Deliverable D1.2.3, type R

Transcription

1 TRANSPHORM Transport related Air Pollution and Health impacts Integrated Methodologies for Assessing Particulate Matter Collaborative Project, Large-scale Integrating Project SEVENTH FRAMEWORK PROGRAMME ENV Transport related air pollution and health impacts Deliverable D1.2.3, type R Emission factors for shipping final dataset for use in Transphorm emission inventories Due date of deliverable: project month 18 Actual submission date: project month 24 Start date of project: 1 January 2010 Duration: 48 months Organisation name of lead contractor for this deliverable: IVL Scientist responsible for this deliverable: Jana Moldanová Revision: [1]

2 D1.2.1 TRANSPHORM Deliverable Contents Emission factors for shipping final data for use in Transphorm emission inventories 3 Introduction 3 International legislation on emissions from shipping 5 Legislation for inland waterways in Europe 7 Abatement techniques for reduction of air pollution 8 Comparing to Emission factors for biofuels with conventional fuels 9 Emission factors for SO 2 and CO 2 11 Emission factors for VOC and CO 12 Emission factors for NO X 13 Emissions factors for PM mass 15 Emissions factors for particle number concentration 20 Emissions factors for PAHs 22 Conclusions 25 References 26 2 of 27

3 Deliverable TRANSPHORM D1.2.1 Emission factors for shipping final data for use in Transphorm emission inventories Jana Moldanová 1, Erik Fridell 1, Andreas Petzold 2, Jukka-Pekka Jalkanen 3, Zissis Samaras 4 1 IVL, Swedish Environmental Research Institute, Box 5302, Gothenburg, Sweden 2 Deutche Zentrum für Luft- und Raumfahrt, Institut für Physik der Atmosphäre, Oberpfaffenhofen, Wessling, Germany 3 Finish Meteorological Institute, P.O. Box 503, Helsinki, Finland 4 Department of Mechanical Engineering, Aristotle University, Thessaloniki, Greece Introduction This report is a final version of review and recommendations of emission factors for particulate matter (PM) expressed as mass, number size-distribution as well as some characteristics of the particle emissions such as black carbon, organic carbon and PAHs. Emission factors for a number of other compounds gas-phase compounds are described briefly. Effects of engine type, fuel quality as well as emission cleaning technologies on emission factors are described. Chapter is also devoted to legislation for emissions from inland and national and international maritime shipping. The recommended emission factors include data reported in open literature, data measured and reported by Transphorm partners within other projects and a new dataset produced in Transphorm measurement campaign. Emissions from a fleet of ships are usually calculated by means of quantifying the fuel consumption by power production first and then multiplying the consumption by emission factors. Some inventories use the bunker sales statistics as a direct estimate of the fuel consumption together with an assumption of a distribution of ship and engines types. Others, including the bottom-up inventories, estimate the power production, and thus the fuel consumption of individual ships, from fleet movement statistics. Emission factors (EF) used are then related either to the generated power EF p (g (species) /kwh) or to the fuel consumed EF f (g (species) /kg (fuel) ), where the first one multiplied by the specific fuel consumption (SFC, unit g (fuel) /kwh) is equal to the second one. Emissions from a marine engine will depend on the type of fuel used as well as on characteristics of the engine. The most important fuel parameters are if the fuel is heavy fuel oil (residual fuel, HFO) or marine distillates (marine gasoil, MGO or marine diesel, MDO) and the sulphur content (FSC). The emissions likely depend on the viscosity and the aromatics content of the fuel but there is not sufficient data to link emission factors to these parameters. There are some other fuels that are much more uncommon such as biodiesel, coal and natural gas. Emissions have been found to vary significantly between engines. Probably the maintenance and age of the engine are important for certain emission factors. For calculating emissions one usually considers the engine power, the engine speed and the emissions standard. The latter applies to nitrogen oxides only. However, one can suspect that the emissions standard also will influence the emissions of, e.g. particles and hydrocarbons, although there is, for most cases, not enough data available to draw conclusions about this. The engines on ships are usually one or several main engines, used for propulsions and a number of auxiliary engines, used for thrusters, electricity generation, pumps etc. The main engine is often equipped with a shaft generator that produces electricity when the main engine is in operation. In addition there are usually a number of boilers for steam and hot water production fuel heating etc. Dominating sources of emissions at open sea are the main engines. At berth these are usually turned off while the auxiliary engines still are being used. For those engines on ships operating with a 3 of 27

4 Relative fuel consumption rate [*100%] D1.2.1 TRANSPHORM Deliverable power plant principle, main engines are used for electricity generation and electrical motors are used for propulsion. In these cases main engines are used both for propulsion and all additional equipment power needs thus obviating the need for separate auxiliary engines. Here, main engines are run also during harbour visits. The engines are usually divided by engine speed into slow speed ( rpm), medium speed ( rpm) and high speed ( rpm). Most modern cargo ships use slow speed, two stroke engines or medium speed, four stroke engines. Later are common in passenger vessels. Some smaller vessels may use high speed, four stroke diesel engines. The specific fuel consumption (SFC), expressed as mass of fuel per unit of work by the engine (g/kwh), depend on the engine type and on the type of fuel used. Typical values can be found in Table 1. Note that the SFC varies between different engines and will typically be lower for larger engines than for smaller. The fuel consumptions in Table 1 are for typical design speeds which usually correspond to an engine load of 80-85% of the maximum engine power. If the engine is used at lower or higher loads the specific fuel consumption is typically higher. Variation of SFC with engine load and engine speed can be seen in Figure 1. Table 1. Specific fuel consumption for marine engines (Cooper and Gustafsson, 2005) Engine type Fuel type SFC (g/kwh) Slow speed Residual oil 195 Marine distillates 185 Medium speed Residual oil 215 Marine distillates 205 High speed Residual oil 215 Marine distillates Wärtsilä CAT MAN Engine Load [*100%] Figure 1. The relative specific fuel-oil consumption (SFOC) as a function of the relative engine load, based on the data of three engine manufacturers: Wärtsilä, Caterpillar and MAN. Figure from Jalkanen et al. (2011). 4 of 27

