorisation sous la rubrique 2910 de la nomenclature des installations classées pour la protection de l'environnement 10 provides emission limit values for internal combustion engines from 20MWth upwards. It is assumed that the limits apply to installations of engines because no single engine exists with thermal capacity >100MWth. Note: this study has not identified any diesel engine combustion plants >50MW th operating in France July

30 Table 2.8 Emission limits for diesel engines in France (5% oxygen) Operating time (h/year) Capacity (MW th ) NO x (mg/nm 3 ) Liquid fuel Dual Fuel CO (mg/nm 3 ) NMHC (mg/nm 3 ) Dust (mg/nm 3 ) > > > Note 1 -The NO x limit for plants operated up to 500 hours per year is multiplied with the coefficient 2.5. Note 2 -If the plant is operated as a combined heat and power generation plant, the respective limit value in the table above can be exceeded by 30 mg/m n 3. Note 3 -The NO x limit for plants which have been licensed before 4 December 2000 and which consume liquid fuel is 1900 mg/m n 3 (independent from the power of the plant). Note 4 -The NO x limit for plants that have been licensed before 4 December 2000 and that consume natural gas can be defined by a person in charge from the responsible regulating authority up to 500 mg/m n 3 if the operator of the plant can prove by a technoeconomic analysis that it is impossible to observe the emission limit in the table above. Note 5 -The limit for VOC in the exhaust gas is 20 mg/m n 3, for plants with more than 50 MW th and a mass flow of organic compounds of more than 0.1 kg/h. Note 6 -The SO 2 limit value for the different fuel types is 300 mg/m n 3 for fuel oil and 1500mg/m n 3 for heavy fuels. GERMANY The Technische Anleitung zur Reinhaltung der Luft (TA Luft) is a common administrative regulation of the German government referring to the Bundes-Immissionsschutzgesetz (BImSchG) which includes limit values for combustion power plants using diesel engines. Note: this study has not identified any diesel engine combustion plants >50MW th operating in Germany. Table 2.9 Emission limits for diesel engines in Germany (5% oxygen) Operation MW th Dust (mg/nm 3 ) CO (mg/nm 3 ) NO x (mg/nm 3 ) Normal operation (4 stroke) 800 (2 stroke) Emergency plants, plants operated 300hr/yr to cover peak demands (e. g. power generation, gas or water supply) 3 80 (no limit) (no limit) ITALY The emission limit values applicable for stationary internal combustion engines in Italy are identified in VDMA (2011). VDMA (2011) identifies that separate limit values apply for engine plants below 50 MW th from those above 50 MW th. The limit values for those above 50 MW th are reproduced below for liquid fuels. July

31 Table 2.10 ELVs for engine plants 50 MWth in Italy (3% oxygen) Fuel type Power P (MW th) PM (mg/nm 3 ) CO (mg/nm 3 ) SO 2 (mg/nm 3 ) NO x (mg/nm 3 ) 50 P < liquid 100 P P 200 P > NETHERLANDS There is legislation governing the emissions from diesel engine combustion engines in plants less than 50MW th. This is BEMS 11 and it came into force on 1 April 2010 for new engines, replacing BEES B; existing installations have to comply 1 January Its ELVs are summarised in the table below. Table 2.11 ELVs (3% oxygen) for diesel engine plants <50MW th in the Netherlands (BEMS) Power MW th NO x mg/m n 3 SO 2 mg/m n 3 PM mg/m n 3 < In addition, BEES A, which is the Dutch implementation of the LCP Directive, includes piston engines within its scope, including gas oil fired engines. BEES A includes NO X limit values 12, which depend on the proportion of the fuel that is gaseous and the permit date. No limit values for HFO-fired engines are included. The most relevant limit values are reproduced in the table below, and are further detailed at Table 2.12 ELVs (3% oxygen) for engine plants in the Netherlands (BEES-A) Fuel Permit date Mechanical capacity NO x g/gj Less than 50% gas 29/51987 to 31/ >50kW 1200 times 1/30 engine yield Less than 50% gas 1/1/1990 and after >50kW 400 times 1/30 engine yield BEES A and BEMS have recently been replaced/integrated into the Activity Act. Note: this study has not identified any diesel engine combustion plants >50MW th operating in Netherlands For SO 2 and dust, the ELVs as for other combustion installations apply. July

32 2.3.3 Other countries Contract /ENV/2012/627812/C3 Collection and Analysis of Data for the Review required A review of regulations outside the EU was also performed to assess conditions imposed on diesel engines. This review covered Switzerland, Norway, India, Japan, Korea, Australia, Canada and the USA and identified relevant regulations only in Switzerland, India, Japan, Canada and the USA. Relevant summaries and extracts are provided in the tables below: SWITZERLAND Country Name of legislation Link to legislation Timescale of regulation Scope Switzerland Ordinance on Air Pollution Control of 16 December 1985 (Status as of 15 July 2010) Current Stationary installations Emission limit values Derogations Relevant clause 82 Stationary internal combustion engines Details 821 Reference value The emission limit values are based on an exhaust gas oxygen content of 5% (v/v). 823 Solids Particulate emissions must not exceed 50 mg/m Nitrogen oxides and carbon monoxide Emissions from stationary internal combustion engines with a rated thermal input of more than 100 kw must not exceed the following limit values: Carbon monoxide: 650 mg/m 3 Nitrogen oxides (nitrogen monoxide and nitrogen dioxide), expressed as nitrogen dioxide: o when operating on gaseous fuels as specified in Annex 5 Number 41 letters d and e, if these fuels are used for at least 80 per cent of the yearly operation time: 400 mg/m 3 o when operating on other fuels: 250 mg/m Nitrogen oxides and carbon monoxide For internal combustion engines of emergency generators which are operated for no more than 50 hours per year, the authorities shall specify preventive emission limitation requirements in accordance with Article 4; paragraph 1 and Annex 1 do not apply. INDIA In India the emission limits for diesel engines for power generation are regulated by the Central Environmental Protection Agency, which is mandated by the Ministry of Environment and Forest. The following values refer to diesel engines with more than 800kW rated power. Table 2.13 ELVs for diesel engines in India (15% oxygen) Date of order NO x (ppmv) CO mg/m n 3 NMHC mg/m n 3 PM mg/m n 3 Before 1 July Between 1 July 2003 and 1 July After 1 July July

33 JAPAN The Japanese law regulates emission value limits for diesel engine plants with fuel consumption > 50I/h. However, local limits might be lower (e.g. in Tokyo the NO x limit is 470 mg/nm 3 ). Table 2.14 ELVs for diesel engines in Japan (source: VDMA, 2011) Bore mm NO x ppm (13% O 2) NO x mg/nm 3 (5% O 2) Particulates mg/nm 3 (13 % O 2) Particulates mg/nm 3 (5 % O 2) < * * 200 *In certain regions 80 mg/nm 3 (13 % O 2). CANADA Country Name of legislation Link to legislation Timescale of regulation Applicability Canada Sulphur in Diesel Fuel Regulations (SOR/ ) Other useful Links: Environment Canada Sulphur in Diesel Fuel Regulations Summary, Current and future; some emissions limit values will come into force in 2014 The Regulations set maximum limits for sulphur in diesel fuel for small and large stationary engines (among other sources). A small stationary engine is a diesel engine with a per-cylinder displacement of less than cm 3. A large stationary engine means a diesel engine with a per-cylinder displacement of at least cm 3 Emission limit values Relevant clause Section 3 Details 3. The concentration of sulphur shall not exceed, (f) in diesel fuel produced, imported or sold for use in small stationary engines, 15 mg/kg after May 31, 2014; and (g) in diesel fuel produced, imported or sold for use in large stationary engines, 1000 mg/kg after May 31, July

34 USA Country Name of legislation Link to legislation Timescale of regulation United States New Source Performance Standards (NSPS) for Diesel Engines (Compression Ignition Engines) 40 C.F.R. Part 60, Subpart IIII Current. Most recent amendment from June 2011 amended the 2006 rule to: tighten the standards for engines of liters per cylinder to levels required by marine engines of the same sizes, and align emission standards for engines above 30 liters per cylinder with those for marine engine. Relevant clause Details Applicability and scope The NSPS standards apply to stationary compression ignition internal combustion engines. The rule also covers stationary engines that are used in emergencies, including emergency generators of electricity and water pumps for fire and flood control. The emission standards apply to new, modified, and reconstructed stationary diesel engines (i.e., existing in-use engines are not affected). The emission standards apply to engines whose construction, modification or reconstruction commenced after July 11, 2005 the date the proposed rule was published in the Federal Register. Compliance with Tier 1 standards is delayed to April 1, 2006 for non-fire pump engines and to July 1, 2006 for fire pump engines. Emission limit values to The standards apply to emissions of NOx, PM, CO, and NMHC. They are expressed in units of g/kwh and smoke standards as a percentage. The standards are not specific for stationary engines; the standards instead refer to ELVs already set out for different mobile engines, depending on the engine size and application: Engines of displacement below 10 liters per cylinder must meet Tier 1 through Tier 4 emission standards for mobile nonroad diesel engines. Engines used only for emergencies, are exempted from Tier 4. Engines of displacement above 10 liters per cylinder must meet emission standards for marine engines. Two groups of standards have been adopted: (1) for engine manufacturers, and (2) for engine owners/operators. Beginning with model year 2007, engine manufactures are required to emission certify stationary engines, and so they are responsible for compliance. During the transitional period before the model year 2007, engines can be sold that are not emission certified. In that case, the engine owner/operator is responsible for emission compliance. Displacement Power Year Emission limit D < 10 L per cylinder > 3000 hp (2.2MW) Nonroad Tier Nonroad Tier 2/4 10 D < 30 L per cylinder All Marine Cat. 2 Tier 2/3/4 (Tier 3/4 proposed) D 30 L per cylinder All Marine Cat. 3 Tier 1 equivalent to MARPOL Annex VI Tier I for NO X (see Table 2.2) Marine Cat. 3 Tier 2/3 - equivalent to MARPOL Annex VI Tier II/III for NO X (see Table 2.2). In addition to the NOx limits, HC ELV of 2.0 g/kwh and CO of 5.0 g/kwh. July

35 2.4 Data collection exercise for this study Overview The existing LCP emission inventory for 2009, which has been provided to ICF for the purposes of this study, is not expected to include diesel engines among the LCPs inventoried because the LCP Directive specifically excludes engines from its scope under Article 2(7). As such, the LCP emission inventory cannot be used as a data source of diesel engines in the EU. A survey has been undertaken by Euromot, the European Association of internal Combustion Engine Manufacturers, of its members to identify stationary diesel engine plants in the EU for which the engines have been manufactured by Euromot s members 13. This survey is considered to form the basis of an inventory of diesel engine plants in the EU. The data consist of individual installations for each EU Member State, the type of diesel engine (4 stroke diesel, 2 stroke diesel, dual fuel or gas diesel), the design fuel type (HFO, LFO, bio-oil, or natural gas), their electrical generating capacity, year of commissioning, location and for some plants additional information on what usage role the plant has. A second data source that has been utilised is of data submitted by operators to the EIPPCB ( EIPPCB dataset ) as part of the LCP BREF review process that is currently on-going. These data have been provided as anonymous data for the purposes of this study 14, and are noted as being not yet validated. Indeed, communications are on-going between the EIPPCB and operators in order to clarify or validate certain data points. The data should therefore be considered draft and only indicative. Further refinements are expected to be available upon the publication of the first draft of the LCP BREF in These data list at a unit level the abatement techniques, rated thermal input, fuel type, participation of the fuel in the overall rated thermal input (%), reference O 2 content (% dry basis), net electrical output per year (MWh e ), electrical efficiency (%), total fuel utilisation (%), total operating time under normal operating conditions (hours), equivalent full load operating factor, annual average emission concentrations of CO, NO X, SOx, and dust (mg/nm 3 ) and the year of first commissioning. Straight diesel engines are identified separately from dual fuel engines Methodology The inventory of diesel engine plants from EUROMOT has been taken as the starting point. The key Member States identified through this EUROMOT survey have been consulted to confirm the status of the diesel engine plants, in particular to confirm thermal capacities and aggregation in order to re-categorise plants according to the common stack approach where possible. In addition, the EIPPCB dataset was used to consolidate the list of diesel engine plants: although the EIPPCB data included engine plants commissioned more recently than the EUROMOT survey would suggest (one in 2010, and four in 2012), consultation with the JRC 15 concluded that these plants were not additional to the EUROMOT-identified plants. Furthermore, consultation with the Portuguese authorities confirmed that the dual fuel engine 13 Euromot s membership includes many of the world s internal combustion engine manufacturers. Industrial engine based plants based on engines manufactured by manufacturers other than those associated with Euromot will therefore have not been captured by this survey. However, it is considered that such plants will be few or even zero. 14 Personal communication with DG JRC, 12 December Personal Communication 15 th February July

36 plant that runs in liquid mode (firing HFO) on the island of Madeira does not meet the aggregation rule threshold for inclusion under Chapter III of the IED. Table 2.15 Status of consultation with Member States on diesel engines Member State Cyprus Czech Republic Finland Greece Italy Status / summary Not confirmed. Not confirmed. Confirmation received. No diesel engine power plants (including DF engines) that would fall under the scope of Chapter III of the IED, also taking into account the aggregation rules of article 29 of the IED, exist in Finland. Confirmation received of all the diesel engine combustion plants with thermal input >50MWth, including installations with common stacks and separate ones (based on the aggregation rules of Article 29.1 of the 2010/75/EC Directive). Authorities also indicated where units with thermal input <15MWth are included because the common stack exceeds 50MW th without these units being calculated. Partial confirmation received. Confirmed three plant closures, one revision of capacity and two confirmation of correct details. No details on remaining plants. Malta No response, but details confirmed via plant permit. 16 Portugal Spain UK Confirmation received. Confirmed that Vitoria power plant, on Madeira:, all diesel engines have their own stack, and all have less than 50 MW th Caniçal Power plant, also in Madeira: there are two combustion plants (with diesel engines) with more the 50 MWth. All other plants with diesel engines, located in the Azores and on Madeira island have less than 50 MWth, so are excluded from IED, chapter III. Confirmation received. Indicated that the installation at Ibiza, which includes 7 diesel engines of total rated thermal input 367MW th, has separate stacks for each diesel engine which implies that none of the diesel engines count as combustion plants >50MW th. However it is unclear if the four emission points of the four more recent engines could be considered as a common stack 17. Contact has promised a response but not received. NB: Member States not listed in this table are estimated to not have any diesel engine plants >50MW th. CAPACITY The EUROMOT data source is understood to be on a basis of plant (stack) level, i.e. based on minimum rated thermal input of 50 MW. In addition, for purposes of this analysis, the survey has been conducted assuming a minimum unit (engine) capacity of 15 MW th. Three plants were included by EUROMOT but which were not expected to meet the 50MW th threshold; these plants have been excluded from the analysis. The rated thermal input of the plants has been estimated from the EUROMOT data on electrical capacity by assuming that the electrical efficiency of the plants is related to the age of the plant. A linear relationship between year of commissioning and electrical efficiency was derived from the EIPPCB data for this purpose, as shown below in Figure 2.1. The relationship suggests an increase in efficiency over time July

37 Figure 2.1 Electrical efficiency of compression ignition engines (data source: EIPPCB) Note: data not validated OPERATING HOURS The EIPPCB data included the number of operating hours of the plants for which data were submitted. These show that approximately one third of total capacity are standby plants with annual operating hours less than 1500/yr, and a further third have high annual utilisation of over 6,000 hours per year. The dataset from EUROMOT indicates for some plants their usage profile, whether as standby, baseload or industrial. The annual operating hours for standby plants have been assumed to be the average of those marked as standby plants in Figure 2.2 (approximately 500 hours per year), and the average of the remaining plants (approximately 5,000 hours per year) has been assumed for all other plants as it has not been possible to establish different usage patterns for industrial from baseload generation. July

38 Figure 2.2 Operating hours based on cumulative rated thermal input (data source: EIPPCB) Approx. one third of capacity: High number of operating hours per year (max = 8760) Annual operating hours Approx. one third of capacity: Varying number of operating hours per year Standby plants Average: 527hrs Combined average for upper two categories: 5002hrs Cumulative rated thermal input (MW) Note: data not validated FUEL USE The majority of the fuel reportedly in use by diesel engines is heavy fuel oil (HFO). EUROMOT report Italian engine plants as using liquid biofuel ( bio oil ) and one plant in the Czech Republic as using light fuel oil (LFO). The EIPPCB dataset also lists the majority of plants as running on HFO, and one plant running on LFO. One diesel engine reported by EUROMOT is running on high pressure natural gas. The EIPPCB data included a factor for each plant that represented the equivalent full load operating factor, calculated from fuel energy input during the reference year divided by the total operating time under normal operating conditions, and then further divided by the total rated thermal input of the whole combustion plant. The weighted average of the plants equivalent full load operating factors is 71%, as derived from the EIPPCB dataset; this factor has been assumed for all the plants. This allows the fuel energy input per plant to be estimated using the operating hours assumed above by multiplying the equivalent full load operating factor by annual operating hours and by rated thermal input. The fuel is assumed to be 100% liquid fuel for those engines marked as not being dual fuel or gas diesel engines. The flue gas volumes associated with the combustion of the fuels are needed to estimate the total emissions from the plants when only data on emission concentrations is available (as is the case for the EIPPCD dataset). The flue gas volumes have been estimated from the total fuel energy estimates using fuel specific flue gas volumes. The specific volume for combustion of HFO in engines (15% oxygen, dry) has been estimated from the figure quoted in AMEC (2012) for liquid fuel fired boilers with adjustment made for the % oxygen content, resulting in a figure of 837 Nm 3 /GJ. The specific volume for combustion of natural gas in engines is assumed to be 760 Nm 3 /GJ (taken from AMEC, 2012). July

39 SO 2 EMISSIONS The EIPPCB data includes SO 2 emission levels for every plant (except the plant firing LFO) and information regarding abatement equipment at the plants. The EIPPCB dataset of reference plants does not include any diesel engines situated in Italy. 18 The data on emission levels are reproduced in Figure 2.3 below. SO 2 emissions are a function of the sulphur content in the fuel and whether end-of-pipe clean-up is utilised. Figure 2.3 shows wide variation in SO 2 emissions among plants without end-of-pipe flue gas desulphurisation (FGD), ranging from the same level as those plants with FGD to emission levels over five times higher. This suggests that the data reflect a wide variation in the sulphur content of the HFO. Given that 1% S content of HFO corresponds with around 570 mg SO 2 /Nm 3 at 15% oxygen content 19, the implied sulphur contents of the plants without FGD appears to range between around 0.2% and 1%. This variation is considered to be realistic. Figure 2.3 SOx emission levels of compression ignition engines (mg/nm 3, 15% O 2, dry) (data source: EIPPCB) Note: data not validated The plants noted in EIPPCB dataset as having FGD installed are those five plants that were commissioned more recently and which were added to the EUROMOT dataset. It is therefore possible to assume that the plants in the EUROMOT dataset do not have end of pipe clean-up for SO 2 emissions. These plants are assumed to have the average SO 2 emission level of 376 mg/nm 3 (weighted by capacity) shown in Figure 2.3, which corresponds with an average S content of 0.7%, whilst the added five plants are known to have FGD installed and so are assumed to have an SO 2 emission level of 90 mg/nm 3. These assumptions rely on the data from EIPPCB as being representative of diesel engine emission levels across the EU. The difference in the two assumed emission levels is equivalent to a 75% reduction in emission concentration. Whilst this is less than the emission reduction capability of FGD (normally considered to be 90 to 95% for wet FGD), the 75% may either reflect the use of dry sorbent injection (which is an FGD technique with lower 18 Personal communication with DG JRC 15 th February At 1% S, 1 tonne HFO contains 10kg S, which upon full oxidation to SO 2 with no release as H 2S, is equivalent to an unabated emission of 20kg SO 2. At 15% oxygen, the specific flue gas volume assumed for liquid fuels combusted in diesel engines is 837Nm 3 /GJ (see above). The flue gas volume from combusting 1 tonne HFO (assuming a net calorific value of 42GJ/tonne) is therefore 837x42=35154 Nm 3. The emission concentration from combusting 1 tonne of HFO that has 1%S is therefore 20,000,000 mg / Nm 3 = 569 mg/nm 3. July

40 removal efficiencies) or may reflect the case that those plants without FGD may to some extent already be using lower sulphur HFO. No information is available as to the types of FGD that have been applied. The diesel engine plants in Italy are noted by EUROMOT as firing crude palm oil or a derived biodiesel. The sulphur content of this fuel is assumed to be zero, thus it is assumed these plants do not emit SO 2. (No Italian diesel engine plants are in the EIPPCB dataset to corroborate this assumption). On the assumption that the LFO (gas oil) fired plant will have a fuel with maximum sulphur content of 0.1%, this corresponds with an SO 2 emission concentration of 57 mg/nm 3 (on the basis of 1% S 570 mg/nm 3 ). (No data on SO 2 emission concentrations were provided in the EIPPCB dataset for the LFO fired plant.) Table 2.16 Assumptions for annual average SO 2 emission concentrations (15% O 2, dry basis) of diesel engine plants (data source: average of unvalidated data from EIPPCB) SO 2 emission levels of HFO plants without FGD 376 mg/nm 3 SO 2 emission levels of HFO plants with FGD 94 mg/nm 3 SO 2 emission levels of LFO plants 57 mg/nm 3 SO 2 emission levels of bio oil plants 0 mg/nm 3 SO 2 emission levels of natural gas fired gas-diesel plants 0 mg/nm 3 SO 2 emissions in tonnes per annum are estimated by multiplying the annual flue gas volume with the annual average emission concentrations. NO X EMISSIONS The EIPPCB data includes NO X emission levels per plant and information regarding abatement equipment at the plants, specifically noting the following: Five plants using SCR plants commissioned in 2010 and Four plants that have SCR installed but which have not used this since There is no indication as to why the SCR has not been used since These plants are lean burn and use in-cylinder NOx reduction techniques All remaining plants do not have end of pipe NOx reduction installed. The data on the NO X emission levels of each of these categories of plant are reproduced in Figure 2.4 below. July

41 Figure 2.4 NOx emission levels of compression ignition engines (mg/nm 3, 15% O 2, dry) (data source: EIPPCB) Note: data not validated Since no information is available in the EUROMOT dataset to distinguish between plants employing primary NO X reduction techniques (in cylinder, lean burn) from those that do not, a capacity weighted average NO X emission level from all the plants that do not operate with functioning SCR has been applied (2,048 mg/nm 3 ). The difference between the assumed emission levels for plants without SCR to those with SCR (i.e. implied abatement efficiency for SCR) suggests that the plants with SCR may well be demonstrating additional NO X control from primary measures. Table 2.17 Assumptions for annual average NO X emission concentrations (15% O 2, dry basis) of diesel engine plants (data source: weighted average of unvalidated data from EIPPCB) NO x emission levels of HFO/bio oil plants without SCR 2,048 mg/nm 3 NO x emission levels of HFO plants with SCR 125 mg/nm 3 NO x emission levels of LFO plants 1,143 mg/nm 3 NO x emission levels of natural gas fired gas-diesel plants 138 mg/nm 3 Note: the assumption for LFO plant is based on data for one plant. NOx emissions in tonnes per annum are estimated by multiplying the annual flue gas volume with the annual average emission concentrations. DUST EMISSIONS The EIPPCB dataset includes dust emission levels per plant and information regarding abatement equipment at the plants, specifically noting the following: Five plants using bag filters plants commissioned in 2010 and All remaining plants do not have end of pipe dust reduction installed. The data on the dust emission levels of each of these categories of plant are reproduced in Figure 2.5 below. July