5 Deliverable TRANSPHORM D1.2.1 Emissions of some species like SO 2, CO 2 and metals are directly proportional to the SFC and fuel composition, regardless the type of engine or its operation regime (abatement techniques not accounted). Others, like NO X, VOC, CO and PM are dependent on combustion regime and thus on type of engine, its power setting and on physical properties of the fuel. International legislation on emissions from shipping Legislation is in force to control the emissions from shipping through Annex VI of the Marine Pollution Convention (MARPOL) that was adopted in 1997 by the Marine Environmental Protection Committee (MEPC) of the International Maritime Organisation (IMO) and came into force in May 2005 (IMO, 2006). Annex VI with its amendment from October 2008 put limits on emissions of SO 2 and NO X globally and contains provisions allowing establishment of Emission Control Areas (ECA) with more stringent reductions of fuel sulphur content and of emissions of NO X, or both (IMO, 2009). Emission Control Areas for PM are mentioned in Annex VI as well, however no regulation for PM as such is given and the PM reduction is expected to come from the reduction of fuelsulphur. Globally the average sulphur content in fuel is today around 2.7% while the IMO limit value is 4.5% which will be reduced to 3.5% after January 1, From the 1 st of January the fuel sulphur content will be below 0.5% which will effectively reduce the emissions of SO 2 and of sulphate particles (Figure 2). 5 % 4 % Baltic Sea ECA North Sea & English Ch. ECA IMO global limit ECA limit Global average HFO Global average MDO 3 % North America s coasts ECA F S 2 % 1 % 0 % Figure 2. The maximum fuel sulphur content (F S, in mass %) for marine fuels allowed globally and in Emission Control Areas (ECAs) given by IMO (years when different ECAs enter in force are shown) and the current average F S of HFO and MDO used by the global fleet (average fuel composition from Endresen et al., 2005) For emissions of NO X all ships newly built or with installed engine manufactured after year 2000 and prior to 1 st of January 2011 must meet the Tier I emission standard and after 1 st of January 2011 the Tier II standard (Figure 3). In addition, after the 1 st of January 2016 the Tier III standard must be met by ships with engines installed after this date when operating in NO X -emission control areas. The revised Annex VI expanded the Tier I rules on engines built between 1 st of January 1990 and 1 st of January 2000 for ships equipped with engine with a power output of more than 5,000 kw and cylinder displacement at or above 90 litres provided that an approved and certified method for reduction of NO X emissions exists for the engine. 1 This regulation may be postponed to 2025 if there is risk of fuel shortage 5 of 27

6 EF NOx (g/kwh) D1.2.1 TRANSPHORM Deliverable TIER I TIER II TIER III Engine speed (rpm) Figure 3. The maximum emission factors for NO X (g/kwh) for marine diesel engines given by IMO In Europe establishment of ECA for SO X in the Baltic Sea entered into force in May 2005, in the North Sea and English Channel in November 2006 and both ECAs entered in effect 1 year later. In these areas the allowed sulphur content has been reduced from the initial limit value of 1.5% to 1% after 1 st of July 2010 and will be further reduced to 0.1% 1 st of January In North America s coastal waters the ECA for SO X and NO X will enter into effect 1 st of August 2012 (see Figure 2). Extension of the European ECAs to NO X emissions is under the discussion. In October 2010 proposal of ECA for waters around Puerto Rico and the Virgin Islands was approved on the 61 st session of the Marine Environmental Protection Committee (MEPC 61). Establishment of this ECA will enter into force in 2012 ( So far the IMO does not specifically regulate particle emissions and there are studies showing that even with low-sulphur marine diesel, the PM emissions will still be significant (Winnes and Fridell, 2009). The rules governing the maximum permitted content of sulphur in fuels used for international shipping as determined by Annex VI of the MARPOL73/78 have been transposed into EU law and complemented by directive 2005/33/EC on the sulphur content of certain liquid fuels which in August 2005 amended directive 1999/32/EC. Directives 1999/32/EC and 2005/33/EC (EC, 1999; EC, 2005) provide fuel sulphur content regulations for vessels operating in EU territorial seas as prescribed in Annex VI (not its year-2008 amendment). In addition all passenger vessels on regular services in EU territorial seas, also those operating outside the ECAs, must from 11 August 2006 comply with the 1.5% sulphur limit. However, in the Mediterranean Sea passenger vessels are not using 1.5 % S fuel even if they are required to do so by the directive. FSC in the Mediterranean is closer to 2.7 % than 1.5 %. These directives provide also sulphur limits for marine gas oils (MGO) and marine diesel oils (MDO) sold in the EU member states. From July 1 st 2010 the more stringent 1% FSC limit of IMO applies in European ECAs while EC is preparing legislation that will transpose the 2008-amendment of Annex VI into EU law (proposal by EC published in June 2011 is now being debated in the European Parliament and the Council). The sulphur limits for marine gas oils (MGO) and marine diesel oils (MDO) sold in EU member states provide by EU Fuel directives are the following (Figure 4): Until 10 August 2006 applied the 0.2% sulphur limit to all marine distillates used in EU territory excluding ships in the territory of Greece, the French DOM-TOM, Madeira, the Azores and the Canary Islands. Between August 2006 and December 2007, the 0.2% sulphur limit for lower grade marine diesel oils was dropped, and a less stringent limit of 1.5% sulphur was introduced to allow use of the marine diesel oils in order to 6 of 27

7 Deliverable TRANSPHORM D1.2.1 comply with the SO X Emission Control Areas, in case supplies of 1.5% S heavy fuel oil were insufficient. The exemption for Greece and the outermost regions continued to apply. Between January 2008 and December 2009 a more stringent 0.1% sulphur limit applied to high grade marine gas oils used in EU territory while the 1.5% sulphur limit for the low grade marine diesel oils continued to apply. The exemption for Greece and the outermost regions continued to apply. From 1 January 2010, the provisions originating from directive 1999/32 and relating to the use of marine gas oils in EU territory (described above) were deleted. Instead a 0.1% sulphur limit was introduced for all marine gas oils placed on the market in EU Member States territory. At the same time a 0.1% sulphur limit started to apply to all types of marine fuel used by ships at berth in EU ports and by inland waterway vessels. This applies to any use of the fuel e.g. in auxiliary engines, main engines, boilers. This legislation goes beyond IMO s Annex VI. There are following exemptions from this 0.1% limit: for ships which spend according to published timetables less than 2 hours at berth, for hybrid sea-river vessels while they are at sea, and for ships at berth which switch off all engines and use shore-side electricity. The outermost EU regions continue to be exempt from this provision, but Greece does not, apart from a 2-year derogation for 16 named Greek vessels until % Marine fuels used in EU ECAs (as established) Marine fuels used by passenger vessels in all territorial seas Marine fuels used in EU ports by ships at berts& in inland waters MGO sold in EU lower grade MDO&MGO (transient) MDO sold in EU 1.5% F S 1.0% 0.5% 0.0% Figure 4. The maximum fuel sulphur content (F S, in mass %) for marine fuels allowed in EU territorial waters and EU inland waterways given by Directives 1999/32 and 2005/33/EC. Legislation for inland waterways in Europe There are regulations regarding emissions for ships on inland waterways in Europe for NO X, HC, CO and PM as well as for the sulphur content in the fuel used. The emission limits are regulated in Directive 97/68/EG and the sulphur content in the fuel in Directive 97/70/EG. The permitted emissions are expressed as mass of emissions per engine work (g/kwh) and depend on the cylinder volume and the net power of the engine. The regulations, given for CO, HC + NO X and PM, can be found in Table 2. Fuel sulphur content used by inland navigation has been limited in EU since 2008 when a limit of 1000 ppm-wt. = 0.1 % was set. From Jan. 1 st 2011, the maximum allowed sulphur content in the fuel is 10 ppm-wt. although 20 ppm can be accepted in some cases. 7 of 27

8 D1.2.1 TRANSPHORM Deliverable Table 2. Regulations on emissions from engines for inland waterways Category: volume/net power (SV/P) (liter per cylinder/kw) Date CO (g/kwh) HC + NO X (g/kwh) PM (g/kwh) V1:1 SV<0.9 och P 37 kw V1:2 0.9 SV V1:3 1.2 SV V1:4 2.5 SV V2:1 5 SV V2:2 15 SV 20 och P <3 300 kw V2:3 15 SV 20 och P kw V2:4 20 SV < V2:5 25 SV < Abatement techniques for reduction of air pollution Emission factors need to take into account effects of abatement techniques. Some reductions of emission factors are summarized in Table 3. Most abatement techniques focus on the emissions of NO X. The most effective technique is selective catalytic reduction (SCR) where NO X react with an added reducing agent (normally urea) over a catalyst to produce nitrogen gas. The process is very efficient and reduction factors of 95% can be reached. There is however a certain minimum exhaust temperature needed for the reaction to take place. Pre-turbo installations of SCR on 2-stroke engines can be problematic. Many SCR installations are equipped with an oxidation catalyst in order to minimise the ammonia slip. This will also lead to the oxidation of CO and hydrocarbons thus reducing those emissions. Further, an SCR will also influence the PM emissions although the details are not quite clear. Other techniques for NO X reduction includes exhaust gas recirculation (EGR), engine modifications and different techniques to introduce water into the engine (humid air motor, HAM, direct water injection, DWI, emulsifier). These techniques can be used in combinations as e.g. Fuel-Water emulsion and EGR which together can reduce 90% of NOx. Scrubber techniques can be used to reduce the emissions of sulphur oxides to the atmosphere. The scrubbers can operate either with seawater or with freshwater under the addition of an alkaline compound. The scrubbers will trap the SO X as sulphates in the water. The efficiency will depend on, among other things, the alkalinity of the water and the volumes. Scrubbers will also capture particles but the efficiency varies between different reports. Table 3. The various abatement techniques and their evaluated emission reduction efficiencies. Abatement technique EF NOx EF SOx EF CO EF VOC EF PM EF NH3 Low NO X engine technologies 1 20% ±0 * ±0 ±0 Exhaust gas recirculation % Direct Water Injection % ±0 ±0 ±0 Humid Air Motor % ±0 ±0 ±0 Selective Catalytic Reduction 1 91% ±0 ±0 ± g/kwh SCR + oxidation catalyst 2 90% 70% 80% Sea Water Scrubber 3 ±0 95% ±0-80% Fuel Emulsifier 3 10% Wetpac 3 50% * Some increase possible Unconfirmed up to 50 % reduction Value from Jalkanen et al. (2011). According to Corbett (2010) reductions range from -98% to -45%, largest fractions of PM are reduced more effectively than the small ones. 1 Lövblad and Fridell, Cooper and Gustafsson, Jalkanen et al., of 27