42 Figure 2.5 Dust emission levels of compression ignition engines (mg/nm 3, 15% O 2, dry) (data source: EIPPCB) Note: data not validated Those plants noted in EIPPCB dataset as having bag filters installed are those five plants that were commissioned more recently and which were added to the EUROMOT dataset. It is therefore assumed that the plants in the EUROMOT dataset do not have end of pipe clean-up for dust emissions. These plants are assumed to have the capacity-weighted average dust emission level of 59 mg/nm 3 shown in Figure 2.5, whilst the added five plants are known to have bag filters installed and so are assumed to have a dust emission level of 3 mg/nm 3. Table 2.18 Assumptions for annual average dust emission concentrations (15% O 2, dry basis) of diesel engine plants (data source: weighted average of unvalidated data from EIPPCB) Dust emission levels of HFO/bio oil plants without bag filters 59 mg/nm 3 Dust emission levels of HFO plants with bag filters 3 mg/nm 3 Dust emission levels of LFO plants 0 mg/nm 3 Dust emission levels of natural gas fired gas-diesel plants 0 mg/nm 3 Dust emissions in tonnes per annum are estimated by multiplying the annual flue gas volume with the annual average emission concentrations. CO EMISSIONS The EIPPCB data set includes data on the CO emission levels per plant. No information on abatement techniques used to reduce CO emissions is noted, although it is recognised that oxidation catalysts are probably widely used but not noted specifically. CO emission levels can be related to NO X abatement if excessively lean conditions are sought for NOx control as this can lead to incomplete and unstable combustion and high CO levels. There does not appear to be a correlation between fuel type and CO emissions in the EIPPCB data, so a single assumption of CO emission levels is assumed for all diesel engine plants of 108 mg/nm 3. The range of annual average CO emission levels was from 6 to 199 July

43 mg/nm 3. The average of all the EIPPCB plant CO emission levels is assumed for all the plants in the EUROMOT dataset. 2.5 Overview of diesel engines in the EU The combined dataset from EUROMOT and EIPPCB as described in section 2.4 has been used to develop estimates for the prevalence of diesel engine plants in the EU27 in terms of their number, total capacity, location, operational mode, current emissions and emission reduction potential Number The combined dataset indicates that there are 40 diesel engine plants in the EU that would meet the 50 MW th threshold for inclusion in the IED. Geographically, the majority of these diesel engine plants are situated in southern EU Member States around the Mediterranean: Spain, Italy, Portugal, Greece, Cyprus and Malta together make up 90% of the total number of plants. The 40 plants are split into 4 stroke (and medium speed) diesel units (60%), two stroke (and slow speed) diesels (38%), and one gas-diesel engine. Nearly three quarters of plants are reportedly firing HFO, with most of the remainder firing bio-oil. The plants firing bio oil are reported as being all the diesel engine plants in Italy. The bio-oil is imported renewable crude palm oil it is understood that this is the typical bulk fuel used as legally the plants may not be permitted to combust fossil fuels, and that subsidies are available for electricity produced from plants fired with this fuel. Figure 2.6 Estimated numbers of diesel engine plants in the EU (source: this study) split by country split by engine type split by fuel type Split by capacity class The LCP inventory lists 3,283 LCPs across the EU-27 in Assuming that the existing LCP inventory excludes diesel engine plants, the estimated number of diesel engine plants represents 1.2% of the number of LCPs in the EU. The split of plants by their age is shown below in Figure 2.7. The plot shows that the distribution of diesel plant ages is bimodal with significant numbers of plants less than 10 years as well as greater than 20 years old. July

44 Figure 2.7 Estimated numbers of diesel engine plants in the EU split by age (source: this study) Capacity The combined dataset estimates that the total rated thermal input of diesel engine plants in the EU is 3.6 GW. The Figure below sets out how this thermal input is split by capacity class and by engine type. A little more than half of the capacity is in the capacity class of MW th and the remainder in the MW th category. The split of capacity by engine type closely matches the split of number of plants across engine type i.e. there does not appear to be a distinct difference in the average plant size of 2-stroke diesel plants versus 4 stroke diesel plants (note that this may not mean that the unit engine sizes are necessarily the same). Figure 2.8 Estimated rated thermal input (MW th ) of diesel engine plants in the EU (source: this study) split by capacity class split by engine type The LCP inventory includes a total rated thermal input of all 3,283 LCPs across the EU-27 in 2009 as being 1,354 GW th, so the estimated capacity of diesel engine plants represents 0.3% of the total rated thermal input of all LCPs in the EU Geographical location A large portion of the plants reported by EUROMOT are situated on islands as shown in Figure 2.9. The figure also shows the split of capacity, fuel input and emissions estimates across location categories. Between 50% and 60% of the diesel engine plants in terms of numbers, rated thermal input, energy input and emissions are on islands (including Malta, Cyprus, and Sardinia). July

45 Figure 2.9 Proportions of diesel engine plants, capacities, energy input and emissions across location categories (source: this study) The Gothenburg Protocol includes a term remote islands in its footnote to the table Limit values for NOx emissions released from new stationary engines in reference to applicability of SCR. However, the Protocol does not define this term remote. More remote locations may rely on the use of diesel engines for power supply due to diesel engines characteristics of: No need for pipeline gas supply No need for large quantities of water Black start capability (ability to begin operation without an external electricity supply) Rapid start-up times Ability to match high variations in load efficiently (through use of banks of engines) Eurelectric (2011) note a number of aspects related to operation of diesel engine plants in remote locations, including: Island diesel engine plants may employ a larger number of smaller capacity engines rather than a smaller number of large engines, due to need to have higher levels of redundancy. This may affect aggregation rules, and may lead to lower efficiencies of plants. The use of end of pipe abatement may have complications, e.g.: The variable load profile and frequent start-stop cycles can lead to insufficient exhaust gas temperatures for effective operation of SCR. Geographically remote locations incur high costs for the supply of reagents and transport of spent catalysts for SCR. Similarly for the use of end of pipe measures that produce by-products, these have higher transit costs too. Within this study, no criteria have been applied to distinguish plants island locations into remote or not remote. In addition to the term remote islands, Article 34 of the IED includes a provision for to allow combustion plants in small isolated systems to be exempt until 2020 from compliance with the ELVs referred to in Article 30(2) and the rates of desulphurisation referred to in Article 31. However, there are currently no ELVs for diesel engines, which renders the Article 34 provision currently not applicable to diesel engine plants. July

46 2.5.4 Operational mode Contract /ENV/2012/627812/C3 Collection and Analysis of Data for the Review required Some partial information has been captured on the use of the diesel engine plants, whether the plants operational mode is baseload electricity generation (either as single cycle, or as a combined cycle with the exhaust heat of the engine flue gases recovered in exhaust gas boilers to produce steam feeding a steam turbine to produce additional electricity), in industrial situations or as standby. This information is summarised in Figure 2.10 below. The majority are used in baseload electricity generation, with a significant proportion being set up with combined cycles to improve total fuel efficiency. Only a small proportion of the total rated thermal input is used in industrial or standby situations (although with resolution of the unknown modes these proportions could change). The operational mode is important as it affects the potential abatement and its cost-effectiveness. The application of abatement measures to standby plants will be least cost effective, but considering they are identified as a small source, the abatement scenarios in Section do not take account of standby plants. Figure 2.10 Operational mode of diesel engine plants rated thermal input in the EU (source: this study) Current emissions The current annual emissions for diesel engine plants have been estimated according to the methodology noted in Section It is important to reiterate here that some of the assumptions that were necessary to make have significant uncertainties, e.g. the annual operational load of the plants, and the emission concentrations of the plants. Consequently, analysis of total emission load is best assessed at EU level rather than subdivided further. The annual emission concentration data are based on the reported data provided by EIPPCB; since these data are of unknown reporting year but likely to be a recent year these annual emissions are labelled as being current. The table below presents the estimated SO 2, NO X, dust and CO emissions for diesel engine plants > 50MWth in the EU for the present day, and compares the estimates to total emission estimates reported for LCPs from the 2009 LCP inventory, and against total industrial combustion emissions 20 from inventories reported for 2009 under NECD. 20 Total industrial combustion is taken as being the sum total of the following NFR codes (per AMEC, 2012): 1A1a (Public Electricity and Heat Production); 1A1b (Petroleum refining); 1A1c (Manufacture of solid fuels and other energy industries); 1A2a, 1A2b, 1A2c, 1A2d, 1A2e, 1A2fi (Stationary combustion in manufacturing industries and construction: Iron and Steel, Non-ferrous metals,: Chemicals, Pulp, Paper and Print, Food processing, beverages and tobacco, and Other). July

47 Table 2.19 Estimated SO 2, NO X, dust and CO emissions (kt per annum) from diesel engine plants in the EU, compared to total LCP emissions and total industrial combustion emissions Annual pollutant emission (kt) Diesel engine plants (estimated in this study) Comparison with 2009 LCP inventory emissions Comparison with NECD emissions Total Industrial combustion Bio HFO LFO Natural gas Total 2009 LCP % of LCP 2009 NECD % of NECD 2009 NECD % of NECD SO , % 4, % 3, % NO X , % 9, % 2, % Dust % N/A N/A N/A N/A CO N/A N/A N/A N/A N/A N/A Diesel engine plants are estimated to have NO X emissions that represent about 5% of the total NO X emissions from LCPs reported for year 2009, around 2.5% of total NO X emissions reported under NECD from industrial combustion activities and less than 1% of total NECD NO X emissions in For dust, the estimates in the table suggest that diesel engine plants emissions are less than 2% of the total dust emissions from LCPs reported for year The estimates presented in the table above suggest that diesel engine plants contribute less than 1% to total SO 2 emissions of LCPs. The estimate for CO emissions from diesel engine plants in the EU represents around 0.02% of total EU CO emissions Emission reduction potential As explained in Section 2.4, based on data from the EIPPCB, a certain proportion of the plants have been assumed to have end of pipe abatement equipment for SO 2, NO X and dust control fitted already. These are new plants and so the choice of NO X abatement may have been influenced by the Gothenburg Protocol which sets emission limits for NO X from new engines. The EIPPCB has indicated that, for four of the five plants with end of pipe SO 2, NOx and dust abatement fitted, the driving force for implementing the abatement equipment was not of local AQ origin, but was EU and national emissions legislation; modular and flexible plant with a high conversion efficiency. 22 These new plants make up 421 MW th which represents 12% of the estimated total installed capacity of diesel engine plants. For the remainder, a further 17% are assumed to have in-cylinder NOx reduction, and operate in lean burn mode. This 17% are also noted by EIPPCB data as having SCR already fitted but which is not used (but can operate at 90% NOx reduction). NOX The BAT conclusion for NO X emissions from diesel engines as set out in the LCP BREF suggests that SCR is a technique that has the potential to be applied to all engines (new and existing) subject to applicability limitations. It is supplementary to the use of primary 21 Sum of EU27 LRTAP reported National totals for the entire territory for year 2009 is kt CO Personal communication with DG JRC 15 th February July

48 measures. The NO X limits in the Gothenburg Protocol (for new diesel engines) are based on the use of SCR and application of the technique is considered necessary to meet them. Primary measures are expected to be sufficient to meet the limits set for exemption cases in the Gothenburg protocol (for engines >20 MW th 1850 mg/nm 3 ; for 5-20 MW th engines, 1300 mg/nm 3 ). 23 Due to the lack of information on the existing uptake of primary measures for NO X control, it is not possible to estimate the potential NO X emission reductions from solely applying primary measures to remaining plants. 24 What is more certain in the underlying dataset is the uptake of SCR (set out on page 30). The potential annual emission reductions in tonnes have been estimated by assuming that all (or some) of the diesel engine plants operate at NO X emission levels the same as those with SCR (as described in Section 2.4.2) scenario 2. The potential limitations of applicability for remote islands are noted in section 2.5.3; however as no indication is available on which island plants are remote, a separate estimate is produced in which SCR is not applied to those diesel engine plants located on islands, i.e. assuming that all plants on islands are isolated systems (Scenario 1). This assumption is likely to overestimate the classification of plants as remote (and underestimate the potential SCR uptake), as it may not be the case that all island plants are remote. Indeed, the most recent plants on Malta (i.e., an island) have SCR fitted for example. This scenario should therefore be interpreted as a lower limit of potential SCR uptake. In all scenarios the plants firing natural gas (one plant) and LFO (one plant) retain their baseline estimated NO X emissions. The table below presents the estimated total EU emission reductions of these scenarios. The costs, which are based on those presented in Section 2.2.8, take into account the identified four engines that already have SCR installed but not operating (i.e. zero capital cost, but operational costs). Table 2.20 Estimated potential NO X emission reductions with SCR Scenario Estimated annual NO X emissions (kt) NOx emission reduction (kt) NOx emission reduction (%) Total costs ( m) Baseline estimate Abatement scenario 1: All HFO plants not known to be on islands with SCR Abatement scenario 2: All HFO plants with SCR % % The total emission reductions shown appear to be likely to be a small overestimate as the implied percentage NO X reduction is 94% whereas the NOx removal efficiency of SCR is normally quoted as being up to around 90%. This was discussed in Section as being likely representing a combination of primary measures and SCR. 23 Personal communication with EUROMOT 22 nd February 2013 suggests that 1300mg NO X /Nm 3 is not yet attained by engines presently on the market. This claim has not been verified. 24 However, the EIPPCB questionnaire does allow operators to indicate which primary measures are implemented for the abatement of NO X emissions out of: In-cylinder NOx reduction, lean burn concept, exhaust gas recirculation (EGR), emulsified fuel, water injection, steam injection, charge air humidification. This information (if reported) was not provided to ICF. July

49 Fuel switching to gas is not considered as an abatement scenario due principally to reasons for operators for choosing diesel engines as the generation type and restrictions that may exist for operators from using gas. DUST As recognised in Section 2.2.8, the BAT-AEL for four stroke engines is 50 mg dust /Nm 3. In section it was set out that the assumed dust emission level of HFO-fired plants that did not operate end of pipe dust control was 64mg/Nm 3. Accordingly, based on this assumption (noted as being itself based on not yet validated data), emission levels of 50mg/Nm 3 dust ought be possible from the existing engines through in-engine measures in combination with the use of low ash and low sulphur fuel. The potential emission reductions from meeting this BAT-AEL are estimated in scenario 3 and shown in the table below. This is considered applicable to existing plants. No cost data are available for this scenario. The use of end of pipe dust control measures (e.g. bag filters) would lead to even greater dust emission reductions. The potential emission reductions from use of bag filters is estimated in scenario 4 by assuming all plants operate at the dust emission level assumed for bag filter equipped plants (see Section 2.4.2). This assumes that technical developments in bag filters and diesel engines will allow them to be retrofitted to all diesel engine plants. This may not be possible and so this scenario is likely to overestimate the potential emission reductions. The table below presents the estimated total EU emission reductions of these scenarios. Table 2.21 Estimated potential dust emission reductions using primary measures and with bag filters Scenario Estimated annual dust emissions (kt) Emission reduction (kt) Emission reduction (%) Baseline estimate Abatement scenario 3: (At least) meeting BAT-AEL of 50 mg/nm % Abatement scenario 4: All use of bag filters % SO 2 The BAT conclusion for mitigating SO 2 emissions from diesel engines as set out in the LCP BREF is to use low sulphur fuel oil, or in its absence, use end of pipe solutions, i.e. FGD. As identified above, the data suggests that the sulphur content of the HFO used in the diesel engines ranges from 0.2% to 1%, with a weighted average of 0.7%. A broad assumption that fuel oil with half the existing sulphur content could be sourced for the diesel engines, i.e. corresponding with an average sulphur content of 0.35%. A simple estimate is made of the potential SO 2 emission reductions from this approach (scenario 5). Regarding the use of FGD, one estimate is given in the table below for the application of FGD to all plants (scenario 7). A second scenario (scenario 6) to give an indication of limiting uptake of FGD to those plants that are not isolated systems is made by only applying FGD to plants that are not known to be located on islands. This is not a perfect criterion since, as identified earlier, some plants on islands may not be remote; however it provides a first order indication of impacts.. The estimate is based on the emission level identified in Section for plants with FGD, without any assumptions about low sulphur HFO in combination with FGD. No assumptions are applied to the natural gas fired plant or LFO fired plant. The table below presents the estimated total EU emission reductions of these scenarios. July

50 Table 2.22 Estimated potential SO 2 emission reductions from low sulphur HFO or FGD Scenario Estimated annual SO 2 emissions (kt) SO 2 emission reduction (t) SO 2 emission reduction (%) Baseline estimate Abatement scenario 5: Lower sulphur HFO Abatement scenario 6: All HFO plants not known to be on islands with FGD Abatement scenario 7: All HFO plants with FGD % % % 2.6 Future trends The use of diesel engines in remote locations is expected to continue over the next 10 to 15 years. No readily foreseen gas supplies are expected to change this. As time goes on, generating equipment does require renewal, and this provides an opportunity for alternative engines to be introduced. It has been seen from the recent renewal of generating capacity at the Vitoria power station on Madeira, Portugal, that a dual fuel engine was chosen as the basis for the generation. Whilst it is unclear whether this was due to known or perceived intelligence on potential gas supplies on the island (whether from imported LNG or from biogas sources), these units provide flexibility, as well as delivering efficiency savings compared to the units they are replacing. It remains though that pure diesel engines can offer higher efficiencies than dual fuel engines. With efficiency gains, for a given electrical output, lower fuel input will be required and consequently total emission loads (tonnes) would be expected to drop over time. These marginal gains in engine efficiencies will be in parallel to efficiency gains expected to be achieved by other combustion techniques (e.g. boilers, gas turbines), and so over time with efficiency gains by all combustion techniques, the relative share of emissions etc. attributable to diesel engines compared to other generating techniques may not be expected to change. The recently amended Gothenburg Protocol sets limit values for NO X emissions for new stationary diesel engines. The limit values that are set will require the use of secondary abatement equipment in new engines that operate more than 1500 hours per year, and (after the transition period of 10 years) in those cases where SCR cannot currently be applied for technical or logistical reasons such as on remote islands. In these cases described, the use of secondary aftertreatment may lead to less emphasis on primary NO X reduction measures; as the engines as manufactured will not need to meet specific limit values in the Gothenburg Protocol. But during the transition period of 10 years for diesel engine plants where SCR cannot currently be applied for technical or logistical reasons, and for those diesel engine plants which operate between 500 and 1500 hours per year, only primary measures for NO X reduction will be necessary. However, the industry expects that additional research and development will be necessary for slow and medium speed diesel engines between 5 and 20 MW th to meet a NO X emission level of 1,300 mg/nm 3 through using primary measures alone. For areas located within reach of natural gas networks, it may be possible that dual fuel engines could increase in market extent, but this is unlikely to progress much more quickly than natural turnover / renewal rates, and given that gas turbines will continue to provide competition. This process could be accelerated however if the relative gas/fuel oil price differential changes. The choice between a dual fuel engine and a pure diesel engine is an economic decision that will take into consideration not only the upfront investment costs but the potential changes in operational savings that could occur within the plant s long lifespan July

51 (at least 15 years, potentially up to 30 years), including e.g. expectations about future availability of gas. The potential operational savings from running on gas could easily justify the higher technological cost of a dual fuel engine. Particularly for isolated systems, for example on islands, in which electricity generation is from a combination of renewables and diesel engines, as the share of renewables increases over time (in response to other policy drivers), there is expected to be demand for greater flexibility from diesel engine generating plant to operate with higher variations of duty cycle (Eurelectric, 2011), which is not expected to lead to reductions in NO X emissions. Given that the major driver in the market for large diesel engines is in the maritime sector (Eurelectric, 2011), to a certain extent developments in the stationary diesel engine market will follow the developments occurring in the development of marine diesel engines. Whilst developments of regulations covering exhaust gases from marine diesel engines are being developed through MARPOL Annex VI and the NO X Technical Code, the requirements of these regulations are less stringent than land-based limit values in the Gothenburg Protocol. Therefore, there is not an expectation that NO X emission levels of stationary diesel engines will drop considerably due to development of marine diesel engines. The limit values for diesel engines in the non-road mobile machinery (NRMM) legislation is similarly not expected to influence the emissions of large stationary diesel engines due to the differences in sizes of engines (the largest locomotive engines are <5 MW) and operational profiles. 2.7 References AMEC (2012) Analysis and summary of the Member State s emission inventories and related information under the LCP Directive (2001/80/EC). CIMAC (2005) Position of the CIMAC WG 5 Exhaust Emissions Control on "Prime Mover Technique Specific Emission Limits Need Stationary Reciprocating Engine Plant". International Council on combustion engines tem_feb_2005rev.pdf EC (2006) Integrated Pollution Prevention and Control Reference Document on Best Available Techniques for Large Combustion Plants July Eurelectric (2011) Emissions from diesel generation in Small Island Power Systems - Recommendations for the revision of the Gothenburg protocol. A EURELECTRIC briefing document, July 2011 MAN (undated) MAN Diesel & Turbo SE Fuels. Webpage accessed 4 March UNECE (2012) Guidance document on control techniques for emissions of sulphur, NOx, VOCs, dust from stationary sources: 7.42 New Stationary Engines. E.zip VDMA (2011) Exhaust emission legislation: Diesel and gas engines, chapter: Stationary power plants. VDMA Engines and Systems. Available online at: Wärtsilä (undated) Wärtsilä dual-fuel engines. Webpage accessed 26 November July

52 July

53 3. Category (b) recovery boilers within installations for the production of pulp 3.1 Introduction The IED (Chapter III, Annex V) does not currently include emission limit values for recovery boilers used in the pulp and paper industry, but lists these plants amongst the combustion plants where the setting of such emission limits should be considered (Article 30(8)). The purpose of this chapter is focussed on looking at emission levels of currently operating recovery boilers in the EU and how these may change in the future. The version of the Production of Pulp, Paper and Board BREF that has been used for this chapter is the draft dated May 2012 (JRC, 2012). This chapter is set out with the following section: Section 3.2 describes recovery boilers by way of sectoral context, an overview of combustion sources in pulp plants, emissions and abatement measures. Section 3.3 notes legal definitions and other aspects of recovery boilers and how they are covered in legislation Section 3.4 describes the data collection process in this study and the methodology for assimilating the datasets for recovery boiler; and Section 3.5 summarises estimates of recovery boilers in the EU based on data collected and extrapolated as described in Section Description and background Sector overview (summarised from JRC, 2012) There are five main process routes for the production of pulp: Kraft (or sulphate) pulping, sulphite pulping, mechanical pulping, semi-mechanical pulping and recycled fibre processing. Of these, the Kraft process makes up around 60% of European pulp production, and the sulphite process makes up around 5% of production. Pulp may be produced from virgin fibre (via chemical or mechanical process routes) or from repulping recovered paper. Of the various pulp production routes, only the chemical pulping processes of Kraft and sulphite are within the scope of this study since recovery boilers are only used in these processes; consequently the scope of the pulp sector covered is part of virgin pulp production. 25 The market distribution of pulp production routes in Europe is shown below in Figure Some pulp mills produce both chemical pulp and mechanical pulp; only the chemical pulp production utilises recovery boilers. Recovery boilers in these sites are within the scope of this study. July

54 Figure 3.1 Chemical pulp production in Europe (data source: JRC, 2012) As the raw material for virgin pulp production, various species of wood are used, categorised as hardwoods and softwoods. The type of wood and its composition can ultimately affect the emission levels of a recovery boiler; in particular the use of eucalyptus. The amount of pulp produced from eucalyptus is around 50% in the Kraft process. Finland and Sweden dominate EU pulp production, with other significant pulp production in Spain, Portugal, Germany, Austria and France. According to JRC (2012), a total of 77 Kraft mills in Europe are distributed across Sweden (21), Finland (17), Spain (9), France (8), Portugal (6), Poland (3) and Austria (3); and a total of 16 sulphite pulp mills are split across Sweden (4), Germany (4), Austria (3) and single mills in five other countries Overview of the chemical pulp process In order to contextualise recovery boilers within a pulp production process, the figure below sets out the processes of a Kraft pulp mill. Figure 3.2 Overview of the main processes in a Kraft pulp mill (source: JRC, 2012); recovery boiler highlighted July

55 In brief summary the core processes of the chemical pulp production are the following. The raw materials are fed in through the wood handling unit that undertakes debarking, wood chipping and screening. The chips are then cooked in a solution (white liquor) of sodium hydroxide (NaOH) and sodium sulphide (Na 2 S) in order to separate the wood fibres, dissolve (and remove) the lignin component of the wood and remove extractives such as resin, fatty esters and esters. Following the cooking, whether in continuous or batch digesters, spent cooking liquor (weak black liquor) is removed from the pulp, destined for the recovery boiler. The waste liquor used in the post oxygen washing of the pulp is normally condensate from the black liquor evaporation. Delignification, undertaken with oxygen, is undertaken after cooking and washing of the pulp. The organic material that is dissolved during oxygen delignification is normally recovered via the recovery boiler. Bleaching is the next process, which is undertaken principally to increase brightness of the pulp; however the bleaching process wastes are discharged as effluent rather than recovered. A final process of drying is undertaken for market pulp; no drying is undertaken if the pulp is used directly within the integrated paper mill. Prior to incineration in the recovery boiler, the spent black liquor from the cooking process needs to be concentrated in order to increase its calorific value. This is undertaken through multiple stages of evaporation that increase the dissolved solid content of 14-18% to dry solid content of 70-85%. This evaporative process also yields a non-condensable gas (NCG) stream that includes H 2 S and other malodorous gas compounds; these gases are normally routed to the recovery boiler for incineration. The closed loop of the system is that the smelt containing recovered chemicals resulting from incineration of the concentrated black liquor in the recovery boiler is dissolved in water to produce green liquor. This green liquor which is composed predominantly of Na 2 S and sodium carbonate (Na 2 CO 3 ) is then causticised with lime where Na 2 CO 3 is converted to NaOH to produce white liquor for pulping. The figure below summarises this chemical recovery system. Figure 3.3 Overview of the Kraft chemical recovery process (source: Tran & Vakkilainnen, 2008) Combustion plants in pulp mills Chemical pulp mills utilise a number of different combustion plants for on-site steam and/or power generation, which differ in size, fuels, load conditions and purpose, and can be principally split into two categories: recovery boilers and other combustion sources. RECOVERY BOILERS The recovery boiler in a chemical pulp mill forms the central part of the installation s chemical and energy recovery system. The recovery boiler has four key functions: July

56 1. Minimising the environmental impact of waste materials (namely black liquor) from the pulping process, i.e. avoiding what would otherwise be a significant waste water load and its associated treatment; 2. Recovery of inorganic pulping chemicals (NaOH and Na 2 S); 3. Co-generation of steam and power to meet site demand by utilising the energy content of dissolved organic material; 4. Energy recovery of by-products (tall oil soap, crude tall oil and crude sulphate turpentine), which may be combusted in the pulp mill or sold to the chemical industry. JRC (2012) describes a recovery boiler as a steam boiler and chemical reactor in the Kraft and sulphite recovery system [which] burns black or brown liquor at high temperature and generates steam and power. All Kraft mills have a recovery boiler, but not all of the sulphite mills have recovery boilers. Recovery boilers are typically larger than 50MW th in capacity. In general the smaller recovery boilers are located in sulphite mills. The figure below gives an overview of a recovery boiler. Figure 3.4 Complete view of a boiler and schematic view of the lower and middle part of the furnace of a recovery boiler of a Kraft pulp mill (Source: JRC, 2012) OTHER COMBUSTION SOURCES In order to help meet core heat and electricity requirements, pulp installations include a number of other combustion sources, some exceeding 50MW th. These sources are not part of the chemical recovery system and so are distinct from recovery boilers, and not subject to the Article 30(9) review. July

57 ROUTING OF FLUE GASES In terms of applying the legal term combustion plant as defined in the IED through the common stack concept (Chapter III, article 29), a combustion plant at a chemical pulp mill could conceivably consist of: (i) solely recovery boiler(s), (ii) solely other boiler(s), or (iii) both recovery and other boilers. Most recovery boilers will have their own flue, and may or may not share a common stack with other combustion sources. It is not considered likely (but may be possible) that some chemical pulp mills may mix flue gas streams of recovery boilers with other sources. Since measurements of emission concentrations occurs in flues, flues that mix the flue gases of recovery boilers with other combustion sources could potentially pose an implementation issue with regards to compliance assessment of emission limit values set specifically for recovery boilers in these instances Fuels used in recovery boilers The black liquor, whilst being a fuel with energy content that is normally enough to make the Kraft pulp mills more than self-sufficient in heat and electrical energy (JRC, 2012), also contains valuable cooking chemicals. Recovery boilers in both Kraft and sulphite mills utilise the black liquor resulting from the pulping process as a primary feedstock. As indicated above, the black liquor is first concentrated through evaporative processes before being combusted in the recovery boiler. The actual concentration of the black liquor, expressed as a percentage of dry solid content, varies across Kraft pulp plants from 65% to 85%; the dry solid content affects SO 2 and NO X emissions as well as the plant s energy efficiency. In sulphite pulp plants, the dry solid content is lower, at around 58 to 60% which corresponds to a net calorific value of around 7,000 kj/kg (JRC, 2012). In addition to the black liquor, as indicated above, it is common for other fuels to also be combusted in the recovery boiler, including the NCGs and dried sludge; fuel oil or gas is normally used for start-up /shut-down or for supplementary firing for stability at low loads. Gas may also be used as a supplementary fuel. NCGs include both strong malodorous gases given off during evaporation plant when raising the black liquor dry solid content and weak malodorous gases that are emitted from other processes. The figure below shows reported feedstocks for recovery boilers (more information on the data source is in Section 3.4). Figure 3.5 Feedstocks in EU recovery boilers other than black liquor and supplementary conventional fuels (data source: EIPPCB) July

58 3.2.5 Emissions to air Contract /ENV/2012/627812/C3 Collection and Analysis of Data for the Review required Recovery boilers are a major source of atmospheric emissions in Kraft mills for NO X, dust (sodium sulphate and sodium carbonate), carbon monoxide, SO 2, and to a lesser extent, H 2 S. Recovery boilers in Kraft mills are similarly a major source of SO 2 emissions, with NO X, dust and CO emissions also being important. SULPHUR: SO 2 AND TRS EMISSIONS SO 2 emissions are predominantly formed through the oxidation of hydrogen sulphide (H 2 S) and carbonyl sulphide (COS) in the lower furnace of a recovery boiler. Total Reduced Sulphur (TRS) emissions from recovery boilers are primarily minor concentrations of malodorous compounds (hydrogen sulphide) (JRC, 2012). Sulphite mill recovery boilers have varying SO 2 emission profiles. Two different operating conditions can be differentiated: normal operating conditions, and periods in which cleaning of the scrubbers is undertaken. Cleaning is undertaken one scrubber at a time. During cleaning, SO 2 removal efficiencies are reduced. Higher SO 2 emission levels occur when the final washer is not in operation (cleaned around 15 days per year). For an example plant, annual average SO 2 emission levels that include all operating time may be 35% higher than annual average SO 2 emission levels of normal operating conditions only (JRC, 2012). Due to this variation, permit limit values expressed as daily averages may differentiate by operation mode in this regard. Table 3.1 lists the key operating variables affecting sulphur emissions from recovery boilers. Table 3.1 Operating variables affecting sulphur emissions from recovery boilers Operating variable Description SO 2 emissions Temperature of the different zones Sulphur to sodium ratio (S/Na 2) in the liquor (sulphidity) The load on the furnace The temperature in different zones is influenced by the DS content of the black liquor and the amount of combustion air. Higher DS feedstock increases combustion temperatures in the furnace (see load on furnace below) reducing sulphur dioxide emissions. However, higher temperatures may also increase NOx levels. The black liquor sulphidity is determined by the amount of sulphur and sodium needed in the cooking process, which depends on the pulp quality being produced. The sulphur is needed in order to achieve a higher yield and to produce a strong paper pulp. The pulp feedstock also affects sulphidity. Where there are higher levels of sulphidity, the sodium released by the furnace may be inadequate to bind all the sulphur released. This may result in a portion of the sulphur leaving the furnace as sulphur dioxide instead of sodium sulphate. High dry solid content of the black liquor may compensate this effect (JRC, 2012) The temperature in the furnace is directly proportional to the load of the furnace. Some boilers may be sensitive to changes in the load, for example SO 2 emissions may increase when the boiler load is reduced. For recovery boilers with a very high DS content, SO 2 emissions do not react to changes in the boiler load (JRC, 2012). Recovery boilers typically run at fixed loads; however some pulp mills run campaigns using different kinds of feedstock (wood types) and this will have an impact on the recovery boiler load. Recovery boilers will revert to lower loads if there are problems in other parts of the mill, e.g. if the digester is shut down for a couple of hours, the recovery boiler will continue running on a lower load using black liquor from the storage tanks. TRS emissions Dry solids content Load on furnace Air distribution For recovery boilers with a very high DS content, TRS emissions do not react to changes in the boiler load (JRC, 2012). TRS emissions are extremely low (or even zero) when boilers are stable and operated on a full load (M JRC, 2012). As with SO2 emissions, the load of the furnace is directly proportional to the temperature in the furnace. A recovery boiler operating on an overloaded mode may have adverse effects on emission characteristics, particularly on the quantity of hydrogen sulphide produced. Temporary peaks in TRS loads may result from rapid changes in the operation mode. TRS emissions from the furnace of a recovery boiler are determined by the mixing of oxygen and sulphur-containing gases. If these are adequately mixed they are converted to SO2. Modern air systems are designed to achieve sufficient mixing (JRC, 2012). July

59 NO X NOx emissions from the recovery boiler contribute 65-85% of the total NO X emissions from a Kraft pulp mill (JRC, 2012). Annual average NO X emission concentrations (6% oxygen, standard conditions) from Kraft mill recovery boilers, which use primary NO X reduction measures only, are around mg/nm 3 (JRC, 2012). 26 The nitrogen content in the black liquor is the principal cause for NOx emissions; variations in emissions are attributed to differences in black liquor composition. Although thermal NO X is normally a minority component, increases in NO X emissions can occur when the DS content of the black liquor is increased, since this leads to higher combustion temperatures and therefore greater thermal NO X. The figure below plots EIPPCB data on annual average NO X emission concentrations of EU recovery boilers, regardless of which primary measures are implemented at the recovery boilers, as a function of the dry solid content of the black liquor. The plot supports the relationship indicated in JRC (2012). Figure 3.6 NO X emission concentration as a function of dry solid content in black liquor (data source: EIPPCB) Different wood species have nitrogen in different forms: for example, the use of the hardwood eucalyptus as a feedstock typically used in mills on the Iberian peninsula leads to lower specific NO X emissions due to lower nitrogen content in the black liquor, as well as limiting the dry solid content concentration of the black liquor to around 72%, which affects both NO X and SO 2 emissions. NO X from sulphite mill recovery boilers is often higher than from recovery boilers in Kraft mills due to the higher temperatures of combustion (hence greater thermal NO X ). NO X emissions range from 175 to 500 mg/nm 3 (daily average) or from 170 to 400 mg/nm 3 (annual average) (JRC, 2012). 26 Excluding two outliers in the EIPPCB dataset, (83 mg/nm 3, 690 mg/nm 3 ), both of which are at Spanish mills. July

60 The table below summarises the factors affecting NO X emissions in recovery boilers. Table 3.2 Operating variables affecting NO X emissions from recovery boilers Operating variable Dry solids content of black liquor Excess O 2 and CO content during combustion Description Increases in the dry solid content of black liquor raises the temperature of the lower furnace and reduces the gas flow. Whilst this is undertaken to maximise steam generation, and it also reduces sulphur emissions, it increases NO X emissions, i.e. there is a trade-off between SO 2 and NO X emissions for operators and authorities to note when considering any increase in dry solids content. Low excess air for combustion reduces NO x emission levels, but as a trade-off also leads to moderately elevated CO emission levels. The Draft PP BREF suggests that CO emission concentrations in the range of 250 to 500 ppm are correlated with significantly lower NOx levels. If CO emission levels are maintained < mg CO/Nm 3 this leads to low emission levels of VOC and polyaromatic hydrocarbon (PAH) emissions and allows achieving lower NOx emissions levels. However, data from the EIPPCB do not clearly show the inverse relationship between CO and NO X emissions as a function of excess O 2, as shown in the figure below; this may be due to variation in additional measures taken to reduce NO X emissions at different boilers. The EIPPCB data do however show a correlation between NO X emission levels and real excess O 2: by grouping recovery boilers by real excess O 2 and averaging their reported NO X concentrations it is found that the average NO X emission level at 6-8% O 2 is 178mg/Nm 3, dropping to 168mg/Nm 3 at 4-6% O 2 and falling further to 149mg/Nm 3 at 2-4% O 2. Figure 3.7 NO X emission levels plotted against CO emission levels, separated by category of real excess oxygen concentration (data source: EIPPCB) The air system design and design of boiler Nitrogen content in black liquor Additional air inlets in the upper parts of a boiler reduce NOx emissions. New recovery boilers typically have an advanced air system which reduces their emissions by 20-30% (JRC, 2012). EIPPCB data gathered on recovery boilers includes the number of air level inlets that form the air distribution system in the boiler. The average data show the following results: Number of recovery boilers data represent Number of air inlets NOX emission concentration (annual average mg/nm 3, 6% O2) 31 3 or fewer 181 NA % % 4 More than % % reduction from 3 air inlets The nitrogen content of black liquor influences the NOx level of the black liquor recovery boiler. Approximately 30% of nitrogen in the black liquor is converted into NOx. Softwood generally has a lower nitrogen content than hardwood, therefore softwood black liquor results in approximately 20% less NOx emissions than hardwood (JRC, 2012). The majority of recovery boilers reported use >50% softwood; most of the recovery boilers in southern Member States primarily use hardwoods. July

61 Operating variable Combustion of NCG in the recovery boiler Burning other N- containing fuels Recovery boiler furnace load Pulp yield Description Although NCGs contain nitrogen, correct incineration of NCG may not lead to an increase in NOx emissions. Incinerating other fuels such as methanol, dissolving tank vent gasses (DTVG) or bio-sludge in the recovery boiler may lead to an increase in NOx emissions. NOx emissions are heavily affected by the load on the recovery boiler. A higher load may result in reaching the capacity constraints of the air system resulting in insufficient O 2, higher temperatures, short retention times and increased NOx emissions. A normal load results in effective optimization of all air levels resulting in optimal temperatures and retention times and minimum conversion to NOx emissions. Low loads result in low air demand in the furnace and low temperatures with long retention times, which lead to moderate conversion of NOx emissions (JRC, 2012). Higher yields lead to lower emission loads. For example, unbleached pulp has a higher yield than bleached pulp because it generates less black liquor per manufactured tonne of pulp compared to bleached pulp which gives lower specific emission loads. Pulp yields from hardwood (e.g. eucalyptus) are higher than from softwood (JRC, 2012). DUST Particulate matter (dust) emissions are predominantly formed when black liquor droplets are sprayed and incinerated in the recovery boiler; the primary component of the dust is particles of Na 2 SO 4. The size of the dust particles prior to abatement is <2.5 μm in diameter. All recovery boilers are fitted with electrostatic precipitators for the removal of dust; collected dust is fed back into the recovery boiler via the black liquor. Annual average dust emission concentrations as gathered by the EIPPCB for recovery boilers are in the range of 5 to 190 mg/nm 3 as measured after the dust after treatment. Dust emissions from sulphite mill recovery boilers range from 1 to 25 mg/nm 3 expressed as daily averages. Elevated dust emission levels can occur when the proportion of hardwood increases or when the SO 2 scrubbing system is partially offline (as noted above). JRC (2012) does not mention emissions of polyaromatic hydrocarbons (PAHs) as being a significant issue for recovery boilers. CO Emission of carbon monoxide (CO) result from improper combustion conditions. Elevated levels of CO often coincide with periods in which combustion conditions targeted at lowering NO X emissions are put in place (e.g. lower oxygen levels), i.e. CO can be a trade-off with NO X. Annual average CO emission concentrations (at 6% O 2 ) in Kraft mill recovery boilers range between 10 and 100 mg/nm 3, whilst concentrations in sulphite mills vary more widely and can be higher, ranging between 4 and 150 mg/nm 3 (JRC, 2012) Air emission abatement measures As stated in the introduction, this chapter draws on the text of the May 2012 draft PP BREF. The finalised PP BREF has not been available within the timescales of this study in order to be drawn upon. The information around abatement techniques, in particular the costs and benefits, may change in the final PP BREF. SULPHUR CONTROL: INCREASING THE DRY SOLID CONTENT OF BLACK LIQUOR Description Increasing the dry solid (DS) content in black liquor is an effective way to reduce both SO 2 and TRS emissions because it increases combustion temperatures. This increases the July

62 vaporisation of sodium (Na) which binds the SO 2 and forms sodium sulphate (Na 2 SO 4 ), thus reducing SO 2 emissions from the recovery boiler (JRC, 2012). The DS content of black liquor after the conventional evaporation phase is typically 70%. With the installation of a concentrator to the evaporation plant, the black liquor may achieve a DS content of 80% or above (up to 85%) (JRC, 2012). This may vary according to the species of the wood, for example plants operating with eucalyptus may find it difficult to achieve DS concentrations over 72% in their black liquor (JRC, 2012). Using a concentrator to increase the DS content of black liquor also maximises electricity production. Applicability, costs and benefits A super concentrator can be implemented at existing and new Kraft mills. Any viscosity problems can be controlled with pressurised storage or heat treatment before the last concentrator (JRC, 2012). Data collected by the EIPPCB of recovery boilers and their dry solid content and reported SO 2 emission levels help to demonstrate the potential benefits for SO 2 emissions due to increases in DS content. The figure below plots the data reported for EU recovery boilers of annual average SO 2 emission concentration against dry solid content, plotting those recovery boilers with end of pipe SO 2 abatement separately. The relationship is demonstrated for both recovery boilers with and without scrubbers. Figure 3.8 SO 2 emission concentrations of recovery boilers in the EU as a function of dry solid content of black liquor (data source: EIPPCB) Using the linear relationship estimated in the above figure for recovery boilers without scrubbers installed, the SO 2 abatement efficiency of the measure can be calculated on a per percentage point increase, and is for example 22% SO 2 reduction for an increase in dry solid content from 72% to 80%. JRC (2012) states that for existing mills, the cost to improve evaporation to increase the DS content of the black liquor is dependent on the target concentration. Investment costs of July

63 8-9m are quoted for a 1500 ADt/d kraft pulp mill to increase the dry solid concentration upwards from 63% to 80%; no additional operating costs are cited. Based on this range, and after amortising the investment costs over a period of 15 years (4% interest rate), we estimate the specific annualised costs for this measure to be from to 0.032m per year per DS percentage point increase per 1000 ADt/day (excluding potential cost savings). Increases in dry solids content of black liquor will result in significant savings, increases in energy economy of the mill and lead to gains in recovery boiler capacity. European mills considered under this study have a DS content that ranges from 60% to 82%. Potential Drawbacks A key draw back to increasing DS content of black liquor is that it may result in an increase of NOx emissions, thus a careful weighing up of emission reduction priorities is required. Additionally, this measure increases the emissions of particulates prior to flue-gas cleaning. An electrostatic precipitator would need to be installed to address this. Finally, for strong black liquor with a DS content >80%, a significant amount of sulphur compounds are released at the last evaporation stage, which would need to be collected and incinerated (JRC, 2012). SULPHUR CONTROL: INSTALLATIONS OF SCRUBBERS Description The installation of scrubbers on a recovery boiler can be applied as an alternative to increasing the DS content of black liquor. Flue-gas scrubbers for recovery boiler exhausts remove SO 2. Wet scrubbing technology, used for this process may include three stages; these are illustrated in Figure 3.9. Figure 3.9 Flue-gas scrubber for recovery boilers (source: JRC, 2012) Installations of scrubbers on the recovery boiler will reduce sulphur emissions and particulate matter emissions and lead to heat recovery. SO 2 reacts with the scrubbing liquor and Na 2 SO 3 and also some Na 2 SO 4 is formed. TRS in the form of H 2 S can be removed together with SO 2. However, removal of H 2 S from the flue-gases requires a high ph scrubbing liquor. At such a high ph, carbon dioxide would also be absorbed, which is July

64 unrealistic due to the relatively large amounts of carbon dioxide being formed in the combustion (JRC, 2012). Applicability, costs and benefits Whilst it is more straightforward and lower cost to fit scrubbers from new, it is possible to retrofit scrubbers onto existing recovery boilers. Capital costs of 7.2m to 10.4m are quoted for bleached Kraft mills of production capacities 250,000 and 500,000 t/yr respectively and operating costs of 0.58m/yr and 0.92m/yr respectively. These costs include the scrubber, scrubber liquor pumps, circulation pumps, electrification and instrumentation. Amortising the capital costs over 15 years (4% interest rate) leads to total annualised costs for scrubbers of 0.49m/yr per 100,000ADt/yr on production capacity 250,000 ADt/yr and 0.37m/yr per 100,000 ADt/yr on production capacity 500,000 ADt/yr Scrubbers may be less beneficial for recovery boilers running with high black liquor DS content, since these already have low sulphur emissions, however they may still be useful to reduce particulate matter emissions and for heat recovery. They can be used as a second part of a two stage dust removal facility consisting of ESPs and wet scrubbers (JRC, 2012). TRS CONTROL: INCINERATION OF STRONG AND/OR WEAK GASES IN THE RECOVERY BOILER Description Incineration of concentrated non condensable gases (CNCG) and diluted non condensable gases (DNCG) reduces TRS emissions (H 2 S) and associated odours. Applicability and benefits CNCG can be introduced into the secondary air level of the recovery boiler where its energy content is used to convert reduced sulphur compounds into SO 2. With sufficiently high DS content, SO 2 emissions from the recovery boiler in this case would be practically unaffected and additional abatement (e.g. scrubber) would not be necessary. For lower DS content, a scrubber may be necessary. DNCG can be incinerated in the secondary or tertiary air levels in the furnace of the recovery boiler. However, the flow rate of the tertiary air in the recovery boiler is limited, therefore alternatives may be needed. The measure can be adopted in new and existing mills. Potential drawbacks Incineration of TRS in a recovery boiler may increase NOx emissions due to the ammonia contained in odorous gases. This may be avoided if malodourous gases are injected correctly into the furnace in the right location. NO X EMISSIONS: OPTIMISED BLACK LIQUOR RECOVERY BOILER AIR SYSTEMS Description Modifications of the air stream system have proved effective at reducing NOx emissions, principally through additional levels. Conventionally, recovery boilers have three air levels; however additional feeding levels can be incorporated at a higher elevation in the boiler. Modification of air inlets via the introduction of additional air inlets in the upper section of the recovery boiler can reduce NOx formation. July

65 Applicability, costs and benefits Contract /ENV/2012/627812/C3 Collection and Analysis of Data for the Review required Application of this technique will result in lower NOx emissions. As identified in section 3.2.5, the potential NO X reduction from this measure of air staging could be 10% to 25%; a midvalue of 22% was selected in Entec (2010). New air inlets are applicable to new and existing boilers, however, boiler retrofitting would occur subject to the size and dimensions of the original boiler and restrictions may occur in some cases. Capital costs are quoted in the draft PP BREF as being 1.7m for a Kraft mill of production capacity 250 kt/year or 2.5m for production capacity 500 kt/year. A small additional operating cost due to the electricity consumption of an additional fan may occur but this has not been quantified in the draft BREF nor in Entec (2010). Amortising these investment costs over a 15 year period at 4% interest rate leads to an annualised specific cost range of 0.04 to 0.06 EUR / yr / 100ktpa. It is noted that this is an order of magnitude higher than the specific cost estimate in Entec (2010). NO X EMISSIONS: END OF PIPE TREATMENT Flue-gas treatment measures applied to the recovery boiler have the potential to effectively reduce total emissions. No flue gas treatment systems whether selective non-catalytic reduction (SNCR) or selective catalytic reduction (SCR) which are both end of pipe NO X control options for other combustion plants appear to be in place at EU recovery boilers according to EIPPCB data. No data is available on the use of secondary NO X control systems outside the EU. SNCR could potentially reduce NO X by between 30% and 50%, whilst SCR could potentially reduce NO X by 70% to 90%. Such systems have been tested on RBs in full scale trials. SNCR and SCR are proven technologies in a wide range of applications. SNCR Selective non-catalytic reaction (SNCR) occurs when ammonia is injected into hot flue-gases at approximately ⁰C. The ammonia reacts with the nitrogen oxide converting it into molecular nitrogen (N 2 ). The SNCR process can be undertaken within the upper part of the recovery boiler utilising numerous injection ports for reduction of chemicals at several levels; hence no additional equipment is required downstream of the boiler. The chemical reaction takes place within a narrow temperature band; if temperatures exceed this, ammonia oxidises to NOx and if temperatures are too low the reactions are slow and the ammonia leaves the boiler unreacted (termed ammonia slip ) (JRC, 2012). Changing boiler loads affects the flue gas temperature, such that the positioning of the SNCR within an optimal temperature window is difficult, resulting in varying NO reduction efficiency. This makes the application of this measure difficult. SNCR has been tested on a 200 tds/d recovery boiler. The results showed good reduction rates with limited ammonia slip over the load range % (JRC, 2012). However, outstanding questions remain over corrosion risks, how to evenly distribute chemical injections, and on ammonia dosing and overdose prevention. The avoidance of condensing acid gases to reduce corrosion risks ought to be possible through choice of construction materials and/or operational measures. SCR Selective catalytic reduction (SCR) occurs when ammonia is injected in a special catalytic reactor into cooled flue-gases. JRC (2012) noted that SCR could be considered an option for new recovery boilers under certain conditions (low dust application after the ESP, low SO 2 ) and that at least one supplier offers recovery boilers with SCR. July