9 Deliverable TRANSPHORM D1.2.1 Comparing to Emission factors for biofuels with conventional fuels The modification of emissions of exhaust compounds CO 2, NO x, hydrocarbons, and particulate matter from medium-speed marine diesel engines was studied for a set of fossil and biogenic fuels in the German project BIOCLEAN (Petzold et al., 2011). Applied fossil fuels were the reference HFO and the low-sulphur MGO, biogenic fuels were palm oil, soybean oil, sunflower oil, and animal fat. Emissions of core gaseous species CO 2, CO and NO X do not vary significantly between fossil high-sulphur HFO (FSC 2.7wt.%) and low-sulphur fossil and biogenic fuels. Emissions of gaseous hydrocarbon compounds relative to HFO are significantly increased for MGO while respective emissions for biogenic fuels are similar to HFO (10% load) or reduced at most to 40% (75% load). The increase in HC emissions for MGO may be linked to the measurement method, because for MGO exhaust the FID sensor detects all HC in the gas phase while for HFO some of the hydrocarbons appear in the condensed phase. Neither EF(HC) of Cooper and Gustafsson (2004) nor Transphorm measurements show an increase in EF(HC) from HFO and MGO. Figure 5 summarises the emissions relative to HFO for an engine load of 75% representing cruise conditions. When using low-sulphur fuels (MGO, biogenic fuels), the emissions of particulate matter (PM) by mass is strongly reduced compared to HFO. This effect is of similar magnitude for all lowsulphur fuels of either fossil or biogenic origin. The reduction in PM mass emissions can be attributed primarily to the reduction in sulphate emissions, but also BC emissions are significantly lower. For all investigated fuels including HFO, emissions of PM and BC are strongest at low loading and decrease with higher loading. Considering all investigated engine load conditions, PM emissions relative to HFO are reduced to 6-25% for MGO and to 6-60% for biogenic fuels. Reductions in BC relative emissions vary from 13% to 30%, with MGO showing the strongest reduction to 13%, while soybean oil was found to emit significantly higher BC than the other low-sulphur fuels. Particle size distributions measured in the exhaust are shown in Figure 6. Both exhaust aerosols are characterized by a strong nucleation particle mode (Mode 1) in the size range dp < 10 nm. For palm oil, Mode 2 is centred at dg = 13 nm, while the respective mode for HFO is centred at dg = nm. This mode likely contains primary BC particles. At 100% load the size distribution of HFO exhaust aerosol features a pronounced peak at dg = 55 nm, while the size spectrum for biogenic fuels shows Mode 3 at dg = 85 nm. In this particular size range, Mode 3 is reduced in number density by up to two orders of magnitude at 100% load, and still by a factor of two at 10% load. Modes 3 and 4 are assumed to be made up of BC agglomerates. The reduced soot particle mode coincides with a strong reduction in BC mass emission, and in emissions of nonvolatile PM by number for biogenic fuels. The increased emission of total PM by number (see Figure 5) is mirrored in the exceedance of particle size spectra for biogenic fuels compared to HFO particularly for nucleation mode particles with dp < 20 nm. Emissions of BC and particle number provide a consistent picture of the modification of PM emissions from marine diesel engines when switching from HFO to low-sulphur fuels of biogenic origin. PM emitted from biogenic fuels is composed almost entirely of carbonaceous matter like OM and EC; see Figure 7 for a comparison of relative chemical compositions for fuels HFO, palm oil and animal fat. Sulphate and sulphate-associated water which dominate PM from HFO, do not contribute to PM from biogenic fuels. Changes in emissions were predominantly related to particulate sulphate, while differences between low-sulphur fossil fuels and low-sulphur biogenic fuels were of minor significance. 9 of 27

10 dn / dlog d p cm -3 STP emission per kwh relative to HFO (75% load) D1.2.1 TRANSPHORM Deliverable CO 2 NO x CO HC CH 2 O PM OM BC N total N nonvol Fuel sequence from left to right: MGO - Palm oil - Animal fat - Soybean oil - Sunflower oil Figure 5. Emissions of gaseous (top row) and particulate (bottom row) compounds per kwh of generated power relative to heavy fuel oil (HFO) as the fossil sulfur-rich reference fuel for investigated fuels marine gas oil (MGO) and biogenic fuels at 75% engine load; used abbreviations for particulate matter compounds are explained in the text (a) (b) % load HFO Palm oil particle diameter d p, nm 100% load HFO Palm oil particle diameter d p, nm Figure 6. Particle number size distribution at 10% load (a) and 100% load (b) for HFO as the reference fuel and for palm oil representing biogenic fuels; particle size spectra were measured by DMA (d p = nm) and by OPC (d p > 250 nm) instruments, solid lines represent 4-modal log-normal size distributions fitted to the data. 10 of 27

11 mass fraction, % Deliverable TRANSPHORM D Heavy fuel oil Palm oil Animal fat EC OM Ash SO4 H2O engine load, % Figure 7. Fractional chemical composition of particulate matter emitted from a large Diesel engine operating on fossil sulfur-rich heavy fuel oil (HFO), and on biogenic low-sulfur fuels palm oil and animal fat. Emission factors for SO 2 and CO 2 The emission of SO 2 is proportional to the fuel consumption and the sulphur content in the fuel. This is because virtually all the sulphur in the fuel will be oxidised into SO 2 in the engine. The emission factor expressed as mass of SO 2 emitted per mass of fuel consumed is therefore EF SO2 (g/kg fuel) = f S (%) * 20, (E 1) where f S is the mass fraction of S in the fuel (in weight per cent) and factor 20 (19.97) comes from recalculation of the molar weight from S to SO 2 and from % to g/kg. To express the emission in mass per engine work the specific fuel consumption must be used. In a more detailed analysis one should consider that some sulphur is oxidised further into SO 3 and may form sulphate particles. This is typically on the order of 1-5 per cent of the S-content in the fuel, depending on the engine load (Petzold et al., 2010). Some measurements indicates much higher loss of S from the gas phase (10-20%), the fate of this S is, however, up to date unknown and more research on this issue is needed before any recommendation on EF modification can be made. In a corresponding way the emissions of CO 2 will be dependent on the carbon content in the fuel and the fuel consumption. This then neglect the small fraction of the carbon that will be emitted as carbon monoxide, organic compounds and soot. The sum of these will typically be two to three orders of magnitude lower than the CO 2 -emissions. The carbon content in marine fuels can vary somewhat but is normally around 87%. Table 4 shows the emission factors for SO 2 and CO 2 for different engine types expressed in mass of emission per engine work and mass of emission per mass of fuel consumed. For inland waterways the sulphur content will be around 10 ppm giving an emission factor for SO2 of 20 mg/kg fuel. The CO2 emission factor will be about the same as for marine gasoil. 11 of 27