66 No cost data are presented in the draft BREF. Notwithstanding the potential BAT conclusions on secondary aftertreatment, Entec (2010) estimated some potential costs for the fitting of SCR to recovery boilers; reproduced here in 2012 prices these are: Specific capital costs of 12,600 / capacity tpd Specific fixed operating costs of 420 / capacity tpd Specific variable operating costs of 4.9 / activity tpy DUST: ELECTROSTATIC PRECIPITATORS Description The performance of Electrostatic Precipitators (ESPs) varies according to the design and dimensioning of the electrostatic filters and operational parameters (e.g. the flue-gas flow rate). Material build-up may form an insulating layer on the electrodes reducing the electric fields which would impair performance, thus this needs close monitoring and control. Typical installation divides exhaust gases from the recovery boiler into two strings led in parallel through the ESPs. This ensures that they can operate interchangeably when maintenance cleaning is carried out in either string, without the need to shut down the boiler. Of the data reported to the EIPPCB (which does not cover every recovery boiler in the EU), all recovery boilers that were reported currently have ESPs fitted, albeit with varying numbers of fields (and presumably therefore varying removal efficiencies). Many recovery boilers are also fitted with scrubbers for SO 2 emission control, and these scrubbers also reduce dust emissions. The data on annual average dust emissions from EIPPCB have been plotted separately for recovery boilers with ESPs only from those with both ESPs and scrubbers in Figure 3.10 below. The plot shows that the annual average dust emission concentration of recovery boilers with ESPs only was 50 mg/nm 3 (standard conditions, 6% O 2 ) whereas the average for recovery boilers also fitted with scrubbers was lower at 35 mg/nm 3. For context, overlaid on the same plot is a line at 20 mg/nm 3 which is the emission level indicated in JRC (2012) as being achievable by well-maintained ESP-fitted recovery boilers. It is worth noting that many of the recovery boilers report operating at dust emission levels below 20 mg/nm 3. July

67 Figure 3.10 Dust emission concentrations of recovery boilers (data source: EIPPCB) Applicability and potential benefits ESPs are designed in accordance with the gas flow and removal efficiency required. Electricity consumption is positively correlated with removal efficiency of dust particles. Dust emission levels of 10 to 30 mg/nm 3 can be achieved as annual averages; well dimensioned, optimised and maintained ESPs could achieve 15 to 20 mg/nm 3 as daily average values. Existing ESP installations can be upgraded to limit costs; this includes the fitting of more modern electrodes, installing automatic voltage controllers or upgrading rapping systems. Performance of existing ESPs can also be improved by better alignment of emitting electrodes with connecting electrodes, elimination of gas leakage, ensure optimisation of electrical power and improvements to collecting plate cleaning. Existing ESP size may need to be increased where there has been an increase in gas and dust load to the ESP or where making changes to power supplies and controls may not be sufficient (JRC, 2012) ESP performance may also be affected by other factors, these include; black liquor composition, gas flow distribution and volume, improper rapping system design, outdated power supplies and controls and poor maintenance. Best performance of ESPs can be achieved when the ESP is kept clean and properly aligned and powered with the latest control system (JRC, 2012). Whilst no data on the economics of ESPs is available, given this is a BAU measure it is not necessary. 3.3 Existing legal provisions Industrial Emissions Directive The IED does not include a definition of recovery boilers. The only reference to recovery boilers in the IED is in Article 30(8), as follows: The emission limit values set out in Parts 1 and 2 of Annex V shall not apply to the following combustion plants: ( ); (b) recovery boilers within installations for the production of pulp. Article 30(8) makes clear that recovery boilers are covered by Chapter III. July

68 3.3.2 Member States Contract /ENV/2012/627812/C3 Collection and Analysis of Data for the Review required As part of the consultation on the recovery boilers operational in the Member States (see Section 3.4.2, ICF consulted Member State competent authorities regarding the applicable legislation governing the control of emissions from recovery boilers. Only the Swedish authorities responded to this consultation. SWEDEN In Sweden the local County Administrative Board (Länsstyrelsen) sets the emission limits individually for each RB/mill. They normally base their figures on recommendations from the Swedish Environmental Protection Agency (Naturvårdsverket) and other stakeholders. No country-wide ELVs are in place. The basic rules for setting conditions can be found in the Environmental Code from , which refers to the use of BAT in Section 3 of Chapter II. One dust ELV condition set by the authorities in 2012 for a RB in Sweden was 60 mg/nm 3 as an annual average Non-EU countries UNITED STATES Country Name of legislation Link to legislation Timescale of regulation United States 40 CFR Part 63 National Emission Standards for Hazardous Air Pollutants for Chemical Recovery Combustion Sources at Kraft, Soda, Sulfite, and Stand-Alone Semichemical Pulp Mills (MACT II Standards under NESHAP subpart MM) Rule and Implementation Information for Pulp and Paper Production MACT Federal Register, Rules and Regulations document (Subpart MM) Current Relevant clause Details Applicability Section A MACT II, regulates chemical recovery combustion sources at kraft, soda, sulfite, and standalone semichemical pulp mills. Scope Emission limit values The affected sources that are regulated are each new nondirect contact evaporator (NDCE) recovery furnace and associated smelt dissolving tank (SDT) located at a kraft or soda pulp mill, each new direct contact evaporator (DCE) recovery furnace system and associated SDT located at a kraft or soda pulp mill, each new lime kiln located at a kraft or sulfite combustion unit located at a sulfite pulp mill, each new or existing semichemical combustion unit located at a stand-alone semichemical pulp mill, and each existing chemical recovery system located at a kraft or soda pulp mill. The chemical recovery system is defined as all existing DCE and NDCE recovery furnaces, SDT, and lime kilns at a kraft or soda pulp mill. For the following ELVs it is not clear whether the dry standard cubic metres are equivalent (i.e. temperature and pressure) to Nm 3. For recovery boilers in Kraft mills: PM 0.10 grams/ dry standard cubic metre at 8% oxygen for existing plant PM grams/ dry standard cubic metre at 8% oxygen for new plant Bubble concept for PM emissions from existing Kraft mills Gaseous organic hazardous air pollutants (HAP) kg / tonne of black liquor solids For recovery boilers in sulphite mills: PM grams/ dry standard cubic metre at 8% oxygen for existing plant PM grams/ dry standard cubic metre at 8% oxygen for new plant 27 Year 2000 version English translation available at July

69 Country Name of legislation Link to legislation Timescale of regulation United States New Source Performance Standards Subpart BB Current Applicability Scope Emission limit values Relevant clause Applicability and designation of affected facility Applicability and designation of affected facility Standard for particulate matter Standard for total reduced sulfur (TRS). Details Includes provisions for the following facilities in kraft pulp mills: Digester system, brown stock washer system, multiple-effect evaporator system, recovery furnace, smelt dissolving tank, lime kiln, and condensate stripper system. In pulp mills where kraft pulping is combined with neutral sulfite semichemical pulping, the provisions of this subpart are applicable when any portion of the material charged to an affected facility is produced by the kraft pulping operation. Facilities under construction or modification after September 24, Standard for particulate matter. (a) On and after the date on which the performance test required to be conducted by 60.8 is completed, no owner or operator subject to the provisions of this subpart shall cause to be discharged into the atmosphere: (1) From any recovery furnace any gases which: (i) Contain particulate matter in excess of 0.10 g/dscm (0.044 gr/dscf) corrected to 8 percent oxygen. (ii) Exhibit 35 percent opacity or greater. [43 FR 7572, Feb. 23, 1978, as amended at 65 FR 61758, Oct. 17, 2000] Standard for total reduced sulfur (TRS). (a) On and after the date on which the performance test required to be conducted by 60.8 is completed, no owner or operator subject to the provisions of this subpart shall cause to be discharged into the atmosphere: (2) From any straight kraft recovery furnace any gases which contain TRS in excess of 5 ppm by volume on a dry basis, corrected to 8 percent oxygen. (3) From any cross recovery furnace any gases which contain TRS in excess of 25 ppm by volume on a dry basis, corrected to 8 percent oxygen. [43 FR 7572, Feb. 23, 1978, as amended at 50 FR 6317, Feb. 14, 1985; 51 FR 18544, May 20, 1986; 65 FR 61758, Oct. 17, 2000] July

70 AUSTRALIA Country Name of legislation Link to legislation Timescale of regulation Australia New South Wales Protection of the Environment Operations (Clean Air) Regulation Other Useful Links: NSW Government. Environmental Issues Webpage Current Applicability Scope Emission limit values Relevant clause Part 5 of the Protection of the Environment Operations regulations Division 2 Standards for scheduled premises 32. General grouping of activities and plant Schedule 3: Standards of concentration for scheduled premises: activities and plant used for specific purposes Details 'Air impurities from emitted activities and plant' largely focuses on industrial, agricultural and commercial scheduled activities but also specifies some requirements for non-scheduled activities. The plants are specified according to groups: Group 1 (plants licensed before Jan 1972) Group 2,3 or 4 (plants licensed between Jan 1972 to August 1997) Group 5 ( plants licensed between Aug 1997 and Sept 2005) Group 6 (plants licensed after Sept 2005) Schedule 3 specifies SO2, NOx and dust/pm for Paper, paper pulp or pulp products industries amongst other industries (Petroleum refining, Petrochemical production, Electricity generation). These are detailed in the table below. Compared to the emission limits for solid particles, the dataset developed in this study (summarised later in Section 3.5) suggests that 59% of EU recovery boiler capacity may have dust emission levels in compliance with Group 6, with a further 38% in compliance with the ELV for Group 5. The small remainder may meet the ELV for Groups 2/3/4. For NO X emissions, our estimates suggests that almost all (98%) of EU recovery boiler capacity may have NO X emission levels in compliance with the Group 6 ELV (300mg/Nm 3 ), with the remainder with emission levels under the Group 5 ELV. For TRS emissions, our estimates suggests that 62% of EU recovery boiler capacity may have TRS emission levels in compliance with the Group 6 ELV (4mg/Nm 3 ). Pollutant Standard of concentration for Kraft recovery boilers Solid particles (Total) Group mg/m 3 Nitrogen dioxide (NO 2) or nitric oxide (NO) or both, as NO 2 equivalent Group 2, 3 or mg/m 3 Group mg/m 3 Group 6 50 mg/m 3 Group 1, 2, 3 or 4 2,500 mg/m 3 Group 5 2,000 mg/m 3 Group mg/m 3 Hydrogen sulfide (H 2S) Group 1, 2, 3, 4, 5 or 6 5 mg/m 3 Total reduced sulfides (TRS), as H 2S equivalent Group 6 4 mg/m 3 Dioxins or furans Group ng/m 3 Group kg/t of black liquor solids fired Smoke Group 1, in approved circumstances Ringelmann 3 or 60% opacity Group 1, in other circumstances Group 2, 3, 4, 5 or 6, in approved circumstances Group 2, 3, 4, 5 or 6, in other circumstances Ringelmann 2 or 40% opacity Ringelmann 3 or 60% opacity Ringelmann 1 or 20% opacity July

71 3.4 Data collection exercise for this study Overview The method to most robustly estimate the total emissions and related relevant data for recovery boilers in the EU was to take a bottom-up approach to inventorying recovery boilers. This approach has been attempted for this study on the basis of the data sources available (in particular from EIPPCB) and through the cooperation of the key industry stakeholder, CEPI. The inventory has been sought at the recovery boiler level. The existing LCP emission inventory for 2009, which has been provided to ICF for the purposes of this study 28, does not necessarily include recovery boilers among the LCPs inventoried due to varied implementation issues among Member States. 29 As such, the LCP emission inventory cannot be used as a sole data source of recovery boilers in the EU. For this study, two principal supplementary data sources have been developed and investigated, and used in combination for this study. The first is the development of a list of pulp production installations with recovery boilers across the EU Member States. This list has been developed based on previous work by Entec (2010), consolidated and updated substantially by CEPI, and is in the process of being confirmed with Member States via direct consultation. A second data source that has been utilised is of data submitted by operators to the EIPPCB ( EIPPCB dataset ) as part of the PP BREF review process that is at the time of writing (March 2013) nearing completion. These data have been provided to ICF 30 as anonymous data for the purposes of this study. These data list at a recovery boiler level various technical aspects covering capacity, fuel type, emissions and abatement techniques. The two data sources are at a plant or unit level. Whilst the EIPPCB dataset is anonymous, for the purposes of the estimation matching up the plant level data from the EIPPCB to the recovery boiler inventory has been partially undertaken. Any potential mismatches bear no impact on estimates of total emissions at Member State or EU level due to the methods used Methodology for developing an inventory of recovery boilers LCP INVENTORY The 2009 LCP inventory was investigated for its potential suitability for an inventory of EU recovery boilers However, the LCP inventory data is thought to be inconsistent with regard to inclusion or exclusion of recovery boilers reported by the Member States. This was concluded from comparisons across Member States of known chemical pulp mills and their reported LCPs and noting the fuel types that had been reported. Consultation with Member State authorities in this study confirms that data from the LCP inventory does not generally include recovery boilers and therefore could not be a basis for an inventory of EU recovery boilers. RECOVERY BOILER INVENTORY A comprehensive bottom-up inventory of recovery boilers in the EU was targeted in order to most robustly estimate the emissions and potential emission reductions of recovery boilers. 28 Personal Communication with the Commission, 7 September Based on personal communications with Member State competent authorities. 30 Personal Communication with EIPPCB, 14 th November July

72 An initial list of pulp installations with recovery boilers was taken from the installation database developed in Entec (2010). 31 The list of locations with pulp installation recovery boilers from this Entec dataset was further developed considerably by the Confederation of European Paper Industries (CEPI) through: Expansion of the list where chemical pulp installations were considered by CEPI to be missing in the Entec list (37 mills); Removal from the list of chemical pulp mills that were known to be closed (19 mills); Removal from the list of mills known to not be chemical pulp mills (e.g. mechanical mills without recovery boilers) (circa 50 entries in the dataset); and Other corrections and updates, e.g. names of operating companies. Following this process, the resulting draft consolidated list of mills considered to have recovery boilers has been matched where possible at an installation level by ICF to LCP(s) at the same locations in the 2009 LCP inventory. This provides the possibility to compare recovery boilers at a mill with reported LCPs. Finally, ICF has consulted Member State experts identified in the Pulp and Paper BREF Technical Working Group on the draft consolidated list, which has resulted in a final consolidated inventory of recovery boilers. The table below summarises the status of consultation with Member States. Table 3.3 Member State Austria Belgium Bulgaria Czech Republic Finland France Germany Italy Poland Portugal Romania Slovakia Spain Sweden Status of consultation with Member States on operational recovery boilers Status / summary Confirmation received. Additional data confirming numbers, capacities, locations and emission levels supplied. Confirmation received. No confirmation received No confirmation received Confirmation received. Additional data on steam capacities and plant identification supplied. Confirmation received. Rated thermal input data supplied. Confirmation received. Rated thermal input and activity data supplied. Confirmation received (recovery boiler no longer operating following production cessation). Confirmation received. Rated thermal input data supplied. Confirmation received. Rated thermal input data supplied. Member State not represented on BREF TWG. Member State not consulted. Data confirmed via alternative sources: REC (2005) and Member State not represented on BREF TWG. Alternative contact sought unsuccessfully. Data confirmation attempted through literature review. Contact has confirmed verbally that there are 9 RBs over 50MW th in Spain. This is consistent with assumptions in this study. Confirmation received. Rated thermal input data supplied. NB: Member States not listed in this table are estimated to not have any recovery boilers. 31 Personal Communication with AMEC Environment & Infrastructure UK, 9 October 2012 July

73 In addition, the Croatian authorities have been consulted but no recovery boilers were identified. EIPPCB DATA ON RECOVERY BOILERS The data gathered by way of a questionnaire by the EIPPCB to underpin the PP BREF development has been shared with ICF for the purposes of study. 30 The EIPPCB data have been anonymised and represent a portion of the total data gathered from operators about pulp plants. The data represent a large number of recovery boilers around Europe. The dataset includes data on emission levels, abatement techniques and capacities of the recovery boilers. These data in particular have been identified as the most representative dataset for estimating emissions from the recovery boilers. They are actual data, with the vast majority being for the years 2008 or 2010 The BAU uptake of different abatement measures and related operational characteristics from the EIPPCB dataset are: 36% of RB capacity uses scrubbers 100% of RB capacity uses ESPs 0% of RB capacity uses SNCR or SCR On average the number of air level inlets is 4. 42% of RB capacity reduced NOX emissions through control of O2 supply GENERAL METHODOLOGY The availability of a detailed and recent dataset of emission levels for individual recovery boilers, which covers more than half of all recovery boilers in the EU and thus could be considered to be sufficiently representative of remaining recovery boilers in the EU, allowed to use the method of extrapolation is for the estimation of emissions from recovery boilers. Due to the quality of the underlying datasets the method of extrapolation in this instance is considered superior (in terms of accuracy) to estimating emissions from legislative limit values. The general methodology adopted for the development of an inventory of EU recovery boilers is as follows: Estimate additional data for recovery boilers in EIPPCB dataset of rated thermal input, recovery boiler daily capacity, pulp production activity, annual emission loads Link inventory recovery boilers with individual (but anonymised) recovery boilers listed in EIPPCB and utilise recovery boiler level data from EIPPCB Develop Member State and EU averages for variables such as emission levels based on EIPPCB data For inventoried recovery boilers without linked EIPPCB data, apply Member State average data. For inventoried recovery boilers in Member States for which no data exist in EIPPCB dataset, apply EU average assumptions. Where data reported by Member States on specific recovery boilers has been provided by Member States or found through literature review this is taken preferentially over anonymised EIPPCB data. In order to undertake the above methodology, the following assumptions were necessary: July

74 Assumption Net calorific value of black liquor Description An example gross calorific value (GCV) of black liquor has been taken from JRC (2012), which is for a Finnish softwood Kraft mill, of MJ/kg. A net calorific value (NCV) has been derived from this GCV using the following equation: [ ( )] ( ) Where H and S are the weight fractions of elements hydrogen and sulphur in the black liquor solids respectively. Values for H and S of 3.20% and 6.20% respectively have been taken from JRC (2012) on page 235. Resulting in an assumed NCV of 10.2 MJ/kg. Recovery boiler capacity (tds/day) Recovery boiler capacity or rated thermal input (MW th) Daily activity / production (ADt/day) Annual activity / production (ADt/yr) Annual emissions from recovery boilers (tonnes) Average capacity and activity The EIPPCB dataset included a field for the recovery boiler average liquid firing rate, assumed to be in units of tonnes of dry solids per day (tds/day). Based on the adjacent comments in the dataset these data have been interpreted as the capacity of the recovery boilers in tds/day. Derived from capacity in tds/day using the assumed NCV for black liquor. Estimated from capacity using factors 1,37 tds/adt for hardwood and 1,69 tds/adt for softwood (source: COWI expert)and coupled with plant level data on proportions of wood being soft or hardwood (source: EIPPCB data). Due to a lack of information on which processes follow downstream of the digestor, no assumptions were made on yield losses for these processes: loss in screening after cooking, loss in oxygen delignification, in bleach plant. Estimated from daily activity by assuming production runs at a rate equivalent to maximum capacity 82% of the year. Factor is based on an average of two sources: (1) 77% derived from average of 6 German recovery boilers annual activity levels compared to their capacities; (2) COWI expert judgement that a modern mill will normally run on 90% load as an average when in operation, and once taking account of shutdown/maintenance stops of approximately 10 days per year, makes for a load of 87%. Estimated by multiplying reported (or average) annual average specific emissions in kg [pollutant]/adt with the annual production. For those recovery boilers inventoried without specific links to EIPPCB or Member State provided data, averages based on EIPPCB data are applied as follows: Member State Capacity (t DS/ day) Activity (ADt/day) Sweden 1,926 1,190 All other Member States applied EU averages of: 1,890 1, Overview of recovery boilers in the EU This section provides a summary of the recovery boiler inventory developed for the EU and described by the methodology in Section 3.4. It represents a total for Kraft and sulphite mills combined Number of recovery boilers The total number of recovery boilers in the EU is estimated to be 94. The majority of the recovery boilers are single recovery boilers at a pulp installation and so can be considered to follow a rule of one stack (combustion plant) per recovery boiler. For the few installations with multiple recovery boilers, a clear determination of stack aggregation has not been confirmed by every relevant Member State. However, comparisons with the LCP inventories reported by Member States for year 2009 suggest that the two sites in Austria with multiple recovery boilers can be considered to have a stack each; whereas the two pulp installations in Germany with multiple recovery boilers have been reported as single combustion plants. Therefore the total number of combustion plants that the 94 recovery boilers represent is likely to be slightly lower than this (92 or below). July

75 The auxiliary combustion sources at many pulp installations already form combustion plants of 50 MW th or more, as reported in the LCP inventory of 2009, and it is understood that recovery boilers often share stacks with these other combustion sources. Consequently, depending on the rated thermal input of the other combustion sources at each individual pulp site, if recovery boilers were to be consistently reported as (part of) combustion plants then the total number of combustion plants reported by a Member State may not increase (however, other aspects such as rated thermal input, fuel energy input and emissions would increase). Based on the data gathered and developed, the recovery boilers are spread around the EU Member States according to the pie chart in Figure Figure 3.11 Number of recovery boilers in each Member State (source: this study) The numbers of plants presented in the above figure represents best estimates for currently operating recovery boilers. Specific points to note about the figures: The figures include a recovery boiler in Austria of known rated thermal input 45 MW. It is thought that the total rated thermal input at stack level for this recovery boiler may be less than 50 MW despite considering any additional combustion sources that may have flue gases exhausting through the same stack. It is known that this recovery boiler is expected to be imminently superseded by a new recovery boiler (which may or may not exceed 50 MW th ) that is presently under construction; the superseded 45 MW th boiler will remain as standby capacity. Presently two, and in the near future three, of the recovery boilers listed for Austria operate as standby capacity, and are functioning only in small capacity each year. A recovery boiler operating in Austria of known rated thermal input 21 MW th is not included in the above figures as it is considered likely that this unit (together with other combustion sources at its site) will not exceed the minimum rated thermal input threshold for coverage under Chapter III of the IED. In Sweden, two recovery boilers identified by the authorities as being less than 50MW th in capacity (17MW th and 25MW th ) have been excluded as the estimated total rated thermal input of the recovery boiler together with auxiliary boilers is less than 50MW th. A further recovery boiler of rated thermal input 43MW th has been included in July

76 Contract /ENV/2012/627812/C3 Collection and Analysis of Data for the Review required the figures as the combined rated thermal input (assuming common stack) of the RB with additional combustion sources does exceed 50MW th. One recovery boiler in each of Portugal, Romania and Spain are estimated to have rated thermal inputs of 49MW th, 44MW th and 43MW th respectively. These boilers have not been excluded from the figures since it is considered possible that with additional combustion sources routed via the same stack, they may form combustion plants greater than or equal to 50 MW th. In terms of context of these numbers of recovery boilers compared to the total number of combustion plants 50MW th in the EU they represent a small proportion (around 3%). As indicated above, due to the varied interpretation by Member States of whether to include recovery boilers within the LCP inventories, the figure of 94 recovery boilers would not be exactly additional to the total number of LCPs. A response has not yet been received from all relevant Member States as to whether their LCP inventories include recovery boilers Capacities of the recovery boilers The total estimated rated thermal input of the recovery boilers in the EU is 18 GW th. The split of this total capacity across capacity classes is shown below in Figure The figure shows that we estimate around five-sixths of recovery boiler installed capacity to be in the 100 to 500 MW th range. Figure 3.12 Estimated rated thermal input of recovery boilers in the EU (source: this study) These estimates are for recovery boilers only and not any additional potential combustion sources that could form part of a combustion plant (common stack). The total EU capacity estimate compared to the EU total capacity of LCPs shows that recovery boiler capacity in terms of rated thermal input is approximately 1.3% that of all EU combustion plants of 50 MW th or more Current emissions The total current annual emissions for recovery boilers have been estimated according to the methodology noted in Section 3.4. It is important to reiterate here that some of the assumptions that were necessary to make have significant uncertainties, e.g. the estimation of annual activity of the plants which was used in combination with specific pollutant loads. The annual emission concentration data are based on the reported data provided by EIPPCB which are a mixture of reporting years 2008 and 2010; for comparison the LCP and NECD inventories of 2009 has been selected. July

77 Table 3.4 below presents the estimated current SO 2, NO X, dust and TRS emissions for recovery boilers in the EU, and compares the estimates to total emission estimates reported for LCPs from the 2009 LCP inventory, and against national totals and total industrial combustion emissions 32 from inventories reported for 2009 under NECD. The comparisons suggest that SO 2 emissions are very small compared to other industrial combustion sources, whilst NO X emissions are an appreciable contribution. Dust emissions are estimated to form a more significant component of total LCP dust emissions: around 6%. Table 3.4 Estimated SO 2, NO X, dust and TRS emission levels (mg/nm 3 ) and emissions (kt per annum) from recovery boilers in the EU, compared to reported annual total LCP emissions, NECD total and total industrial combustion emissions Pollutant Recovery boilers (estimated in this study) Comparison with LCP inventory emissions Comparison with NECD emissions Total Industrial combustion EU average of emission concentrations (mg/nm 3, 6% O 2, annual average, standard conditions) Total EU annual emission (kt/yr) 2009 LCP (kt/yr) % of LCP 2009 NECD (kt/yr) % of NECD 2009 NECD (kt/yr) % of NECD SO , % 4, % 3, % NO X , % 9, % 2, % Dust % N/A N/A N/A N/A TRS N/A N/A N/A N/A N/A N/A The estimates of emissions at a Member State level is presented below in Table 3.5. This table shows that of the total EU estimated dust emissions from recovery boilers of 6.5 kt, 68% are estimated to originate from Finnish and Swedish recovery boilers combined. Compared with these Member States share of total capacity (61%) of recovery boilers, this implies that estimated dust emission levels from these plants are on average slightly higher than for other Member States. Table 3.5 Estimated SO 2, NO X, dust and TRS emissions (kt per annum) from recovery boilers per Member State (source: this study) Member State Annual SO 2 emissions (kt/yr) Annual NO X emissions (kt/yr) Annual dust emissions (kt/yr) Annual TRS emissions (kt/yr) Austria Belgium Bulgaria Czech Republic Total industrial combustion is taken as being the sum total of the following NFR codes, per AMEC (2012a): 1A1a (Public Electricity and Heat Production); 1A1b (Petroleum refining); 1A1c (Manufacture of solid fuels and other energy industries); 1A2a, 1A2b, 1A2c, 1A2d, 1A2e, 1A2fi (Stationary combustion in manufacturing industries and construction: Iron and Steel, Non-ferrous metals,: Chemicals, Pulp, Paper and Print, Food processing, beverages and tobacco, and Other). July

78 Member State Annual SO 2 emissions (kt/yr) Annual NO X emissions (kt/yr) Annual dust emissions (kt/yr) Annual TRS emissions (kt/yr) Finland France Germany Poland Portugal Romania Slovakia Spain Sweden EU Emission reduction potential SO 2 EMISSIONS As identified above in Section 3.2.6, two main methods for the reduction of SO 2 emissions have been identified: increasing the dry solid content in the black liquor and fitting scrubbers. In order to assess the potential SO 2 emission reductions from increasing the dry solid content, the potential increase in dry solid content has been assessed at the recovery boiler level: i.e. noting that (i) this measure would typically not need to be undertaken with scrubbers already in place and so has been assumed to be applied only to recovery boilers without scrubbers; and (ii) that recovery boilers based on eucalyptus have limited potential to increase dry solid content and so the measure is not applied to these recovery boilers. The impact on NO X emissions has been quantified. Fitting scrubbers has been modelled only against those recovery boilers that do not already have scrubbers installed; abatement efficiencies of 90% SO 2 and 50% dust have been assumed. The table below presents the estimated total EU emission reductions and total EU costs of these scenarios. Table 3.6 Estimated effectiveness and costs of SO 2 abatement scenarios Scenario SO 2 emissions (t/yr) SO 2 emission reduction (t/yr) SO 2 emission reduction (%) NOx emission change (t/yr) Dust emission change (t/yr) Estimated costs ( m/yr) Baseline estimate 5,504 Abatement scenario 1: increase dry solid content of black liquor at plants without scrubbers and which do not use eucalyptus. Abatement scenario 2: fit scrubbers to recovery boilers that do not already have scrubbers fitted 4, % + 2,823 NA ,423 3,080 56% NA -2, July

79 NO X EMISSIONS As identified above in Section 3.2.6, methods for the reduction of NO X emissions from recovery boilers consist of optimising various operating variables. Since only costs are available for the optimisation of the air system (air staging) this variable is assessed quantitatively. Air staging is assumed to be applicable to all recovery boilers with 4 or fewer air inlets, and a NO X abatement efficiency of 22% is assumed. The potential application of end of pipe treatment SCR is also modelled, although it is recognised that this measure may be unproven in commercial recovery boiler application. The table below presents the estimated total EU emission reductions and total EU costs of these scenarios. Table 3.7 Estimated effectiveness and costs of NOX abatement scenarios Scenario NO X emissions (t/yr) NO X emission reduction (t/yr) NO X emission reduction (%) Estimated costs ( m/yr) Baseline estimate 35,157 Abatement scenario 3: optimising air systems. Abatement scenario 4: fitting SCR 28,957 6,200 18% 13 7,031 28,126 80% 358 DUST EMISSIONS All the recovery boilers are estimated to already have ESPs fitted. Section identified that well-optimised and maintained ESPs can achieve 20mg/Nm 3 as an annual average dust emission concentration. A scenario is produced which assumes that dust emission concentrations of 20mg/Nm 3 are achieved by all recovery boilers regardless of whether scrubbers are additionally fitted or not. 33 (). No estimates of costs are developed since this is not available in JRC (2012). The table below presents the estimated total EU emission reductions of this scenario. Table 3.8 Estimated effectiveness and costs of dust abatement scenarios Scenario Dust emissions (t/yr) Dust emission reduction (t/yr) Dust emission reduction (%) Estimated costs ( m/yr) Baseline estimate 7,694 Abatement scenario 5: optimise and maintain ESPs to achieve 20 mg dust/nm 3. 3,379 4,315 56% N/A 3.6 Future trends Methods of pulp production As noted in section 3.2.1, pulp production from virgin fibre via the chemical routes of Kraft (sulphate) or sulphite process is not the only method for pulp production. In its 2050 roadmap document, CEPI identifies that by 2050 the balance between use of recycled fibre versus virgin fibre is expected to shift more in favour of recycled fibre (CEPI, 2011). Whilst not explicitly stating the consequence of this on the chemical pulp sector, the 2050 pathway 33 Figure 3.10 identified that 20mg/Nm 3 is already achieved by recovery boilers both with and without scrubbers. July

80 would suggest that, for fixed total levels of pulp production, this would see a decline in the chemical pulp sector. However, pulp is a globally marketable product, and total production levels are not necessarily projected to decline. Research is being undertaken in Sweden and Finland regarding new pulping methods aiming for new products from cellulose. Dissolving pulp, used for production of viscose to replace e.g. cotton in clothing is one trend. Also, as an example, some mills are looking at (and are doing it in same cases) extracting and upgrading valuable chemicals from the black liquor. Tall oil can for instance be used as a raw material for green diesel. Other methods of removing lignin from the black liquor is also available, e.g. the Lignoboost technology marketed by Metso Black liquor gasification Black liquor gasification (BLG) could have the potential to replace the use of recovery boilers in the future. Expected application is between and could be of direct benefit to the pulp and paper sector (CEPI, 2011). The technique converts concentrated black liquor into inorganic compounds which can be used to recover cooking chemicals as per a recovery boiler as well as combustible fuel gas consisting mainly of hydrogen, carbon monoxide and carbon dioxide (which can be removed downstream of the gasifer if the syngas is to be used to e.g. produce motor fuels,. The syngas can be fired in a gas turbine for power generation. Hot flue gases from this process generate steam in a waste heat boiler and the resulting high pressure steam generates additional power in a steam turbine. BLG can also produce syngas as a raw material for making higher value chemicals (JRC, 2012) or biofuels for the road transport sector (Nykomb, 2005). Future trends in the pulp technology sector may favour BLG technology, since 12% of mills have a capacity greater than ADt/yr (minimum capacity needed to achieve power efficiency) and this figure is set to increase. Additionally, as recovery boilers reach the end of their operational lifecycle, they will need to be replaced (JRC, 2012). Possible environmental benefits from BLG include increased electric power generation; low emissions to the atmosphere and increases in production capacity. However, the developments in technology may be hindered by several factors including capital (the investment cost for a BLG combined cycle is of the order of 50% higher than for a recovery boiler); no long term affordable and reliable refractory system 34 to protect the gasifier from severe conditions; achieving adequate purification of syngas required for the gas turbine to work and finally increased pressure on the causticising department Partial Borate Autocaustising The primary autocausticising reaction to take place in a recovery boiler furnace is when sodium metaborate (NaBO2) and sodium carbonate (Na2CO3) form trisodium borate (Na3BO3) in the recovery boiler smelt. Trisodium borate reacts with water in the smelt dissolving tank to form sodium hydroxide (NaOH) and regenerate NaBO2. The partial borate autocausticising 34 The new refractory material used by Chemrec is performing better. July

81 process occurs when sodium borates are added to the Kraft liquor at substoichiometric levels [ 449, M.Björk et al ]. A portion of the sodium carbonate (Na2CO3) is causticised in the recovery boiler. The causticisation of the remaining (Na2CO3) is completed in a conventional recausticising plant of the pulp mill with a reduced quantity of lime JRC (2012). APPLICABILITY This technology is particularly applicable to Kraft mills where incremental causticing and lime kiln capacity are required. The technique has been tested in several mills since 2000 with positive results. As such, it is being used successfully by Kraft pulp mills in the US, Canada and Sweden. Since the autocaustising reaction takes place in the recovery boiler no additional capital investment is required (JRC, 2012). Environmental benefits as a result of this technique include reduced emissions from the lime kiln in addition to reduced energy consumption by lime kiln. Furthermore studies have suggest borate present in cooking liquor increases pulp yield, may decrease rejects, improves the selectivity of lignin removal and can increase pulp viscosity at the same kappa number (JRC, 2012). 3.7 References AMEC (2012a) Analysis and summary of the Member State s emission inventories and related information under the LCP Directive (2001/80/EC). AMEC (2012b) Collection and analysis of data to support the Commission in reporting in line with Article 73(2)(a) of Directive 2010/75/EU on industrial emissions on the need to control emissions from the combustion of fuels in installations with a total rated thermal input below 50MW. Final report for the European Commission. AMEC Environment & Infrastructure UK _combustionpdf/_en_1.0_&a=d CEPI (2011) unfold the future The Forest Fibre Industry 2050 Roadmap to a low-carbon bio-economy. Available at Entec (2010) Assessment of the Possible Development of an EU-wide NOx and SO2 Trading Scheme for IPPC Installations. Final Report for the European Commission. June Entec UK Limited JRC (2012) Best Available Techniques (BAT) Reference Document for Production of Pulp, Paper and Board. Draft May Nykomb (2005). High Efficient Motor Fuel Production from Biomass via Black Liquor Gasification. Presentation at ISAF XV International Symposium on Alcohol Fuels September 2005, San Diego, USA by Tomas Ekbom of Nykomb Synergetics. Available at REC (2005) Short presentation of Somes Pulp&Paper Mill. Regional Environmental Center for Central and Eastern Europe. Available at s/short-presentation-somes-dej.doc July

82 Tran, H. and Vakkilainnen, E (2008) The Kraft Chemical Recovery Process. July

83 4. Category (c) combustion plants within refineries firing the distillation and conversion residues from the refining of crude-oil for own consumption, alone or with other fuels 4.1 Introduction This category of combustion plant covers plants greater than 50MW th within refineries that combust one or more liquid residues (or by-products) that result from certain crude oil refining processes. The IED includes emission limit values (ELVs) for these combustion plants (Section 4.2.2), which were maintained from the LCP Directive. There is an obligation for the Commission to review the need to amend or complement these ELVs. This chapter is aiming to provide an overview of these combustion plants and their emission levels in order to inform that review. During the course of this study ICF has sought to gain access to the relevant data gathered by the EIPPCB for the Refinery BREF review to underpin an analysis of combustion plants in refineries. ICF received data from the EIPPCB in January 2013, but the data have not been considered sufficiently large or robust in order to underpin an analysis of combustion plants in refineries to the same level of detail as adopted for the other sectors covered by this study. BAT conclusions for combustion plants in refineries are being considered specifically in the revision of the REF BREF, which is due to be finalised imminently (at the time of writing March 2013). 4.2 Description and background Sector background The energy system of refineries supplies heat (directly or indirectly) and electricity to the refining processes. Direct heating is from the combustion of fuels in heaters or furnaces; indirect heating is through steam production in boilers. Combined steam and electricity generation (CHP) is becoming increasingly common in refineries. Power supply can be outsourced to a separate company (whether produced on site or imported locally), or in the case of refineries generating their own power with surplus, electricity can be exported to the grid and steam can be sold to local demand. Combustion plants in refineries are not exclusively over 50MW th (stack) in capacity; many smaller sources are also used References in the IED The relevant references in the IED to combustion plants within refineries firing the distillation and conversion residues from the refining of crude-oil for own consumption, alone or with other fuels, are set out in the following subsections: DETERMINATIVE FUEL Article 3(30) of the IED defines a determinative fuel thus: determinative fuel means the fuel which, amongst all fuels used in a multi-fuel firing combustion plant using the distillation and conversion residues from the refining of crude-oil for own consumption, alone or with other fuels, has the highest emission limit value as set out in Part 1 of Annex V, or, in the case of several fuels having the July

84 same emission limit value, the fuel having the highest thermal input amongst those fuels; Therefore the determinative fuel is declared at a plant level for each pollutant of SO 2 35, NO X and dust, based on which fuels have the highest ELV, or if equal ELV which fuel contributes more to the thermal input. The determinative fuel concept is used in the determination of ELVs for multi-fuelled combustion plants, as described in the next subsection. Based on analysis of the relevant ELVs for refinery combustion plants, the determinative fuel at each capacity class are shown in the table below, assuming that both liquid fuels and gaseous fuels are used in the plants. The table shows that liquid fuels, i.e. distillation or conversion refinery residues, are the determinative fuel in most cases (this will not be the case for plants not firing any liquid fuels). Table 4.1 Determinative fuels per pollutant and per capacity class Pollutant Plants permitted before 27 November 2002 Plants permitted after 27 November 2002 (Note 1) MW th MW th MW th >500 MW th MW th MW th SO 2 (Note 2) RFG (Note 3) RFG (Note 3) RFG (Note 3) RFG (Note 3) Coke (where applicable) or liquid fuels Liquid fuels or coke (where applicable) NO X Liquid fuels Liquid fuels Liquid fuels RFG or coke (where applicable) Liquid fuels Liquid fuels or RFG Dust Liquid fuels Liquid fuels Liquid fuels Liquid fuels Liquid fuels or coke (where applicable) Liquid fuels or coke (where applicable) Note 1: there are no refinery combustion plants greater than 300 MW th in the EU27 that were permitted after 27 November 2002 according to the 2009 LCP inventory. Note 2: the determinative fuel concept is not applicable for SO 2 where the bubble concept is used (Article 40(3)). Note 3: assuming that RFG is counted as a low calorific gas from gasification of refinery residues. MULTI-FUEL FIRING COMBUSTION PLANTS Article 40 of the IED sets out the process for determining the ELVs for multi fuel firing combustion plants. Article 40(1) sets out the default method, which is based on weighting the ELVs according to the fuels consumed. Article 40(2) provides an alternative method for derivation of ELVs for multi-fuel fired existing combustion plants firing the distillation and conversion residues from refining crude oil. The method relies on the determinative fuel concept as set out above. In the method, if the determinative fuel makes up 50% or more of the thermal input then the ELV for the determinative fuel applies. As indicated above, for the plants of relevance to this chapter which are firing liquid residues, the liquid residues are in most cases the determinative fuel, so the ELV would be for liquid fuels. If the determinative fuel makes up less than 50% of the thermal input then the ELV is calculated through a weighting method for each of the fuels, except that the ELV for the determinative fuel is calculated as twice the normal applicable ELV minus the ELV for the fuel co-fired with the lowest ELV. For a typical refinery combustion plant less than 500MW th firing refinery fuel oil and refinery fuel gas, this method effectively provides for a 35 Not applicable for SO 2 where the bubble concept is used. July

85 determinative fuel ELV (before fuel weighting) of 600 mg/nm 3 for NO X and 95 mg/nm 3 of dust. Ultimately, due to the complexity of the calculation, which leads to different applicable ELVs plant by plant, it is difficult to model the overall EU picture of emission levels compared to ELVs. As indicated later in this chapter in Section 4.4.1, the vast majority of refinery combustion plants that fire liquid fuels are also firing other fuels, and so are classed as multi-fuel fired combustion plants. AVERAGE SO 2 EMISSION LIMIT VALUES FOR MULTI-FUEL FIRING COMBUSTION PLANTS WITHIN A REFINERY Part 7 of Annex V of the IED sets out average SO 2 emission limit values for multi-fuel firing combustion plants (excluding gas turbines and engines) within a refinery with a rated thermal input of 50 MW or more, which use the distillation and conversion residues from the refining of crude-oil for own consumption, alone or with other fuels. This is what is referred to as the bubble concept for SO 2 ELVs, in which the bubble ELV can apply to these multi-fuelled plants, irrespective of the mix of fuels used. The ELVs concern existing plants permitted prior to 2013: the limit values are set at 1000 mg/nm 3 for plants permitted before 27 November 2002 and 600 mg/nm 3 for newer plants (based on 3% oxygen for liquid fuels). The averaging periods for these bubble ELVs are no different to those applicable for the standard ELVs,. These ELVs are referenced from Article 40(3), which is the article concerning the calculation of ELVs for multi-fuel fired combustion plants; in that Article it is stated that these ELVs may be applied (i.e. it s not obligatory) instead of those described in Article 40(1) or Article 40(2). Discussions with stakeholders have indicated that the ELVs of Article 40(3) would be chosen in preference to the ELVs that otherwise apply under Article 40(2). NOX EMISSION LIMIT VALUES FOR COMBUSTION PLANTS USING LIQUID FUELS Point 4 of Part 1 of Annex V sets out the NO X emission limit values for combustion plants using solid or liquid fuels for combustion plants referred to in Article 30(2) [existing plants]. In this point a derogation from the standard ELVs for liquid fuel-fired plants of rated thermal input between 100 MW and 500 MW which were permitted before 27 November 2002 sets the NOx ELV to be 450 mg/nm 3. Without this derogation the applicable limit values would be 200mg/Nm 3 for plants of rated thermal input 100 to 300 MW, or 150 mg/nm 3 for plants of rated thermal input greater than 300 MW. DUST EMISSION LIMIT VALUES FOR COMBUSTION PLANTS USING LIQUID FUELS Point 7 of Part 1 of Annex V sets out the dust emission limit values for combustion plants using solid or liquid fuels for combustion plants referred to in Article 30(2) [existing plants]. In this point a derogation from the standard ELVs for liquid fuel-fired plants of rated thermal input above 50 MW which were permitted before 27 November 2002 sets the dust ELV to be 50 mg/nm 3. Without this derogation the applicable limit values would be 200mg/Nm 3 for plants of rated thermal input 100 to 300 MW, or 150 mg/nm 3 for plants of rated thermal input greater than 300 MW Defining the fuels in scope of distillation and conversion residues from the refining of crude oil An important aspect of the refining energy system is the effective use of internal residue streams as part of the energy mix, i.e. by-products of refining processes. The energy July

86 systems of refineries (including combustion plants of 50 MW th or more) use the following fuels for on-site power and/or steam generation (in order of most used in combustion plants 50MW th ): Refinery fuel gas (RFG) Refinery fuel oil Natural gas Coke generated in catalytic cracking Coke generated in catalytic cracking is not a residue from distillation or conversion, so is not considered within the scope of this chapter. This study is considering only refinery fuel oil as an applicable distillation and conversion residue. This is because the only references regarding ELVs and their derogations are made for liquid fuels in the IED; refinery fuel gas is covered in the next chapter on gases other than natural gas. REFINERY FUEL OIL JRC (2012) describes refinery fuel oil as being normally a mixture of the residues from atmospheric and/or vacuum distillation and conversion and cracking processes, where cracking processes are noted (page 345) as including thermal cracking, catalytic cracking and hydrocracking of residues. The sulphur content of RFO, which varies greatly, is determined by feedstock choice, except for residues from hydrocracking. The table below shows the sulphur, nitrogen and metal content of different fractions suitable to be used as RFO. Table 4.2 Sulphur, nitrogen and metal content of fractions suitable for liquid refinery fuels (source: JRC, 2012) Fraction suitable to be used as liquid refinery fuel Crude oil origin S (%) N (%) Metal content (%) Atmospheric residue North Sea Atmospheric residue Middle East Vacuum residue North Sea Vacuum residue Middle East Cracked residue Middle East CONCAWE (2010) in their sulphur survey for 2006 suggests that there has been a trend for decreasing use of refinery fuel oil between 1998 and 2006 shown in Figure 4.1 below. CONCAWE (2010) implies that the trend of decreasing use of liquid residues (corresponding with increasing use of refinery fuel gas) has been more prevalent in smaller combustion plants (stacks less than 50 MW th ) than in combustion plants of 50MW th or more. This is not however consistent with another driver which has been the determinative fuel concept in the LCPD and now IED, which is cited as providing an incentive to limit refinery fuel oil use to 50% of the thermal input of plants 50MW th. July

87 Figure 4.1 Refinery fuel oil firing as a fraction of overall refinery fuel combustion (source: CONCAWE, 2010) Commercially available liquid fuels are not used at all in combustion plants in refineries. 36 Therefore, for LCPs in refineries as reported in the LCP emission inventories, the fuel category of liquid fuels can be interpreted to be solely refinery fuel oil. 37 The 2009 LCP emission inventory indicates that refinery combustion plants of rated thermal input at least 50MW th used 28% refinery fuel oil in 2009, as shown in Figure 4.2 below. This figure is higher than that suggested by CONCAWE (2010), which indicates that, when only considering plants >50MW th, the contribution of refinery fuel oil to the total energy input has dropped from 31% in 1998 to 22% in Figure 4.2 Refinery LCP fuels in 2009 (data source: 2009 LCP inventory) 36 Personal communication with CONCAWE, 20 th September Petrochemical complexes may rely on liquid residues from refineries, but may potentially supplement with commercially available fuels. Petrochemical combustion plants that are not categorised as refineries in the LCP inventory are not captured in this study. July