12 D1.2.1 TRANSPHORM Deliverable Table 4 Emission factors for CO 2 and SO 2 from Cooper and Gustafsson (2004) Engine type Fuel type FSC EF CO2 (g/kwh) EF CO2 (g/kg fuel ) EF SO2 (g/kwh) EF SO2 (g/kg fuel ) Slow speed Residual oil 2.7% Residual oil 1% Marine distillates 0.5% Marine gas oil 0.1% Medium Residual oil 2.7% speed Residual oil 1% Marine distillates 0.5% Marine gas oil 0.1% High speed Residual oil 2.7% Residual oil 1% Marine distillates 0.5% Marine gas oil 0.1% Emission factors for VOC and CO The emissions of hydrocarbons and carbon monoxide represent incomplete combustion of the fuel. These emissions from marine diesel engines are typically small due to the lean burning conditions and stable engine loads, but sharp increases may occur during rapid load changes of engines (acceleration/deceleration phases) because of incomplete combustion of fuel. Typical emission factors can be found in Table 5. Emission of CO and HC also increase at lower load. The details in the emissions at lower loads will depend on the operation and on the individual engine. Figure 8 summarizes variation of EF CO with engine load. For CO the trend found in Transphorm 2 campaign were lower comparing to Sarvi et al. and more similar to Cooper and Gustafsson (2004). For HC the variability of EF was found to be large in the Transphorm campaign and the trend for EF(HC) dependence on the engine load was drawn through the data of Sarvi et al. (2008) and Cooper and Gustafsson (2004). The emission factors for engines for inland waterways will be similar to those in Table 5. The emission regulation for CO on 5.0 g/kwh will likely have little impact on the emissions. Table 5. Emission factors for CO (EF CO ) and HC (EF HC ) from Cooper and Gustafsson (2004). Engine type Fuel type Operational mode EF CO (g/kwh) EF CO (g/kg fuel ) EF HC (g/kwh) EF HC (g/kg fuel ) Slow speed Residual oil At sea Manoeuvring Marine distillates At sea Manoeuvring Medium speed Residual oil At sea Manoeuvring Marine distillates At sea Manoeuvring High speed Residual oil At sea Manoeuvring Marine distillates At sea Manoeuvring of 27

13 EF (g/kwh) EF (g/kwh) Deliverable TRANSPHORM D1.2.1 a) Sarvi et al (HFO) Cooper et al (HFO, MGO) Transphorm 1 HFO Transphorm 2 HFO Transphorm 1 MGO Transphorm 2 MGO b) Engine load, % of max Figure 8. Effect of engine load on emission factors (in g/kwh) for CO(a) and HC(b). All tested engines are fourstroke MSD. a - The thick grey line is polynomial trend in all datapoints (EF(CO) = *EL *EL , EL = engine load in % of max, R 2 = 0.64), the dashed line is trend drawn through Cooper and Gustafsson (2004) and Transphorm 1 HFO datapoints. b - The thick grey line is polynomial trend drawn through Sarvi et al (2008) and Cooper and Gustafsson (2004) (EF(HC)= 5E-05*EL *EL , R 2 = 0.72). Emission factors for NO X The larger part (~90%) of the nitrogen oxides emitted from marine engines is formed from nitrogen in the air at the high temperatures prevailing in the combustion zones in the cylinders. The emissions of nitrogen oxides is as mentioned earlier regulated for engines manufactured after the year 2000 and for engines with a power output of more than 5,000 kw and cylinder displacement at or above 90 litres after The emission standards define the maximum allowed NOx emission factor (in g/kwh) determined by the year of installation of the ship engine and by its rated speed n. This NOx emission factor is a weighted emission factor for a certain driving cycle at standard engine inlet air humidity (10.71 g/kg) and temperature (25ºC). The driving cycle depends on type of engine (cycles C1 for variable-speed, variable-load auxiliary engine, D2 for constant-speed auxiliary engine, E2 for Constant-speed main propulsion application including diesel-electric drive and all controllable-pitch propeller installations, E3 for propeller-law-operated main and auxiliary engines) and the measured emission factor is corrected to the standard conditions. Determination of the NOx emission factors is in detail described in the NOx Technical Code (Annex 14 i.e. revised Annex VI of MARPOL from 2008, for Tier I is until 2011 possible to use Annex VI from 1997). Emission factors in Tier I standard represents engine standard of year 2000, for older vessels an engine upgrade may be needed but that is obligatory only provided that an approved and certified 13 of 27

14 D1.2.1 TRANSPHORM Deliverable method for reduction of NO X emissions exists. Tier II standard represents c.a. 20% emission reduction from Tier I and is expected to be met by internal engine combustion optimization measures. The parameters examined by engine manufacturers include fuel injection timing, pressure, and rate (rate shaping), fuel nozzle flow area, exhaust valve timing, and cylinder compression volume. Tier III standard represents c.a. 80% reduction and requires dedicated NOx emission control technologies such as various forms of water induction into the combustion process (with fuel, scavenging air, or in-cylinder), exhaust gas recirculation, or selective catalytic reduction. Typical emission factors for the different Tiers and engines speeds can be found in Table 6. Emission factors measured on Transphorm campaigns were generally lower than Tier I standard, also for engine of pre-tier I standard. Table 6. Emission factors for NO X from Cooper and Gustafsson (2004) (no Tier) and IMO regulations. The emission factors in g/kg fuel assume that the SFC will not change between the Tiers. Engine type Fuel type Emission class EF NOx (g/kwh) EF NOx (g/kg fuel ) Slow speed Residual oil No Tier Tier Tier Tier Marine distillates No Tier Tier Tier Tier Medium speed Residual oil No Tier Tier 1 * Tier Tier Marine distillates No Tier Tier 1 * Tier Tier High speed Residual oil No Tier Tier Tier Tier Marine distillates No Tier Tier Tier Tier * EF for engine speed n = 1000 to 2000 rpm, EF = 45 n (-0.2) in Tier I. EF for engine speed n = 1000 to 2000 rpm, EF = 44 n (-0.23) in Tier II. EF for engine speed n = 1000 to 2000 rpm, EF = 9 n (-0.2) in Tier III. For inland waterways the regulations put limits on the NO X emissions. For older engines (older than 2007 or 2009, see Table 2) the values in Table 6 can be used. For more recent engines the regulation limits for HC+ NO X can be used (Table 2). If the engine characteristics are not known a value of 9.0 g/kwh can be used corresponding to approximately 40 g/kg fuel. 14 of 27

15 EF [g/kg-fuel] Deliverable TRANSPHORM D1.2.1 Emissions factors for PM mass Particles emitted by marine engines consist of a volatile and non-volatile fraction. Volatiles are mainly sulphate with associated water and organic compounds. Non-volatiles consist of elemental carbon (soot, char) and of ash and mineral compounds containing Ca, V, Ni and other elements. Because of the high content of condensable matter in the exhaust the methodology of sampling impacts the PM mass found. Sampling directly in the hot exhaust captures to a large extend only the non-volatile part of PM while sampling in the diluted and cooled exhaust captures also some of the volatiles. The amount, however, depends on the dilution and temperature program of the sampling. Figure 8 shows the difference between PM sampled in the hot and diluted exhaust from a slowspeed diesel engine running on HFO with 1.9% sulphur unidentified sulph. assoc. water sulphate OC EC ash 1 0 dilluted hot Figure 9. Composition of PM (as EF) collected on filters in the diluted and hot exhaust gas (Moldanová et al., 2009). Emissions of PM varies with fuel type, fuel sulphur content and engine operation mode. Table 7 shows EF PM for cruise conditions published by Cooper and Gustafsson (2004) for different marine engines for HFO and MDO fuels together with EF PM from the Lloyds emission database (European Commission, 2002). These emission factors are based on larger number of measurements (c.a. 45 measurements at IVL database and 25 in Lloyd s database). The mean FSC in Cooper and Gustafsson (2004) was 2.3% and in EC (2002) 2.7%. The PM measurements reviewed in these reports were performed using the partial dilution equipment, i.e. corresponding to the PM in diluted exhaust in Figure 8. When more recent data from individual engines are added, one can see a span of EF PM for engines using RO between 1 and 13 g/kg fuel with the mean around 7, and for engines using MD between 0.2 and 1 g/kg fuel. Figure 10 shows a plot of the available data on EF PM at cruise conditions (engine load 75-90%) against the fuel-sulphur content (FSC). We can see a clear positive trend in emission factor for PM against the FSC for data measured on engines using RO. Emission factor for MSD engines from Cooper and Gustafsson (2004) are lower comparing to other data, however there are only few more individual measurements for this engine and fuel category available. The Lloyds EF for SSD engines are included (EC, 2002) assuming FSC 2.7% which is Lloyd s estimate of the global mean FSC for RO. 15 of 27

16 EF PM [g/kg fuel] D1.2.1 TRANSPHORM Deliverable Table 7. Emission factors for PM mass (EF PM ) for cruise and manoeuvring conditions from Cooper and Gustafsson (2004) for different marine engines for residual oil (RO) and marine distillates (MD). The mean FSC is 2.3 wt.% for RO and 0.4 wt.% for MDO. EF PM for cruise conditions from the Lloyds emission database (European Commission, 2002) for the global fleet, these EF PM are weighted for fuels used by the engine category. The global FSC in this study is 2.7 wt.%. Cooper & Gustafsson (2004) EC (2002) Engine type Fuel type EF PM at sea EF PM manoeuvring EF PM at sea g/kwh g/kg fuel g/kwh g/kg fuel g/kg fuel SSD MD SSD RO * MSD MD MSD RO * HSD MD HSD RO * Mixture of MD and RO RO RO Tr. MD MD Tr EC (2002) Cooper (2004) Cooper (2004) FSC [wt.%] Figure 10. Emission factors for particle mass EF PM as a function of FSC (in wt. %). EF PM for RO is plotted in blue, EF PM for MD is plotted in green. Datapoints with crosses (Tr.) are from the Transphorm measurement campaigns (the dashed line fitted through the RO data has equation EF(PM) = 2.084*FSC , R 2 = 0.75) PM mass emission factors change with the engine load. In the stack typically between 1 and 5% of sulphur is oxidized to SO 3 (Moldanová et al., 2009; Petzold et al., 2010; D2.1.4) and contributes to the exhaust PM. Petzold et al. (2010) showed a positive correlation between the SO 2 in-stack oxidation and the engine load for engines using HFO with similar fuel sulphur content (between 2 2.5%) (Figure 11a), Transphorm measurements showed an increase in S oxidation from 0.2 to 1.4%. While EF for sulphate is positively correlated to the engine power, i.e. contributes most to the PM emissions at high engine loads, emissions of black or elemental carbon and of organic carbon are higher at low engine loads and have their minima at loads around 50% and increases somewhat at cruise conditions (Figure 11b, Petzold et al., 2010). The resulting dependence of EF PM on engine load thus varies with fuel sulphur content and potentially also with fuel type. One should also remember that the fuel consumption of course also varies with the engine load making the emissions (in g/hour) higher at cruising that at low loads. 16 of 27