88 REFINERY FUEL GAS (RFG) Contract /ENV/2012/627812/C3 Collection and Analysis of Data for the Review required RFG is the main fuel combusted in refinery combustion plants. RFG is collected within the refinery gas system a site-wide manifold that collects off-gas streams from multiple processes around the refinery site. As a gaseous fuel, RFG is not considered as a relevant distillation and conversion residue from the refining of crude oil 38, and is discussed further in Chapter 5. RFG, which itself represents a range of different off-gases collected in the RFG manifold (e.g. syngas, coke gas, hydrogen), covers all gaseous fuel other than natural gas used in refineries. Therefore, for LCPs in refineries reported in the LCP emission inventories, the fuel category of Other gases can be interpreted to be solely RFG. This assumption is relevant to the methodology of this chapter (see Section 4.3) Petrochemical complexes The definition of this category of combustion plants stipulates that the residues are for own consumption. On complex integrated petrochemical sites, residues from refining processes and/or from chemical processes may be utilised individually or jointly in refining or chemical combustion plants or processes. The boundaries between the refinery and chemical installations are on such sites difficult to establish. Consequently, for the purposes of scope, for petrochemical sites that combust distillation and conversion residues from the refining of crude oil 39, but which do so in chemical plants, it is assumed that such uses are not falling under IED Article 30(9). This is further underlined by the text in Article 30(9)(b) that specifies that the review covers combustion plants within refineries. Furthermore, due to the principal data source being relied upon of the LCP emissions inventory the classification of refinery plants in this inventory will be adopted. Refineries generally follow a simple philosophy for the use of fuel to their furnaces and boilers. Firstly they maximise use of RFG, as they typically cannot export it if they did not burn the RFG, they would have to flare it. Then they burn the low grade refinery fuel oil, which is kept distinct from sales fuel oil Specificity of the energy systems The review clause of Article 30(9) establishes that the Commission should take into account the specificity of the energy systems of refineries during its review of the ELVs set out in Annex V for the relevant combustion plants. It is understood that this term is making reference to the individuality and variability of the fuels and feedstocks within refineries. This variability is of relevance to this study; however as noted below in Section 4.3.1, insufficient data has been available through this process and supplied to ICF to enable a robust analysis on this basis. For example, any fixed emission limit values should be set at high enough levels to allow for consequent variations in the emission levels that result (only) from variations in the liquid fuel used (which in turn is affected by feedstocks of the refinery) and the ratio of gas to liquid 38 CONCAWE have suggested that they consider RFG to be a distillation and conversion residue from the refining of crude oil. 39 Considered likely to be solely RFG, but this needs to be further checked. July

89 fuels used. 40 In order to establish the specificity of the energy systems and the implications for emission levels, it will be important to utilise sufficient data that demonstrates this relationship between fuel/feedstock variability and emission level variability, and isolate liquid refinery fuel variation from refinery fuel gas variation. The term specificity of the energy systems may also refer to the specific design and operating characteristics of the combustion equipment, which can have a significant impact on thermal NOx generation from combustion of gaseous fuels Size / capacity This review is limited to combustion plants meeting the capacity thresholds for inclusion within the scope of Chapter III of the IED, i.e. with minimum total rated thermal input at the stack level of 50MW or more. In determining the rated thermal input of a combustion plant, the aggregation and de minima rules of Article 29 are applied which mean that only combustion units exhausting to the stack of 15MW th or more are counted, but that this does not necessarily exclude units less than 15MW th from being within scope of the IED Combustion techniques Article 28(a) of the IED excludes from the scope of Chapter III plants in which the products of combustion are used for the direct heating, drying or any other treatment of objects or materials. Furnaces, in which no direct contact between waste gases and material takes place, are included within the scope of the study, along with boilers, engines and turbines. 4.3 Methodology for assessment of plants in the EU Data sources ICF reviewed the following sources for potential use in describing and assessing combustion plants within refineries firing the distillation and conversion residues from the refining of crude-oil for own consumption: LCP emission inventory CONCAWE (2010) sulphur survey Selection of EIPPCB data submitted for the revision of the Refineries BREF LCP EMISSION INVENTORY The LCP inventory is considered to be a comprehensive list of combustion plants of 50 MW th or more in the EU27. The inventory includes all such refinery combustion plants identified separately, with their rated thermal input (MW), liquid fuels consumption (TJ, among other fuels also stipulated) and annual emissions in 2009 (tonnes of SO 2, NO X and dust). The emissions data represent the total emissions from all the fuels fired at the combustion plant, i.e. not just due to single fuels. This means that, for multi-fuel plants, there is not a separate emission level associated with liquid fuel firing only (or any other fuel). 40 Emission limit values are set separately for liquid and gaseous fuels, except where the bubble concept is used. 41 Personal communication with DG JRC 15 th January July

90 For those refinery combustion plants that are single fuelled in this dataset, it is possible to estimate average emission levels associated with separate refinery fuels. This is a useful step as it forms a set of assumptions on which to split the emissions associated with combustion of multiple fuels into each fuels. For the purposes of this analysis, single fuelled is assumed to mean greater than 90% of the thermal input to the combustion plant. On this basis, the table below summarises the emission factors (g/gj) of single fuelled refinery combustion plants from the 2009 LCP inventory. This clearly does not make any assumptions around the abatement equipment installed at a refinery to reduce emissions, and it also is only intended to represent possible but indicative EU average values. In particular for RFG-fired refinery combustion plants, the data are based on a large number of plants (112) and so could perhaps be interpreted as an EU average that is of sufficient statistical significance that takes account of the variations in RFG composition (of which sulphur and hydrogen content principally affect the pollutants of interest). Table 4.3 Single fuel type Single fuelled refinery LCPs: number, energy input and emission factors (data source: 2009 LCP inventory) Number of plants Energy input (TJ) Emission factor (g/gj) Number % of total refinery LCPs TJ % of total refinery LCPs SO2 NOx Dust Other solid fuels / coke Liquid fuels / refinery fuel oil 7 2.5% 95, % % 49, % Natural gas 8 2.8% 28, % Other gases / RFG % 317, % CONCAWE (2010): 2006 SULPHUR SURVEY The CONCAWE sulphur survey reports on the results of a detailed survey (via questionnaire) of around two thirds of refining capacity in OECD Europe regarding sulphur pathways from crudes to products taking account of sulphur recovery and SO 2 emissions. The 2006 sulphur survey was the most recent published result. Of relevance to this chapter, CONCAWE (2010) includes data from year 2006 detailing the use and characteristics of refinery fuel oil (as well as other refinery fuels) such as sulphur content, split of fuel types and other technical data. CONCAWE (2010) does not cover pollutants other than SO 2. In particular, CONCAWE (2010) estimates the average SO 2 concentration for refineries large combustion plants (i.e. all combustion plant 50MW th per refinery, excluding other emission sources such as catalytic crackers, flares, sulphur recovery units and small combustion plants). These are shown below in Figure 4.3; the weighted average figure for 2006 is close to 600 mg/nm 3, with values ranging from zero to over 3,000 mg/nm 3. July

91 Figure 4.3 Estimated distribution of annual average refinery LCP SO 2 concentrations in 1998, 2002 and 2006 (source: CONCAWE, 2010) EIPPCB DATASET FROM THE REVISION OF THE REFINERIES BREF This dataset is based on responses to a questionnaire sent out by the EIPPCB and received from 21 refinery operators distributed across 5 Member States. The responses represent 22 sites, of which four are part of petrochemical complexes. The following questions of relevance to this study were captured in the questionnaire (for some of the responses): Energy production capacity (MW th ), split by furnaces, boilers, engines, turbines, CHP. It was not indicated as to how the capacity was split into single fuel firing or multi-fuel firing Quantities of fuels consumed for energy production, split by fuel type (gaseous, heavy liquid, light liquid, and solid) and whether internally generated or imported, with sulphur and nitrogen content. Flue gas volumes per energy production type (furnaces, boilers etc.). This is however not split by fuel type Annual absolute emissions (tonnes) of SO 2, NO X and PM from each energy production source. This is not identifiable by fuel type. SO 2 and NO X equivalent concentrations. Techniques affecting emissions, including energy management techniques, heat integration and recovery, primary measures on furnaces and boilers, SO 2, NO X and particulate abatement techniques. As not all the questionnaires have included responses to the questions of relevance to this study, this reduces the statistical significance of the dataset. Consequently, ICF has agreed with the Commission not to use this dataset as the basis for the analysis. However, some of the responses have been used to provide a secondary or backup data source. July

92 4.3.2 Methodology Contract /ENV/2012/627812/C3 Collection and Analysis of Data for the Review required The aim of this chapter is to establish estimates for the prevalence of those combustion plants in refineries that are firing distillation and conversion residues from crude oil refining, including how numerous they are, their capacities, fuel consumption and emissions. An assessment of abatement equipment of these plants is also made but this is much less robust than the assessments made in earlier chapters. Insufficient statistically significant data was obtained regarding the emission levels associated with different abatement techniques such that broader assumptions were necessary together with interpolation from the LCP inventory. The number, capacity and fuel consumption of combustion plants in refineries firing refinery fuel oil are taken directly from the LCP inventory for 2009, as this is deemed the most comprehensive data source available for the EU27. The LCP inventory for 2009 was filtered for refinery plants only, and also filtered to exclude refinery combustion plants that did not fire liquid fuels in It is important to note that this assessment has been made using the LCP inventory for 2009 only; changes may occur over time for a refinery combustion plant, for example using no refinery fuel oil in one year but may use the fuel in other years. Additionally, the number and capacity of relevant plants in Croatia 42 have been determined through consultation with the Croatian authorities. Fuel consumption for these Croatian plants has been estimated based on the average fuel mix of EU27 refinery combustion plants (source: LCP inventory) and the annual operating hours for the plants (supplied by Croatian authorities). EMISSIONS Total emission concentrations for each refinery combustion plant were estimated by dividing the total annual emissions by the sum of the flue gas volumes estimated for each fuel type. The flue gas volumes for each fuel type were estimated using specific flue gas volumes and calorific value assumptions taken from CONCAWE (2010), and reproduced below in Table 4.4. Table 4.4 Fuel characteristics assumptions made (source: CONCAWE, 2010) Fuel Specific flue gas volume (Nm 3 /kg) Gross calorific value (MJ/kg) Oxygen content (%) Specific flue gas volume (Nm 3 /GJ) Coke (Note 1) Refinery fuel oil Natural gas RFG Note 1: Adjusted to 6% oxygen from 3% oxygen as quoted in CONCAWE (2010) For the estimation of emissions and emission levels from refinery fuel oil only, it was necessary to apportion the total emissions for each LCP across the fuels that were combusted, similarly to that adopted in AMEC (2012). This was undertaken by assuming 42 Ttwo refineries operating in Croatia at Sisak and Rijeka, each with two combustion plants over 50MW th July

93 emission factors for each fuel type, as given in Table 4.3, together with implied fuel specific flue gas volumes (Nm 3 /GJ) derived from flue gas volumes (Nm 3 /kg) and calorific values (MJ/kg) taken from CONCAWE (2010) and reproduced in Table 4.4. The emission levels estimated from the LCP inventory were then compared to other data sources, including the relative sulphur contents of different site-produced fuels from EIPPCB data and from CONCAWE (2010). Emissions from the four plants in Croatia were estimated based on the emission factors set out in Table 4.3 in light of insufficient data from the Croatian authorities on emissions at the refineries due to combustion processes only. 4.4 Overview of combustion plants within refineries firing the distillation and conversion residues from the refining of crude-oil for own consumption, alone or with other fuels The estimates in this section are for the EU27 plus Croatia Number of combustion plants Out of the 289 combustion plants 50 MW th in refineries in the EU27 plus Croatia, 168 plants (i.e. 58% of the total number) use refinery fuel oil. The 168 plants are distributed across all the Member States that reported refinery LCPs, but there are some refineries that reported no LCPs using liquid refinery fuel. In these latter refineries, there may be utilisation of refinery fuel oil in combustion plants smaller than 50 MW th. The vast majority of the combustion plants fire refinery fuel oil with other fuels rather than alone. This is shown in the table below. Table 4.5 Number of refinery combustion plants split by capacity class MW th MW th MW th >500 MW th Total EU27 Alone With other fuels Croatia Alone With other fuels Total The 168 combustion plants represent 5% of the total of 3,283 LCPs reported in the 2009 LCP inventories for EU27 (i.e. excluding other combustion plants in Croatia). Three of the refinery LCPs firing refinery fuel oil are opted out plants (2 wholly opted out, 1 partially opted out) under Article 4(4) of the LCPD so this capacity would be expected to cease operation by July

94 4.4.2 Capacity of combustion plants The rated thermal input of the 168 combustion plants that use refinery fuel oil is shown in the table below split by capacity class. Of these, 704 MW th are opted out under Article 4(4) of the LCPD and so would be expected to be no longer operating from Table 4.6 Rated thermal input (MW th ) of refinery combustion plants split by capacity class MW th MW th MW th >500 MW th Total EU27 3,454 11,671 11,323 24,273 50,721 Croatia Total 3,760 11,671 11,323 24,273 51,027 This total capacity of 51 GW th represents 3.8% of the total LCP capacity of 1,354 GW th reported in the 2009 LCP inventories for EU27 (i.e. excluding other combustion plants in Croatia) Fuel consumption The total fuel consumption of all refinery LCPs, as well the fuel consumption of just those LCPs firing distillation and conversion residues are shown in the table below. Table 4.7 Fuel consumption of refinery combustion plants Other solid fuels / coke Liquid fuels / refiner fuel oil Natural gas Other gases / RFG Total TJ % TJ % TJ % TJ % TJ All refinery LCPs EU27 41,199 3% 365,117 28% 87,513 7% 808,205 62% 1,302,034 Croatia 216 3% 1,917 28% 460 7% 4,244 62% 6,838 Total 41,415 3% 367,034 28% 87,973 7% 812,449 62% 1,308,872 Refinery LCPs that fire refinery fuel oil EU27 17,135 2% 365,117 37% 49,507 5% 544,907 56% 976,666 Croatia 216 3% 1,917 28% 460 7% 4,244 62% 6,838 Total 17,352 2% 367,034 37% 49,966 5% 549,152 56% 983,504 The total fuel consumption of 984 PJ of the combustion plants firing refinery fuel oil represents 5.9% of the of the total LCP fuel consumption of 16,566 PJ reported in the 2009 LCP inventories for EU27 (i.e. excluding other combustion plants in Croatia). The 367 PJ of refinery fuel oil represents 2.2% of the EU27 total LCP fuel consumption. The proportion that refinery fuel oil makes up of the total energy input varies for each combustion plant. Figure 4.4 shows the distribution across the combustion plants of the proportion that refinery fuel oil makes up of total energy input; i.e. the distribution behind the overall figure of 37% refinery fuel oil. July

95 Figure 4.4 Refinery fuel oil consumption as a proportion of total energy input (data source: 2009 LCP inventory) Current emissions The estimates of emissions made in this study from only the combustion of refinery fuel oil in refinery combustion plants are presented below in Table 4.8. These are distinct from the total emissions from refinery combustion plants that fire at least some refinery fuel oil, as these emissions would include those arising from the combustion of other fuels (RFG, natural gas and coke where relevant). The table presents totals for EU27 and Croatia combined due to high uncertainty in Croatia emission estimates. Table 4.8 Estimated SO 2, NO X and dust emission levels (mg/nm 3 ) and emissions (kt per annum) from the combustion of refinery fuel oil in refinery combustion plants in the EU27 and Croatia, compared to reported annual total LCP emissions, NECD total and total industrial combustion emissions Pollutant Combustion of refinery fuel oil in refinery combustion plants (estimated in this study) Comparison with LCP inventory emissions Comparison with NECD emissions Total Industrial combustion Average emission concentration (mg/nm 3, 3% O 2, annual average, standard conditions) Total annual emission (kt/yr) 2009 LCP (kt/yr) % of LCP 2009 NECD (kt/yr) % of NECD 2009 NECD (kt/yr) % of NECD EU27 and Croatia SO , % 4, % 3, % NO X , % 9, % 2, % Dust % N/A N/A N/A N/A The average emission concentrations shown in Table 4.8 above are weighted average emission concentrations across all relevant plants (i.e. plants which fire at least some refinery fuel oil). The full ranges of emission concentrations estimated at a plant level for these plants are shown in Figure 4.5, Figure 4.6 and Figure 4.7 below separately for SO 2, July

96 NO X and dust respectively. The three Figures plot both the total LCP emission concentrations (i.e. resulting from the combustion of all fuels at these plants) as well as the estimated effective emission concentration arising solely from combustion of refinery fuel oil. It is important to note that, since almost all the plants that fire refinery fuel oil do so with other fuels, the emission concentration associated with a single fuel is a theoretical construct as in practice the flue gases are mixed and inseparable. The construct is however valid for the purposes of discussions of emission limit values which are set separately per fuel type. Figure 4.5 Estimated distribution (across cumulative total energy content) of SO 2 emission concentration from refinery plants firing refinery fuel oil in 2009 (source: this study, based on 2009 LCP inventory) July

97 Figure 4.6 Estimated distribution (across cumulative total energy content) of NO X emission concentration from refinery plants firing refinery fuel oil in 2009 (source: this study, based on 2009 LCP inventory) Figure 4.7 Estimated distribution (across cumulative total energy content) of dust emission concentration from refinery plants firing refinery fuel oil in 2009 (source: this study, based on 2009 LCP inventory) July

98 The plot for SO 2 (Figure 4.5) shows that the weighted average SO 2 emission concentration for the LCPs overall is 662mg/Nm 3, which compares to the figure for just RFO combustion of 1558 mg/nm 3 and the SO 2 bubble from LCPD and IED for plants permitted before 2003 of 1000 mg/nm 3 and 600 mg/nm 3 for more newly permitted plants (IED, Annex V, Part 7). It is important to note that the estimates assume no end of pipe abatement which is consistent with the small number of questionnaire responses to EIPPCB. Figure 4.5 also indicates the range of implied sulphur content of refinery fuel oil based on the estimates of the SO 2 emission concentrations for this fuel is from close to 0% to 4%, with a weighted average of 1.0% S. This is slightly lower than the estimates in two other sources, although these are noted to be different years of data, and important to note that from 2008 gas oil was required to have lower sulphur content, which may impact on the sulphur content of RFO. The two other sources of data that have been identified concerning the sulphur content of refinery fuel oil are: CONCAWE (2010). From its sulphur survey for the year 2006, CONAWE (2010) indicates that the (weighted) average sulphur content of refinery fuel oil was 1.39%. The EIPPCB dataset included 12 refineries reporting the sulphur content of siteproduced heavy liquid fuels consumed for energy production 43, with sulphur contents ranging from 0.37% to 2.1%, with an average weighted by consumption of 1.32%. 1.32% S is considered to be equivalent to an SO 2 emission concentration of around 2,100 mg/nm Figure 4.6 shows that estimated NO X emission levels from the combustion of RFO vary widely between 100 mg/nm 3 up to around 800 mg/nm 3, with a few exceptional cases (<2% of energy consumed) of emission levels up to 1400 mg/nm 3. Around 80% of the energy consumed (representing 75% of the plants) are estimated to have NO X emission levels between 270 mg/nm 3 and 620 mg/nm 3. Figure 4.7 shows estimated dust emission levels from the combustion of RFO to vary widely between close to 0 mg/nm 3 up to around 150 mg/nm 3, with a few exceptional cases (<2% of energy consumed) of emission levels up to 265 mg/nm 3. A major component of the dust emissions is heavy metals. Refinery fuel oil has a heavy metal content four to five times higher compared to the metals content of the crude, depending on the yield of residue and the residue content of the crude (JRC, 2012). The metal content can vary significantly depending on the crude, generating (unabated) particulate concentration in the flue gas of 150 to 500 mg/nm 3 (JRC, 2012). This suggests that dust abatement is likely to be installed at a large proportion of refinery combustion plants that are firing RFO. 43 Excluding imported heavy liquid fuels, which are assumed to fall outside the scope of distillation and conversion residues from the refining of crude-oil for own consumption. 44 At 1.32% S, 1 tonne RFO contains 13.2kg S, which upon full oxidation to SO2 is equivalent to an unabated emission of 26.4kg SO2. At 3% oxygen, the specific flue gas volume assumed for RFO combusted in boilers is 300 Nm3/GJ (source: derived from CONCAWE, 2010). The flue gas volume from combusting 1 tonne RFO (assuming a net calorific value of GJ/tonne) is therefore 300x42.03 = Nm 3. The emission concentration from combusting 1 tonne of RFO that has 1.32 %S is therefore 26,400,000 mg / Nm 3 = 2094 mg/nm 3 (without end of pipe SO 2 abatement). This is equivalent to indicating that around 1600mg/Nm 3 is equivalent to 1% S, i.e. slightly lower than other estimates of 1700mg/Nm 3. July

99 Table 4.9 Metal content of residual fuel oil typically used in refineries (source: JRC, 2012) Metal Concentration range (ppm) Average concentration (ppm) Possible specific load 5 th to 95 th percentile (g/t feed) V Ni Pb Cu < Co Cd < Cr < Mo As < Se Abatement options In light of the current finalisation of the Refineries BREF document at the time of writing, this chapter does not refer to what is BAT for the reduction of emissions from refinery fuel oil. The finalised Refineries BREF should be referred to once published. 4.5 Future trends PRODUCTS Key trends that affect the refined petroleum products industry in Europe: Reduction in demand for heavy fuel oil in land based sources Competing impacts on transport fuel demand: increase in travel demand (a small but growing portion of which in electric vehicles) in line with economic growth against reduction in demand through CO 2 standards in cars and vans and their consequent reduction in fuel consumption (and the associated rebound effect); impact of the Renewable Energy Directive requiring 10% renewable in transport fuels by 2020 thus reducing the share of fossil fuels from refineries; impact of the Fuel Quality Directive including its Article 7a requirements by 2020 on crude slate of refineries; and impact of low sulphur marine fuel standards requiring shifts in product types to the marine sector. Specifically, an increased demand from 2015 for 0.1% sulphur fuel for use by vessels in the sulphur emission control areas (SECAs) of the North Sea, English Channel and Baltic Sea. In addition, a change demand for sulphur content of residual fuel from 3.5% to 0.5% is due to occur in either 2020 or All of the above impacts on demand will lead to changes in the refining industry to enable the refineries to match their supply (in quantity and product slate). This may include changes in the quality of refinery fuel oil that, as a residue, will be available in refineries for their own energy systems. July

100 CRUDE IEA (2012) forecasts the following changes to the crude slate for Europe in this decade: OECD Europe to decrease FSU crude imports by 1.2 mb/d in 2017 compared to OECD Europe to increase crude imports from Africa by 900 kb/d in 2011 to take 3.0 mb/d by North Sea production is projected to continue to decline, leaving a sweet crude (i.e. low sulphur) deficit that needs to be filled. Production of African crudes is projected to grow, as identified above, providing a source of crude to replace lost North Sea production. Demand for West and North African crudes in the US is expected to decline due to (1) increase in U.S. domestic shale/tight oil production and (2) flat U.S. product demand. ECONOMICS AND CLOSURES Overall, in North West Europe there is little requirement for refiners to add new crude refining capacity due to the current surplus position and the large and growing gasoline surplus. In fact, refiners have already closed some capacity permanently whilst other capacity has been idled temporarily to prevent the buildup of excess product stocks. The economic environment for refiners in NW Europe is forecast to be harsh and further closures are likely. Understandably there are few investments planned to expand refining capacity or upgrading. IEA (2012) forecast projects European refinery throughput to drop more than demand by 2017 compared to 2011; i.e. which would lead to a drop in refinery utilisation. However, increases in capacity are expected in Central and Eastern Europe as the economy recovers and slight increases can be expected in the Mediterranean. The application of a carbon price creates an incentive for refineries to run crudes that require less energy-intensive processing (i.e. medium to light crudes and low sulphur in place of heavier and higher sulphur). 4.6 References AMEC (2012) Collection and analysis of data to support the Commission in reporting in line with Article 73(2)(a) of Directive 2010/75/EU on industrial emissions on the need to control emissions from the combustion of fuels in installations with a total rated thermal input below 50MW. Final Report for the European Commission. Available at: _combustionpdf/_en_1.0_&a=d CONCAWE (2010) Sulphur dioxide emissions from oil refineries in Europe (2006). Available from EC (2012) LCP emission inventories for years 2007 to Personal communication. Entec (2010) Assessment of the Possible Development of an EU-wide NOx and SO2 Trading Scheme for IPPC Installations. Final Report for European Commission. European Commission (2010) SEC(2010) 1398 final. Commission Staff Working Paper on Refining and the supply of Petroleum Products in the EU. IEA (2012) Medium-Term Oil Market Report, 12 October July