17 S conversin, % EC, OM [g/kg fuel] SO4=, PM [g/kg fuel] Deliverable TRANSPHORM D1.2.1 In the STEAM2 model Jalkanen et al. (2011) use EF PM (emission factor for total particulate matter mass) and emission factors for 5 different PM components: EC, OC, sulphate, ash and the sulphate-associated water as a function of engine load and the FSC. The FSC dependence is built on data from the 2nd IMO GHG study (IMO, 2009) and the dependence on engine load on data from Agrawal et al. (2008a), Petzold et al. (2008) and Moldanova et al. (2009). Table 8 and Table 9 show emission factors from STEAM2 for a span of FSC and engine loads. The increasing trend in S oxidation with increasing engine load as shown in the previous paragraph and in Figure 11 is however not present in Jalkanen et al. (2011). The recommended EFs For EC are for HFO 0.5 g/kg fuel at low engine load and 0.2 g/kg-fuel for high engine load and for MGO 0.3 g/kg-fuel at low and 0.1 g/kg fuel at high engine loads (MGO based on D2.1.4). a) b) Ref (1) 2.32% Ref (1) 2.21% Ref (2), 2.05% Ref (3), 2.85% Ref (4), 1.95% Ref (5) Tr1, 0.91% Tr2, 0.96% Tr2, 0.58% Ref (6), 0.16% Tr1, 0.1% EC OM SO4= PM Engine load, % of max Engine load, % of max Figure 11. a - Efficiency for converting fuel sulphur to particulate-matter sulphate at various engine loads and for fuels with different sulphur contents given in wt-%; the dashed lines represent linear relationships between part of sulphur in exhaust converted to sulphate and engine load, the grey for HFO, the orange for MGO. (Ref (1): Petzold et al., 2010, Ref (2): Agrawal et al., 2008a; Ref (3): Agrawal et al., 2008b; Ref (4): Moldanová et al., 2009; Ref (5): Kurok, unpublished, Ref (6): Kasper et al., 2007, Tr1, Tr2: Data from Transphorm campaigns S1 and S2, D2.1.4). b - Mass emission factors for carbon-containing compounds, sulphate and PM in the raw exhaust gas, FSC 2.40wt-% (filled symbols) and 0.91% (open symbols) (EC - elemental carbon, OM organic matter, both analysed by multi-step combustion method) (from Petzold et al. 2010, data in their Table 1 and from D2.1.4). Table 8. Total PM 2.5 emission factors (g/kwh) at selected engine loads and fuel sulphur content (wt-%) Load 0.1% S 1.0% S 1.5% S 2.7% S 3.5% S 20% % % % % of 27

18 D1.2.1 TRANSPHORM Deliverable Table 9. The emission factors of PM 2.5 subcomponents (g/kwh) as a function of engine load. The fuel sulphur content is 1.5 wt-% Load EC OC Ash SO 4 = H 2 O Total PM 20% % % % % In the atmosphere the oxidation of the emitted SO 2 proceeds and in ship plumes sulphate becomes the dominant component of the PM. If all sulphur corresponding to 1% FSC would be oxidised into H 2 SO 4 and this H 2 SO 4 would condense on particles in the ship plume, one would get an emission factor for particulate H 2 SO 4 of 30.6 g/kg fuel and, further, if also the water associated to sulphate is accounted for an emission factor of 67 g/kg fuel is obtained. The plume studies (D1.2.2) have shown that between ~1% (polluted urban air, winter conditions) and 60% (clean background, summer conditions) of emitted SO 2 contributes to the PM, giving EF(H 2 SO 4 nh 2 O) 1-40 g/kg fuel. These numbers can be compared to the typical EF PM that are a few grams per kg fuel. Cooper and Gustafsson (2004) assumed that PM emissions from ship diesel engines are dominated by particles with diameters less than 1 µm (i.e. TSP = PM 10 = PM 2.5 ) based on general consensus at that time. However, measurements of Cooper (2003) indicate that about 50-70% of the total suspended particles (TSP) could be as PM 2.5 and the remainder as PM 10 (Cooper, 2003). Other studies show presence of larger (tar-like and re-entrained) particles in the exhaust from engines using heavy fuel oil (Lyyränen et al., 1999; Fridell et al., 2008; Moldanova et al., 2009) indicating that TSP may be larger than PM 10. One can anticipate that the particle size distribution will be dependent on fuel type, engine type, operation and age of the gas plume. Emission factors of different metals obtained from x-ray fluorescence analyses (XRF) of filter PM deposits are summarised in Table 10. The composition of fuel with respect to S, V, Ni and ash content is shown together with the EFs. One can see a large variability in fuel composition affecting the EFs for metals. This effect is often larger than effect of engine load. Ca, Zn and P are associated with lubricant oil. Emissions from inland waterways have been regulated first in for PM mass (Table 2), in 2010 for FSC with limit 0.1% and in 2011 with FSC limit of 10 ppm. Emission factors for year 2010 would be thus similar to those for MGO, EFs after 2011 are difficult to assess mainly because the regulation is very recent and no measurements have been found with this fuel. Further, it seems unlikely that the emissions are close to the permitted values in Table 2 when considering that measurements on marine engines using gasoil (with 100 times more sulphur) show lower values (Table 7 and Table 8). A linear extrapolation of the data in Table 8 gives an emission factor of about 0.30 g/kwh corresponding to about 1.5 g/kg fuel. PM emitted from large pre-euro and Euro I Diesel engines has 51% of EC and 35% of EC. These numbers can be used as proxy for inland navigation emission factors until new measurements are available. 18 of 27

19 Table 10. Emission factors (in g/kg fuel) for metals and other elements obtained from XRF analyses of PM filter deposits. EF from Ref 1 - Agrawal et al. (2008b), Ref 2 - Agrawal et al. (2008a) and from Transphorm campaigns (D2.1.4) are compared. Column 2-5 are concentrations in fuel. Ash S V Ni Mg Al Si P S K Ca V Cr Mn Fe Co Ni Cu Zn Fuel wt.% Fuel mg/kg g/kg fuel HFO, 25-30% load Ref <d.l <d.l <d.l <d.l. <d.l <d.l <d.l <d.l <d.l. <d.l Tr <d.l. <d.l. <d.l. <d.l <d.l. <d.l <d.l HFO, ~50% load Ref <d.l <d.l <d.l <d.l. <d.l Ref <d.l <d.l <d.l <d.l. <d.l Tr2 a <d.l <d.l <d.l <d.l Tr2 b <d.l. <d.l <d.l <d.l <d.l HFO, 70-85% load Ref <d.l <d.l <d.l <d.l. <d.l Ref <d.l <d.l <d.l. <d.l Tr <d.l MGO 50-75% load Ref 1 < <1 < <d.l. <d.l <d.l <d.l. <d.l <d.l Tr1 < <1 <1 <d.l. <d.l. <d.l. <d.l <d.l <d.l. <d.l. <d.l. <d.l. <d.l. <d.l. <d.l Tr2 < <1 <1 <d.l. <d.l <d.l <d.l <d.l <d.l <d.l MGO 25% load Ref 1 < <1 <1 <d.l <d.l <d.l <d.l. <d.l

20 EF N [10 16 /kg fuel] EF N [10 16 /kg fuel] EF N (airborne) [10 16 /kg fuel] Emissions factors for particle number concentration Emission factors for particle number concentrations are in the order of magnitude of /kg fuel. Measurements, presented by Petzold et al. (2010), on a 4-stroke MSD burning RO show emission factors for total particle number, EF N, between 1 and 4.5 x /kg fuel, with a positive correlation between EF N and the engine load (Figure 12a). Transphorm measurements showed consistent, for 80% engine load operation mode somewhat higher EF N ( x10 16 /kg fuel), also these with positive correlation between EF N and engine load. Airborne measurements in ship plumes have shown EF N of the same order of magnitude (Lack et al., 2009; Petzold et al., 2008; Murphy et al, 2009; Jonsson et al., 2011) (Figure 12a, Jonsson et al. (2011) do not have engine load, 50% is an estimate). Petzold et al. (2010) investigated the volatility of the emitted particles in exhaust from a 4-stroke MSD test engine using a thermo denuder. They found that 2/3 of the particles at high load and 1/3 at low load were volatile and that the number of non-volatile particles did not change for loads >20%. The increase in total particle emissions with load by a factor of 3 was almost entirely attributed to sulphuric acid-water droplets. Figure 12b shows the number concentrations of particles in the accumulation mode, i.e. those with diameters in the range µm, measured on test engines and in ship plumes. The Transphorm measurements have shown lower proportion of volatile particles (~1/3). This is consistent with lower sulphate content of the measured particles partly resulted from lower FSC of the fuel. a) b) total (test engine) total (Tr1) nonvolatile (test e.) nonvolatile (Tr1) total (airborne) total (Tr1, MGO) nonvolatile (airborne) N(0.1-3 μm) (test e.) N(0.1-3 μm) (Tr1) E E E E Engine load, % of max 0.0E Engine load, % of max Figure 12. Emission factors for particle numbers measured on test engine burning RO with FSC 2.21 (test engine), and 0.91 (Tr1) and in airborne measurements in ship plumes. a total and non-volatile particles, b particles in accumulation mode. (from Petzold et al., 2010, D2.1.4 and Jonsson et al., 2011) Jonsson et al. (2011) found in measurements of ship plumes of 4 identified ships performed with and without thermodenuder from the coast next to the shipping line that 1/3-2/3 of particles were volatile. Pirjola et al (2011) found in similar measurements that the volatile particles were responsible for 55-61% of the total particulate number. These measurements also showed the growth of particle diameter by condensation of volatiles (Figure 13). The data of Pirjola et al. (2011) include con-

21 Deliverable TRANSPHORM D1.2.1 tribution of the local scale dispersion. The plots of particle number size distributions in plume at different distances from the source (Figure 14) show that PM concentration levels in ship plumes typically reach background concentration levels in ~12 minutes and at a distance of five kilometres downwind from the source. Figure 13. Number size distributions of total (w/o denu) and non-volatile (with denu) part of PM in plumes from 2 different ships measured next to the shipping lane. Figure according to Pirjola et al. (2011) Figure 14. Particle number size distributions in ship plumes at different distances downwind the source. Lines correspond to distances (in km) between observation point and the vessel. Image from Pirjola et al (2011) There are no data on EF(PN) for inland shipping. The existing data for MGO indicate EF(PN) around 1x10 16 #/kg-fuel. As already mentioned the fuel used in inland shipping differs now from marine fuels, among others in FSC. Petzold et al. (2011) investigated EF(PN) from different biodiesels, fuels with FSC close to that of fuel used in inland shipping but with differences in composition with respect to other compounds. As can be seen in Figure 5, these fuels show EF(PN) higher 21 of 27