101 JRC (2012) Draft Best Available Techniques (BAT) Reference Document for the Refining of mineral oil and gas. Draft 2 (March 2012) July

102 July

103 5. Category (d) Combustion plants firing gases other than natural gas 5.1 Introduction Annex V of the IED currently includes emission limit values for combustion plants firing gases other than natural gas. Article 30(9) requires the Commission to review the need to amend those ELVs. The purpose of this chapter is focussed on looking at the relevant combustion plants in the EU that fire gases other than natural gas in this context. 5.2 Description and background References in the IED Annex V of the IED makes reference to several gases other than natural gas, as listed in the table below with the ELVs that are set out. Table 5.1 References in the IED to gases other than natural gas Gas Applicability ELV (mg/nm 3 ) SO 2 NO X Dust Low calorific gases from the coke oven / coke oven gas Plants referred to in Article 30(2), Excluding gas turbines/engines or 300 (Note 1) Not stipulated Low calorific gases from blast furnace / blast furnace gas Plants referred to in Article 30(2), excluding gas turbines/engines or 300 (Note 1) 10 Gases produced by the steel industry that can be used elsewhere Excluding gas turbines/engines Not stipulated Not stipulated 30 Low calorific gases from gasification of refinery residues Plants referred to in Article 30(2), granted a permit before 27 November 2002 or began operation before 27 November 2003, excluding gas turbines/engines or 300 (Note 1) Not stipulated Liquefied gas Excluding gas turbines/engines 5 Not stipulated Not stipulated Other gases Plants referred to in Article 30(2), excluding gas turbines/engines Not stipulated 200 or 300 (Note 1) Not stipulated Other gases Gas turbines including CCGTs referred to in Article 30(2) Not stipulated 120 or 200 (Note 2) Not stipulated Note 1: 300mg/Nm 3 for plants less than 500MW th that were granted a permit before 27 November 2002 or began operation before 27 November 2003 Note 2: 200mg/Nm 3 for plants operating fewer than 1500h/year that were granted a permit before 27 November 2002 or began operation before 27 November Refineries The combustion of refinery fuel gas in large combustion plants is covered in the REF BREF. REFINERY FUEL GAS RFG is a fuel of varying composition since is a combination of off-gases from a number of different refinery processes. At a site level, the composition of RFG varies considerably over time according to the crude being processed, which processes are operating, the functionality of any clean-up treatment of off-gases prior to injection into the RFG manifold (e.g. amine scrubbers), etc. (Entec, 2009). JRC (2012) indicates that: July

104 Some refinery fuel gases may be sulphur-free at source (i.e. from catalytic reforming and isomerisation processes) or sulphur-containing at source (most other processes, i.e. from crude distillation, cracking and all hydrodesulphurising processes). In the latter case the gas streams are normally treated by amine scrubbing to remove H 2 S before being released to the refinery fuel gas system. RFG is not defined in the IED. The following entries in the above table are considered to apply to RFG: low calorific gases from gasification of refinery residues and other gases. The gasification of refinery residues leads to the generation of syngas (mixture of mainly CO and H 2 ), from which hydrogen is sometimes separated. JRC (2012) also suggests both (in section 2.10) that the syngas is normally directed straight for combustion in combined cycle power generation units, and also (in section ) that the syngas is fed into the RFG manifold. If the syngas, which may have a low calorific value, is fed into the RFG manifold then it is unclear if the combined RFG remains a low calorific gas or not. Low calorific is not defined however in the IED. The calorific value of RFG is typically lower than methane, depending on the mixture, around 18.6 MJ/Nm 3, while the calorific value of syngas might around half this value. In all cases the calorific values will be very site- and operationspecific. Catalytic crackers produce coke as well as off-gases. Coke is outside the scope of this study (as a solid fuel), but the incineration of the off-gases from catalytic crackers if fed into the RFG system is within the scope of this chapter. The incineration of these carbon monoxide rich off-gases takes place in a CO boiler. The composition of the RFG affects the emissions resulting at the combustion stage. Entec (2009) notes that NO X emission levels are affected by the hydrogen (H 2 ) fraction, which is typically between 5% and 10%, but can rise much higher, even up to 60%. The H 2 content affects the flame temperature (higher H 2 results in higher flame temperatures), which in turn affects thermal NO X formation (higher combustion temperatures increase NO x emissions). Other constituents of RFG include unsaturated hydrocarbons and hydrocarbons C 3 +, the latter of which also affects NO X emissions. REFINERY FUEL GAS SYSTEM AND RFG USE The RFG manifold, referred to as the fuel gas system, may include injection of natural gas and/or LPG as supplementary to the RFG. The quantities of gases other than RFG injected into the manifold will vary according to the balancing needs of the refinery and will depend on whether the heating value of the RFG is insufficient or if there is no separate natural gas distribution system. Since LPG is a valuable product, it is unlikely to be added into the RFG manifold in significant quantities. In practice therefore, the fuel piped into the combustion plants in the refinery from the RFG manifold may include mixtures of RFG, natural gas and/or LPG. However, it is considered from ICF expertise that the quantities of natural gas and LPG are minimal. Furthermore, RFG may also be exported to chemical plants in petrochemical sites. RFG can t normally be sold and hence the fuel is typically for own consumption. It is unclear however if there are any (albeit few) instances in which RFG is sold, e.g. in non-jointly owned petrochemical works. Plants with excess RFG production ("gas containment") have typically used the excess to generate power. In some cases, this is done via co-generation of power and steam sometimes in a facility operated by a different company. Therefore there is a potential overlap with category (e) too when considering double counting. July

105 RELEVANT COMBUSTION TECHNIQUES RFG is fired in boilers, process furnaces (heaters) and gas turbines. The majority of installed combustion plants in refineries are heaters, which can include waste heat boilers for additional heat extraction. Gas turbines are not generally fired with RFG; however RFG is used in waste heat recovery boilers attached to the turbine for additional heat extraction (Entec, 2009) Steel industry OVERVIEW Processes in integrated steelworks may generate three types of off gases: coke oven gas (COG), blast furnace gas (BFG) and basic oxygen furnace gas (BOFG), which constitute the basis of the energy system in an integrated steelworks. The three off-gases are described in the next subsection. THE OFF-GASES (SUMMARISED FROM ENTEC, 2009) COG is the result of dry distillation (coking) of specific types of coal into lumps of coal. Important properties for this natural product in this respect are C-content, volatile matter and sulphur content. Coke is an essential ingredient of the iron reduction process in a Blast Furnace. COG typically has a calorific value from 17.4 to 20.0 MJ/m 3, typically comprises hydrogen, methane, carbon monoxide and hydrocarbons. The raw gas yield is of the order of m 3 /hr/tonne of coking coal. BFG is the result of the reduction process of iron oxide with carbon. Hot blast air (optionally enriched with oxygen) reacts with the reducing agents to produce mainly carbon monoxide as a combustible compound, that further reduces iron oxide to liquid iron. The residual gas which still contains carbon monoxide alongside hydrogen, carbon dioxide and other components is BFG. The calorific value of BFG is consequently low (typically in the range 2.7 to 4.0 MJ/m 3, depending on the CO content) but is produced in significant volumes (approximately 1200 to 2000 Nm 3 of BFG per tonne of pig iron) making it in the largest contributor to total energy of all the off-gases. Due to its low calorific value, waste flue gas volumes are high for BFG combustion. BOFG is the result of the steelmaking process in which liquid iron and scrap are relieved of the carbon in the iron by blowing pure oxygen into the liquid hot metal. The BOF process is a batch process and therefore the production of BOF-gas is similarly in batches. BOFG is best used as a general addition to enrich large volume gas flows like BFG. BOFG is being used in LCPs in an increasing number of integrated sites. BOFG comprises approximately 70-80% CO and typically has a calorific value around 9 MJ/m 3. BOFG produced during oxygen blowing leaves the furnace with a temperature of 1200 C and flow rate of Nm 3 /t steel. Site-specific blending of COG, BFG and BOFG is undertaken in order to produce a gas with a more regulated calorific value of around 10 MJ/m 3. USE OF OFF-GASES The off-gases, which have different calorific values, are used as fuels (rather than flared): The gases are prioritised for use in processes, e.g. in the coke oven, blast furnace stoves, sinter plant, BOF plant, re-heating or annealing furnaces, continuous casting, re-heating furnaces in hot rolling mills. Excess gases are used primarily as unconventional gaseous fuel within on-site power plants that provide services to the steelworks. July

106 At some sites there may be agreements for further excess gases to be exported offsite to adjacent power plants that fire multiple fuels (operated by the power sector). Power plants play an important role in an integrated steelworks as they consume the excess process gases and provide the necessary steam and power to all the key processes (JRC, 2012). Most integrated steelworks also supplement the firing of the three mentioned gases with purchased commercial fuels (e.g. oil, natural gas), as well as utilising waste heat from the power plant and importing electricity. If the site s process operations do not require a process gas the on-site combustion plants LCPs have to respond to the situation and utilise this gas to avoid flaring. This response has to be rapid (timescales in terms of minutes) and the LCP has to accommodate changes not only in volumetric terms but also in terms of the gas calorific value. Consequently the load factors of the LCPs vary during operation. As examples: When a blast furnace shuts down, which may be planned or unplanned, volumes of BFG available to the combustion plant suddenly drops significantly. Typically large amounts of COG are used directly in hot rolling mill heating furnaces. When such a furnace is taken off, the combustion plants have to deal with large amounts of COG, which has high calorific value. The use of pulverised coal injection in the blast furnace in place of coke affects BFG: when the injection fails, the calorific value and volume of BFG drops. There are variable volumes of BOFG due to the batch BOF process. Situations such as the above cause temporary disturbances to combustion plants that fire or co-fire the gases. Specifically, if the combination of gases combusted suddenly increases significantly in calorific value, e.g. if the proportion of COG increases significantly, then the combustion temperatures in the combustion plant will rise temporarily leading to greater thermal NO X. Good energy management at the steelworks therefore includes buffering to some extent where possible fluctuations in gas quality and quantity. Flaring is avoided where possible (i.e. except safety flaring) to avoid wasting energy. However, BFG, COG and BOFG cannot be stored in gasholders for longer than a few minutes, because of the large volumes produced. Each steelworks may have different energy management structures. The main variables affecting the efficiency and emissions of combustion plants in integrated steelworks firing process gases are temporal fluctuations in volumes of each gas and the different calorific values of each gas. The two factors combine together to give varying total energy (MJ) available to combustion plants from process gases. COMBUSTION TECHNIQUES JRC (2012) indicates that the use of process gases from iron and steel works in a power plant in an integrated steelworks is usually realised in boilers or gas turbines (in combined cycle gas turbines). Most steelworks power plants are CHP plants optimised for electricity generation, since steelworks heat demand is low. Local private or commercial heat consumers are therefore sought for supply of surplus heat. The process gases are not fired directly in gas turbines, but in CCGT plants the process gases are used to fire the heat recovery boiler for supplementary firing. July

107 ABATEMENT MEASURES The Iron and Steel BREF (JRC, 2012) covers abatement measures for the pre-treatment of off-gases. Regarding SO 2 /NO X /dust emission abatement measures for combustion plants firing steel industry gases, although the Iron and Steel BREF includes new sections for process gasfired power plants in iron and steel works, the BAT and BAT-AELs for these plants are to be included in the LCP BREF which is undergoing revision at the time of writing Chemical industry Chapter 6 of this report concerns liquid production residues from chemical installations. It is also described there that chemical production processes generate a number of gaseous effluent streams that may be recovered as fuel for furnaces, power plants or steam boilers. Burning such residues in combustion plants in the chemical industry allows for recovery of the internal energy of these fuels while avoiding the flaring or the thermal oxidation of such non-commercial fuels. The utilisation of residues can displace the purchase and consumption of commercial fossil fuels (and hence reduce costs) and reduce waste streams, leading to overall site energy optimisation. Any relevant gaseous fuels other than natural gas that these chemical combustion plants are firing are of relevance to the present chapter. Entec (2009) indicates that the use of such gases in the chemical sector is rare but cited specific examples of two Belgian combustion plants that were firing ethylene or propylene based chemical production off-gases Other sectors firing gases other than natural gas Possible other gases used in combustion plants include biogases and landfill gases. However, the use of such gases in combustion plants greater than 50MW th is considered to be limited or non-existent. Landfill sites often utilise gas engines to generate electricity from the landfill gas. Consultation with EUROMOT has shown that landfill sites are likely too small to generate the gas rate necessary for sole firing of a combustion plant of 50 MW th or more. As such landfill gas is not considered to form a significant source of other gases for combustion plants 50MW th or more. Anaerobic digesters also produce biogas but are typically less than 15 MW th in size. The German biogas sector is one of the largest in Europe. ICF estimates that Germany s approximate 7,500 biogas plants have an average rated thermal capacity of around 1MW. 45 As such these plants and hence biogas is not considered to form a significant source of other gases for combustion plants 50MW th or more Abatement measures for combustion plants firing gases other than natural gas The following BREFs / draft BREFs should be consulted: 2006 adopted LCP BREF (section 7.5) or June 2013 Draft 1 of LCP BREF (section ) 45 Based on figures in IEA Bioenergy Member Country Report for Germany July

108 July 2013 Final Draft of REF BREF (section 5.9) 2003 adopted LVOC BREF (section 6.4) 5.3 Existing legal provisions Industrial Emissions Directive References in the IED to other gases were set out in Section Member States Entec (2009, p. 60) included an overview of the ELVs set out in different Member States for the combustion of other gases; it is not reproduced here, except to highlight that the Dutch legislation BEES-A includes adjustment factors for ELVs for the combustion of RFG to try to take account of the variability of the RFG composition. This is outlined in Entec (2009) from page Data collection exercise and methodology Overview of methodology The data collection and subsequent analysis of the use of other gases in this chapter has been split out by sector. This both reflects the different data sources that are distinct for each sector, as well as for the purposes of reporting data consistently with other chapters. As identified in section 5.2, the principal sectors associated with the combustion of gases other than natural gas are refineries and steelworks, and to a lesser extent the chemical industry Data sources and methodology refineries The methodology for the estimation of number, capacity, fuel consumption and emissions from refinery combustion plants firing RFG is as per chapter 4. The fuel category of other gases in the LCP inventory is assumed to solely refer to RFG Data sources steelworks gases Three sources of data have been identified to inform this work on the combustion plants firing steel industry process gases: The LCP inventory for 2009 includes a number of combustion plants that are marked as being in the steel industry. EUROFER have provided ICF with an inventory of combustion plants of 50MW th or more firing steel industry process gases. The EIPPCB have provided ICF with a draft dataset of combustion plants firing steel industry gases that has been gathered under the LCP BREF review. LCP INVENTORY This is described in section Additionally, through consultation with Eurelectric, additional consolidation of this inventory has occurred regarding those plants which are in the power sector but which are using steel industry process gases, i.e. are expected to be located adjacent to steelworks. July

109 EUROFER Contract /ENV/2012/627812/C3 Collection and Analysis of Data for the Review required EUROFER have provided ICF 46 with an anonymised list of LCPs of 50MWth or more that use iron and steel process gases. The list was collated by EUROFER to form an input into the LCP BREF revision that is in process now. The list covers at plant level the rated thermal input according to the common stack definition of combustion plant, the fuels used, as a % of thermal input, and the annual emission loads of SO 2, NOx and dust for year The fuels that are separately identified are natural gas, COG, BFG, BOFG (in some cases only a combined fraction for BFG+BOFG), light crude, heavy oil, tar and coal. The location of each plant is included. This list contains only those combustion plants that participated in the data collection exercise. The list only included those LCPs where the process gases (COG,BFG,BOFG) are considered to be main fuels. The definition used for main fuel was as defined by the EIPPCB in the framework of the LCP BREF revision and expressed in the questionnaire as: Fuel or a combination of fuels used during normal operating conditions, i.e. fuel diet for the combustion plant. During the periods of start-ups and shutdowns, as well as during operation under special permit conditions, other fuels may be used (e.g. 'startup'/'back-up fuel(s)). Therefore the EUROFER list does not include multi-fueled LCPs using I&S process gases as an addition to a fossil fuel. It was estimated by EUROFER that the list contains over 80% of the identified LCP s using I&S process gases as main fuels in the framework of the LCP BREF revision. EIPPCB The EIPPCB provided ICF with a dataset of 41 combustion plants firing steel industry process gases and 14 combustion plants firing unspecified other gases that has been gathered under the LCP BREF review. 47 The dataset is considered to be draft due to in some cases unvalidated data. The emission levels of plants firing steel industry gases from this dataset are shown in Section Results Overview of all LCPs firing gases other than natural gas The 2009 LCP inventory has been analysed for the plants that fire other gases. Summary statistics from the inventory for separate proportions (any, at least 50% and at least 90%) that other gases makes up of total thermal input at an LCP level are shown below in the table, both in absolute figures and expressed as proportions of the total reported EU27 figures. The relative proportions of total EU LCPs shown in Table 5.2 are plotted in Figure 5.8. The results indicate that nearly one in five LCPs (19%) in the EU27 fires at least some proportion of gases other than natural gas, and that for these LCPs, their total rated thermal input and SO 2, NO X and dust emissions make up smaller proportions of the EU27 totals (16%, 14%, 16% and 13%, respectively), suggesting that these plants are smaller than the EU average and with slightly lower emissions factors than the EU average LCP (a 46 Personal communication, 24 January Personal communication 12 December July

110 comparison against only natural gas fired plants may not lead to the same conclusions). For the plants that are firing other gases as their main fuel (>90% of thermal input), the results show that these plants make up 6% by number and 3% by capacity of all EU LCPs, and less than 1% of SO 2 /dust emissions and 2% of EU LCP NO X emissions. Table 5.2 The number of LCPs, their rated thermal input and annual SO 2, NO X and dust emissions for the EU27 for which other gases make constitutes different proportions of total energy input (source: 2009 LCP inventory) Other gas energy input (% of total energy input) Number of LCPs Capacity of LCPs SO 2 emissions NOx emissions Dust emissions No. % (EU) GW th % (EU) kt % (EU) kt % (EU) kt % (EU) >0% % % % % 16 13% >50% % % % % % >90% % % % % % Figure 5.8 The proportions of EU27 LCP totals that LCPs firing varying percentages of others gases make up for number, capacity and emissions (data source: 2009 LCP inventory) Overview of combustion plants in refineries firing RFG NUMBER OF COMBUSTION PLANTS Out of the 289 combustion plants 50 MW th in refineries in the EU27 plus Croatia, 270 plants (i.e. 93% of the total) use at least some quantities of RFG. 48 One third of these plants are fired solely with RFG, with the remainder being primarily co-fired with refinery fuel oil, with some being co-fired with natural gas or solid fuels (which is assumed to be coke). 48 The 19 refinery combustion plants reported to use no RFG in 2009 are distributed across 10 Member States. These 19 plants include 1 plant with main fuel of other solid fuels, 8 plants with main fuel as natural gas, and 10 plants with main fuel of refinery fuel oil. July

111 Table 5.3 Number of refinery combustion plants split by capacity class MW th MW th MW th >500 MW th Total EU27 Alone With other fuels Croatia Alone With other fuels Total The 270 combustion plants represent 8% of the total of 3,283 LCPs reported in the 2009 LCP inventories for EU27 (i.e. excluding other combustion plants in Croatia). Three of the refinery LCPs firing RFG are opted out plants (2 wholly opted out, 1 partially opted out) under Article 4(4) of the LCPD so this capacity would be expected to cease operation by RATED THERMAL INPUT The rated thermal input of the 270 combustion plants that use RFG is shown in the table below split by capacity class. Of these, 704 MW th are opted out under Article 4(4) of the LCPD and so would be expected to be no longer operating from Table 5.4 Rated thermal input (GW th ) of refinery combustion plants split by capacity class MW th MW th MW th >500 MW th Total EU Croatia Total This total capacity of 65 GW th represents 4.8% of the total LCP capacity of 1,354 GW th reported in the 2009 LCP inventories for EU27 (i.e. excluding other combustion plants in Croatia). FUEL CONSUMPTION The total fuel consumption of all refinery LCPs, as well as the fuel consumption of just those LCPs firing RFG are shown in the table below. The part of the table showing fuel consumption of all refinery LCPs is a repeat of Table 4.7in chapter 4. July

112 Table 5.5 Fuel consumption of refinery combustion plants 50MW th or more Other solid fuels / coke Liquid fuels / refinery fuel oil Natural gas Other gases / RFG Total TJ % TJ % TJ % TJ % TJ All refinery LCPs EU27 41,199 3% 365,117 28% 87,513 7% 808,205 62% 1,302,034 Croatia 216 3% 1,917 28% 460 7% 4,244 62% 6,838 Total 41,415 3% 367,034 28% 87,973 7% 812,449 62% 1,308,872 Refinery LCPs that fire RFG EU27 38,335 3% 346,073 27% 69,415 6% 808,205 64% 1,262,027 Croatia 216 3% 1,917 28% 460 7% 4,244 62% 6,838 Total 38,551 3% 347,990 27% 69,874 6% 812,449 64% 1,268,865 The total fuel consumption of 1269 PJ of the combustion plants firing RFG represents 7.7% of the total LCP fuel consumption of 16,566 PJ reported in the 2009 LCP inventories for EU27 (i.e. excluding other combustion plants in Croatia). The 812 PJ of RFG represents 4.9% of the EU27 total fuel consumption by LCPs. The proportion that RFG makes up of the total energy input varies for each combustion plant. Figure 4.4 shows the distribution across the combustion plants of the proportion that RFG makes up of total energy input; i.e. the distribution behind the overall figure of 64% RFG. This 64% compares well with the EIPPCB dataset (combined gaseous portion 66%), the LCP inventory (64% and 6% for RFG and natural gas respectively) and CONCAWE, 2010 (combined gaseous portion 70%). Figure 5.9 RFG consumption as a proportion of total energy input in refinery combustion plants 50MW th or more (data source: 2009 LCP inventory) July

113 EMISSIONS The estimates of emissions made in this study from only the combustion of RFG in refinery combustion plants are presented below in Table 4.8. These are distinct from the total emissions from refinery combustion plants that fire at least some RFG 49, as these emissions would include those arising from the combustion of other fuels (refinery fuel oil, natural gas and coke where relevant). The table presents totals for EU27 and Croatia combined due to high uncertainty in Croatia emission estimates. The table indicates that, compared to total emissions in the LCP inventory, the NO X emissions from RFG combustion in combustion plants in refineries is most significant at 2.9%. Table 5.6 Estimated SO 2, NO X and dust emission levels (mg/nm 3 ) and emissions (kt per annum) from the combustion of RFG in refinery combustion plants in the EU27 and Croatia, compared to reported annual total LCP emissions, NECD total and total industrial combustion emissions Pollutant Combustion of RFG in refinery combustion plants (estimated in this study) Comparison with LCP inventory emissions Comparison with NECD emissions Total Industrial combustion Average emission concentration (mg/nm 3, 3% O 2, annual average, standard conditions) Total annual emission (kt/yr) 2009 LCP (kt/yr) % of LCP 2009 NECD (kt/yr) % of NECD 2009 NECD (kt/yr) % of NECD EU27 and Croatia SO , % 4, % 3, % NO X , % 9, % 2, % Dust % N/A N/A N/A N/A The average emission concentrations shown in Table 4.8 above are weighted average emission concentrations across all relevant plants (i.e. plants which fire at least some RFG). The full ranges of emission concentrations estimated at a plant level for these plants are shown in Figure 4.5, Figure 4.6 and Figure 4.7 below separately for SO 2, NO X and dust respectively. The three Figures plot both the total LCP emission concentrations (i.e. resulting from the combustion of all fuels at these plants) as well as the estimated effective emission concentration arising solely from combustion of RFG. It is important to note that, since many of the plants that fire RFG do so co-fired with other fuels, the estimated emission concentration associated with a single fuel (see methodology section 5.4.1) is a theoretical construct as in practice the flue gases are mixed and inseparable. The construct is however valid for the purposes of discussions of emission limit values which are set separately per fuel type. 49 E.g. as shown in Section for all LCPs firing other gases. July

114 Figure 5.10 Estimated distribution (across cumulative total energy content) of SO 2 emission concentration from refinery plants firing RFG in 2009 (source: this study, based on 2009 LCP inventory) Figure 5.11 Estimated distribution (across cumulative total energy content) of NO X emission concentration from refinery plants firing RFG in 2009 (source: this study, based on 2009 LCP inventory) July

115 Figure 5.12 Estimated distribution (across cumulative total energy content) of dust emission concentration from refinery plants firing RFG in 2009 (source: this study, based on 2009 LCP inventory) The plot for SO 2 (Figure 4.5) shows that the weighted average SO 2 emission concentration for the LCPs overall is 662mg/Nm 3, which compares to the figure for just RFG combustion of 119 mg/nm 3. The estimates show that 90% of plants in terms of energy input may have emissions concentrations in the range of 0 to 200 mg/nm 3, with the remaining 10% having concentrations from 200 up to 4,000 mg/nm 3 (although only the last 0.4% of this energy input is with associated with emission concentrations in the range of 1,100 to 4000mg/Nm 3 ). Figure 4.5 also indicates the range of implied sulphur content of RFG (if it is assumed that no end of pipe SO 2 abatement equipment is fitted), estimated to range from 0.0% to 2.3%, with a weighted average of 0.07% sulphur 50, which appears to indicate that desulphurisation of S-containing gases is widely implemented prior to release of said gases to the RFG system. This percentage is consistent with two other sources: CONCAWE (2010). From its sulphur survey for the year 2006, CONAWE (2010) indicates that the average sulphur content of gaseous fuels (including both RFG and natural gas) used in combustion plants 50MW th or more was 0.05%. the dataset from the EIPPCB from the revision of the REF BREF containing data which are mostly for the year 2008 and suggest that the weighted average of the small number of sites that reported both quantities of RFG and their average S content was 721ppm (0.07%), with sulphur contents ranging from 0ppm to 1500ppm. 50 I.e. 0.07% S in RFG is considered to be equivalent to the SO 2 emission concentration of 119 mg/nm 3 when assuming a specific flue gas volume for RFG combusted in boilers of 309 Nm 3 /GJ (source: derived from CONCAWE, 2010) and a calorific value of GJ/tonne (source: CONCAWE, 2010). This assumes no end of pipe SO 2 removal. July

116 Figure 4.6 shows that estimated NO X emission levels from the combustion of RFO varies widely between 15 mg/nm 3 up to around 700 mg/nm 3, with one exceptional case (<0.1% of energy consumed) of 1650 mg/nm 3. Figure 4.7 shows estimated dust emission levels from the combustion of RFO to vary between 0 mg/nm 3 and 70 mg/nm 3, with 90% (by energy content; 92% by number) estimated to have emission levels less than 15 mg/nm Overview of combustion plants firing steel industry gases NUMBER OF COMBUSTION PLANTS 50 MW TH OR MORE The list of plants from EUROFER contains 45 combustion plants of 50MW th or more, and EUROFER indicated that this captures at least 80% of the plants in their opinion. his excludes e.g. external power generation co-firing with steel industry gases. The estimate of the number of combustion plants firing steel industry gases derived from the LCP inventory is 94 LCPs, which includes those plants which co-fire with steel industry gases. RATED THERMAL INPUT The list of plants from EUROFER contains 12.1 GWth of combustion plants firing steel industry gases in steelworks. EUROFER indicated that this would represent at least 80% of the plants. The estimate derived from the LCP inventory which is around 37.1 GWth. FUEL CONSUMPTION The analysis with the LCP inventory suggests that, as totals from the LCPs estimated to fire or co-fire steel industry gases, their total fuel consumption is 521 PJ, of which around half (277 PJ) is from gases other than natural gas. The EUROFER list gives insight into the possible fuel mix between the different steel industry gases. The data suggest that the mix is 67% BFG, 7% BOFG and 26% COG. EMISSIONS The I&S BREF (JRC, 2012) provides achieved emission levels from more than 20 European integrated steelworks power plants (boilers and turbines) fired with process gases. Table 5.7 Achieved emission values for gas-fired boilers and turbines when using process gases from iron and steel works (source: JRC, 2012) Annual average emission concentration (mg/nm³) SO 2 NO X Dust CO Mean Max Min The EIPPCB dataset (LCP BREF) includes annual average SO 2, NO X, dust and CO emission levels for 39 boilers firing steel industry process gases. The averages across the plants, weighted by the total rated thermal input of the combustion plants, are: 135 mg SO 2 /Nm 3 and 26 mg CO /Nm 3, i.e. considerably higher than the achieved mean concentrations of SO 2 and CO in Table 5.7, and 84 mg NO X /Nm 3 and 4.1 mg dust/nm 3, i.e. slightly lower and significantly lower than the achieved mean concentrations of NO X and dust in Table 5.7. July

117 Figure 5.13 below shows scatter plots of the EIPPCB dataset for the four pollutants of SOx, NO X, dust and CO of emission concentrations against rated thermal input. For annual average SOx emissions levels, Figure 5.13 shows that plants varied from around 25 mg/nm 3 to close to 400 mg/nm 3, with no specific trend for plants with higher capacity. This range is wider than the range of concentrations achieved shown in Table 5.7. For annual average NO X emission levels, Figure 5.13 shows that plants varied from around 20 mg/nm 3 to close to 250 mg/nm 3, with no specific trend for plants with higher capacity. This range is wider than the range of concentrations achieved shown in Table 5.7. For annual average dust emission levels, Figure 5.13 shows that plants varied from 0 mg/nm 3 to 65 mg/nm 3, which is also a wider range than shown in Table 5.7. The distribution of plant shows that only the smallest capacity plants (with rated thermal inputs around 100 MW th or less) have the higher dust emission concentrations. The same distribution of CO emission levels across capacity is demonstrated in Figure Figure 5.13 Annual average emission concentrations of combustion plants firing steel industry process off-gases (data source: EIPPCB dataset) July

118 The emissions estimates from EUROFER are that, of the combustion plants in steelworks firing steel industry process gases, the annual emission loads are 9.9kt SO 2, 8.1 kt NO X and 0.5 kt dust (data primarily for year 2010). As these EUROFER estimates capture between 80% to 100% of all the plants in steelworks, these estimates increase to max kt SO 2, 10.2 kt NO X, and 0.6 kt dust. These appear to be low compared to the estimates from the LCP inventory which are for year 2009 as 42kt SO 2, 37kt NO X and 2.8 kt dust Overview of combustion plants firing other gases in the chemical industry Approximately 100 combustion plants of 50MW th or more have been identified as being in the chemical or petrochemical industry and firing other gases (none of these plants are listed as being refinery LCPs in the LCP inventory). These plants have a total rated thermal input of 20 GW th in the 2009 LCP emission inventory. Around one third of these combustion plants (estimated: 28) in number are co-firing only small proportions (i.e. less than 20% of thermal input) of gases other than natural gas, and around half of them (estimated: 52) are firing large proportions of other gases (greater than 80% of thermal input). In terms of energy input, Figure 5.14 plots the distribution of the proportion that other gases makes up of total energy input for the combustion plants that have been identified as being in the chemical or petrochemical industry. The plants are ordered in terms of their proportion of other gases consumption, and this is plotted against cumulative energy input to the plants. The plot shows that about 25% of the energy input is in plants firing other gases only. A further 15% of energy input is in plants firing primarily other gases (i.e. over 50% of energy input is from other gases). Most of the remaining 60% of energy input is from plants where only small proportions of thermal input are from other gases. Figure 5.14 Proportion of other gases of total energy input in chemical industry combustion plants 50MW th or more (data source: 2009 LCP inventory) For those plants for which the other gases makes up 50% or more of the thermal input, the emissions attributable to the combustion of other gases are likely to strongly influence the total emissions of those LCPs as they are likely to be equal to or slightly lower than the total plant emissions. July

119 Table 5.8 Estimated SO 2, NO X and dust emissions (kt per annum) from combustion plants in the chemical industry firing other gases, compared to reported annual total LCP emissions, NECD total and total industrial combustion emissions Pollutant Total LCP emissions (kt/yr) Comparison with LCP inventory emissions 2009 LCP (kt/yr) Comparison with NECD emissions Total % of LCP 2009 NECD (kt/yr) % of NECD Industrial combustion 2009 NECD (kt/yr) % of NECD Plants firing % other gases by energy input SO , % 4, % 3, % NO X 6.6 1, % 9, % 2, % Dust % N/A N/A N/A N/A Plants firing 0%-50% other gases by energy input SO , % 4, % 3, % NO X 12 1, % 9, % 2, % Dust % N/A N/A N/A N/A The annual emission loads presented in the table above indicate that emissions from the combustion of gases other than natural gas in the chemical and petrochemical sector are small. No abatement scenarios for these plants are considered. Estimates of emission concentration levels of these plants are not undertaken due to the uncertainty of the fuel types Other sectors with combustion plants firing other gases than natural gas For those combustion plants not accounted for in the various analyses above, i.e. which have not been attributed to the refining sector, to using steel industry process gases or known to be in the chemical or petrochemical industry, the Table 5.9 summarises data from the LCP inventory. July

120 Table 5.9 Estimated SO 2, NO X and dust emissions (kt per annum) from combustion plants firing other gases, not included in above analyses, compared to reported annual total LCP emissions, NECD total and total industrial combustion emissions Pollutant Number of LCPs Total LCP emissions (kt/yr) Comparison with LCP inventory emissions 2009 LCP (kt/yr) % of LCP 2009 NECD (kt/yr) Comparison with NECD emissions Total % of NECD Industrial combustion 2009 NECD (kt/yr) % of NECD Plants estimated to be firing biogas SO , % 4, % 3, % NO X , % 9, % 2, % Dust % N/A N/A N/A N/A Plants estimated to be firing non-condensable gases in pulp / paper installations SO , % 4, % 3, % NO X , % 9, % 2, % Dust % N/A N/A N/A N/A Plants estimated to be firing low calorific local methane deposits SO 2 0 2,563 0% 4,798 0% 3,871 0% NO X , % 9, % 2, % Dust % N/A N/A N/A N/A Plants estimated to be firing syngas at IGCC plants SO , % 4, % 3, % NO X , % 9, % 2, % Dust % N/A N/A N/A N/A Plants estimated to be firing sour gases from oil/gas field extraction SO , % 4, % 3, % NO X , % 9, % 2, % Dust % N/A N/A N/A N/A Other Plants firing unknown gases SO , % 4, % 3, % NO X , % 9, % 2, % Dust % N/A N/A N/A N/A Table 5.9 shows that a large portion of the plants firing other gases remain uncategorised in this analysis, i.e. it is unknown what gases these plants are firing. The plants could be firing any of the gases mentioned in this chapter. The emission estimates in the subsections of Table 5.9, e.g. on biogas, are minima, because there may be additional biogas fired plants within the unknown section. Overall the emissions from these plants are not a negligible fraction of the total LCP emissions. No abatement scenarios are described for these plants due to unknown uptake of abatement measures. July

121 5.6 References Contract /ENV/2012/627812/C3 Collection and Analysis of Data for the Review required CONCAWE (2010) Sulphur dioxide emissions from oil refineries in Europe (2006). Available from Entec (2009) Study to inform ongoing discussions on the proposal for a Directive on industrial emissions. Part 1: Combustion Activities Final Report. June 2009 EUROFER (undated) Presentation. Personal communication 4 th February JRC (2012) Best Available Techniques (BAT) Reference Document for Iron and Steel Production. European Commission Joint Research Centre. March July

122 July

123 6. Category (e) Combustion plants in chemical installations using liquid production residues as noncommercial fuel for own consumption 6.1 Introduction This chapter deals with combustion plants in chemical installations that combust residues from chemical processes for the generation of steam and/or power which are not commercially available fuels. Under Article 30(9) of the IED, the need to amend the emission limit values for these combustion plants needs to be reviewed by the Commission. The purpose of this chapter is to technically describe the plants within this category and to present available information on these plants in terms of their number, capacity, fuels, type of plant, emissions, current abatement techniques, potential additional abatement techniques (especially BAT), location, uses etc. to assist the Commission in determining the need to amend Union-wide emission limit values. 6.2 Description and background Various chemical production processes generate a number of gaseous and liquid effluent streams that may be recovered as fuel for furnaces, power plants or steam boilers. Burning such residues in combustion plants in the chemical industry allows for recovery of the internal energy of these fuels while avoiding the flaring or the thermal oxidation of such noncommercial fuels. The utilisation of residues can displace the purchase and consumption of commercial fossil fuels (and hence reduce costs) and reduce waste streams, leading to overall site energy optimisation. It is known from the LVOC BREF (2003) (e.g. section regarding effluent from ethylene oxide recovery section) that liquid effluent streams that are concentrated in hydrocarbons can either be sold or incinerated as a waste (in which case it would be subject to the provisions of Chapter IV of the IED not chapter III). If this exact stream is partially sold, or if it is known to be equivalently commercially available with exact same properties from another site, then the residue is deemed to be outside the scope of this study because the scope of the combustion plants under review in Article 30(9)(d) are those that are using residues as non-commercial fuel for own consumption. Some chemical installations are located in or are a part of larger petrochemical complexes. In such integrated sites, it is may be difficult to separate those processes related to crude oil refining from chemical plants in which a furnace/boiler provides steam to both refinery and chemical plant units. In this case, the fuel use to the boiler can easily be apportioned; in many cases, units in a refinery are operated as profit centers and detailed info should be available to operators on each unit (certainly between refining and chemical production) although such information may not be mandated for provision to regulators. For such integrated sites this chapter is concerned with the consumption of the residues within the same installation as where the residues were produced because the scope of the combustion plants under review in Article 30(9)(d) are those that are using residues as noncommercial fuel for own consumption. There are no definitions associated with liquid production residues in the IED (nor in the adopted 2003 LVOC BREF), as the residues are very much site-specific effluent streams, as well as varying according to the chemical process. The effluent streams may vary temporally in their composition as well as according to the process and/or chemical installation in which they are recovered. However, typical components include sulphur (S)-components, fixed carbon and nitrogen (N)-components. Further information on fuel characteristics are provided in Section 6.6. July

124 6.3 Existing Emission Limit Values European Union No specific exemptions to the general ELVs in the IED for these installations are provided for SO 2 or dust emissions. A specific reference to these combustion plants and their NO X emissions is made in Annex V Part 1 Point 4 of the IED which states that: Combustion plants in chemical installations using liquid production residues as noncommercial fuel for own consumption with a total rated thermal input not exceeding 500 MW which were granted a permit before 27 November 2002 or the operators of which had submitted a complete application for a permit before that date, provided that the plant was put into operation no later than 27 November 2003, shall be subject to an emission limit value for NO x of 450 mg/nm 3. Combustion plants in chemical installations using liquid production residues as noncommercial fuel for own consumption with a total rated thermal input exceeding 500 MW will be subject to the NO X ELV for general liquid fuels, which is 150 mg/nm 3 for existing plants. Furthermore, those relevant plants that were permitted after the dates set out in the exemption (and which are still subject to the ELV requirements of Article 30(2)) will have the NO X ELVs for general liquid fuels applied for all capacity classes (i.e. 450 mg/nm 3 for MW th, 200 mg/nm 3 for MW th and 150 mg/nm 3 for MW th ). For new combustion plants, there are no such derogations Other jurisdictions emission limit values A review of regulations outside the EU have been undertaken to assess conditions imposed on combustion plants using liquid production residues as non-commercial fuel. This review covered Japan, Korea, Australia, Norway, Switzerland, Australia, Canada and the USA and identified relevant regulations only in Switzerland and the USA. Relevant summaries and extracts are provided below. SWITZERLAND ORDINANCE ON AIR POLLUTION CONTROL OF 16 DECEMBER 1985 (STATUS AS OF 15 JULY 2010) 51 In Annex 3, Chapter 7 emission limit values are set for Combustion installations for liquid fuels as specified in Annex 5 Number In Annex 3, Chapter 8, emission limit values are established for Multi- and mixed-fuel combustion installations. Provision 81 Multi-fuel combustion installations states that if a single installation operates alternately on different kinds of fuel, the emission limitation requirements are determined by the fuel used in each case. Provision 82 relating to Mixed-fuel combustion installations states that If different kinds of fuel are burned at the same time in a single installation, the emission concentrations must not exceed the composite limit value. The composite limit value is calculated used a weighted ELV approach Other liquid fuels means liquid organic compounds which can be combusted like extra light fuel oil and meet the requirements specified in Number 132. See: July

125 UNITED STATES Country Name of legislation Link to legislation Timescale of legislation United States Title 40 (Protection of Environment), Part 60 (Standards of Performance for New Stationary Sources) Subpart D Standards of Performance for Fossil-Fuel-Fired Steam Generators Title 40 (Protection of Environment), Part 60 (Standards of Performance for New Stationary Sources) ecfr Subpart D Standards of Performance for Fossil-Fuel-Fired Steam Generators ecfr Current Applicabilit y Emission limit values Clause Subpart D Applicability and designation of affected facility. Subpart D gives standards for particulate matter ( 60.42), for SO 2 ( 60.43) and for NOX ( 60.44). Details (1) Each fossil-fuel-fired steam generating unit of more than 73 megawatts (MW) heat input rate (250 million British thermal units per hour (MMBtu/hr)). (2) Each fossil-fuel and wood-residue-fired steam generating unit capable of firing fossil fuel at a heat input rate of more than 73 MW (250 MMBtu/hr) Standard for particulate matter (PM). 43 nanograms per joule (ng/j) heat input (0.10 lb/mmbtu) derived from fossil fuel or fossil fuel and wood residue. Exemptions: - Natural gas fired plants - Plants that combust only gaseous or liquid fossil fuel (excluding residual oil) with potential SO 2 emissions rates of 26 ng/j (0.060 lb/mmbtu) or less and which do not use post-combustion technology to reduce emissions of SO 2 or PM Standard for sulfur dioxide (SO 2). (1) 340 ng/j heat input (0.80 lb/mmbtu) derived from liquid fossil fuel or liquid fossil fuel and wood residue Standard for nitrogen oxides (NO X expressed as NO 2). (2) 129 ng/j heat input (0.30 lb/mmbtu) derived from liquid fossil fuel, liquid fossil fuel and wood residue, or gaseous fossil fuel and wood residue. Derogation s Emissions and fuel monitoring Continuous emissions monitoring system (CEMS) for measuring SO 2 or NO X are not required in the following cases: - fossil-fuel-fired steam generators that combust only gaseous or liquid fossil fuel (excluding residual oil) with potential SO 2 emissions rates of 26 ng/j (0.060 lb/mmbtu) or less and that does not use post-combustion technology to reduce emissions of SO 2 or PM, and which monitor SO 2 emissions by fuel sampling and analysis or fuel receipts. - fossil-fuel-fired steam generators that do not use flue gas desulfurization and which monitor SO 2 emissions by fuel sampling and analysis. - if the initial performance tests show that emissions of NO X are less than 70 percent of the applicable standards in If an operator is not required to and elects not to install any CEMS for either SO 2 or NO X, a CEMS for measuring either O 2 or CO 2 is not required. July

126 6.4 Data collection and methodology ICF consulted the European Chemical Industry Council (Cefic) regarding the combustion plants that are the subject of this chapter. Cefic confirmed that they do not have an overview of sites with LCPs firing non-commercials fuels. As a result, Cefic surveyed its members with a number of questions to identify combustion plants of rated thermal input 50MW th or more burn liquid production residues as non-commercial fuels in the EU. In total, replies were received from 8 Cefic members. Some of these members operate multiple sites in the EU and some of the responses were aggregated for the sites. Therefore, Cefic estimated that between 15 and 20 sites responded in total to the survey. Cefic has emphasised that whilst the precise numeric values should not necessarily be considered as accurate, they could be considered as indicative of general trends in the EU. A second data source that has been utilised is a Cefic Study about NO x reduction in large combustion plants in petrochemical plants 53 that has been submitted as part of the LCP BREF review process that is currently on-going. The study examines utility boilers in these plants that simultaneously burn non-conventional gaseous and liquid multi-fuels. This study highlights that owing to the highly variable composition of such fuels, these plants have highly variable emission levels and that this is a key parameter to be taken into consideration when determining possible BAT. This study has informed the sections in this chapter on fuels used, emissions and abatement techniques. A third data source was investigated but not used: data gathered by the EIPPCB under the LCP BREF development on four boiler combustion plants firing liquid production residues. 54 However this small dataset was deemed of limited use. As a definitive list of chemical installations firing non-commercial liquid production residues is not available, the existing LCP emission inventory for 2009, has been used alongside the E- PRTR database 55 in an attempt to identify which LCPs in the LCP inventory are both (a) chemical plants and (b) using liquid fuels. The mapping exercise of identifying which LCPs in the inventory are in chemical installations as listed in E-PRTR together with identifying from the LCP inventory of those plants whose energy input is either wholly or partially from liquid fuels provides an upper limit on the LCPs of interest to this chapter. A final data source that was assessed and analysed was information received from the German authorities indicating which LCPs in the German LCP inventory are situated in chemical installations, and of these which are known to use liquid production residues for own consumption. In light of no further detailed data, this German dataset has been used as an indicative result for extrapolating from the complete list of LCPs that are in chemical installations and which fire liquid fuels to those which are combusting liquid production residues. 6.5 Overview Numbers The total number of LCPs in chemical installations in the EU burning liquid production residues as non-commercial fuels is estimated to be currently between 25 and 50. This is a 53 This study discusses the feasibility and the performances of the BAT to reduce the NO x emissions from existing large combustion plant burning multi-fuels within the petrochemical industry (load range = 100 to 300 MWth). 54 Personal communication with DG JRC / EIPPCB, 12 th December July

127 crude estimate is based on an identification of chemical installations from the 2009 LCP inventory (225 LCPs in total). The list is then further refined to select plants using liquid fuels (93 LCPs). Following this, in the absence of any other refined datasets, the German response is used to estimate what proportion of these use liquid residues as fuel. The German case indicates that between 27% and 54% (by capacity) of known chemical LCPs using liquid fuels operate with liquid residues. 56 It is noted that this is a very crude estimation in the absence of more detailed information. The Croatian authorities were also consulted with and according to their inventory, there are no sites in chemical installations using non-commercial fuels in Croatia Rated thermal input of the plants The total rated thermal input of chemical installations in the EU burning liquid production residues as non-commercial fuels is estimated to be between 8 GW th and 16 GW th. This figure represents the total capacity of the relevant plants, not just the participation of the residues in the overall rated thermal input. The split of this total capacity across capacity classes for known chemical installations burning liquid production residues as noncommercial fuels in Germany is shown below in Figure 6.1 (e.g. only includes the 6 LCPs known to be burning liquid production residues). The figure indicates that around two thirds of installed capacity is in the 100 to 500 MW th range. There is one plant in the very large capacity class (>500 MW th ). Figure 6.1 Estimated rated thermal input of chemical installations burning non-commercial liquid fuels in Germany (source: this study, based on data from authorities in Germany) It is estimated that the total rated thermal input of chemical installations in the EU burning liquid production residues as non-commercial fuels represents between 0.6% and 1.1% of EU combustion plants of 50 MW th or more (as listed in the 2009 LCP inventory). 6.6 Fuels used The Cefic survey indicated that in the chemical industry, the three predominant sources for non-commercial fuels are: Purge streams from other upstream and downstream processes; 56 This range results from the categorisation of LCPs in the German case. 6 of the LCPs are known to be burning liquid production residues as fuels, 10 are known to not be using them and the status of remaining 7 is not known. Therefore, the low estimate excludes the latter category, whilst the high estimate includes them. July