FCGEN FUEL CELL BASED POWER GENERATION PUBLISHABLE SUMMARY FINAL REPORT. Executive Summary. Hydrogen generation from diesel

FCGEN FUEL CELL BASED POWER GENERATION PUBLISHABLE SUMMARY FINAL REPORT Executive Summary The FCGEN project was founded European Community Joint Technology Initiative Hydrogen and Fuel Cells as the first of its kind that seeks to integrate PEM FC based diesel fuel powered on board APU systems. The research partners Volvo Technology (Sweden), Johnson Matthey (Great Britain), Modelon AB (Sweden), PowerCell AB (Sweden), Jožef Stefan Institute (Slovenia), Forschungszentrum Jülich (Germany) and Fraunhofer ICT IMM (Germany) have bound to design, develop and demonstrate a proof of concept dieselpowered fuel cell APU system to produce greener and more efficient electricity. The consortium partners have reached their final goal by the end of the FCGEN project and successfully demonstrated its functionality in fully automatic autonomous operation mode. The system was designed for the use in truck, possibly also for recreational vehicle or yacht applications. It generates up to three kilowatts electrical power but can be modified for larger power ranges. Therewith enough output is available to run electric and electronic small consumers as well as power guzzlers like for instance air conditioning, heaters or refrigerators. The system consists of a diesel and a water tank, a hydrogen generating module (reformer) and a fuel cell module including a low temperature PEM fuel cell (LT PEM) with 55 cells, as well as necessary BoP components, a battery, a power converter module and a control unit. In the project the whole system was built from scratch, where the fuel processor, the most complex and essential component was designed to achieve highest possible efficiency, increased lifetime, reduced size and weight, and shortened startup time, while maintaining full controllability and of the system. Hydrogen generation from diesel To extract the hydrogen, which is needed for the fuel cell, out of the Diesel, the Diesel fuel in the tank is converted into a hydrogen rich gas by autothermal reforming. This process was developed at Forschungszentrum Jülich and has already proven its high stability for duration in the range of 10,000 hours. Carbon monoxide which is also generated by the reforming process is initially converted to a remaining low concentration (< 10 ppm) by means of further reactors (plate heat exchangers by Fraunhofer ICT IMM) being part of the fuel processor. This is also valid for the sulphur which is contained in the fuel in minor amounts. The resulting gas is primarily composed of hydrogen, carbon dioxide and steam of which the hydrogen is processed in the fuel cell to generate electricity. Fuel cell system as environmentally friendly option The catalytic processes needed to convert diesel fuel were realized by using catalysts made by the catalyst producer Johnson Matthey. The fuel cell developed by PowerCell is characterized by a high long term stability. The whole system is started up by the combustion of diesel fuel and is running fully automated thanks to the control system and power conditioning modules, both developed at the Jožef Stefan Institute. The battery of the system is recharged automatically by the fuel cell. For efficiency maximization, as large as possible amount of heat is recovered by means of using cooling loops for preheating the reactants. In the power range from three to ten kilowatt so far only gasoline or diesel fuelled electricity generators based on combustion engines (APU) are available at the market. The FCGEN fuel cell system is working with a higher efficiency, a low noise emission and is virtually environmentally compatible. For the near future, when the H2 infrastructure is in its early days, as well as for the remote or mobile applications, such systems represent a very interesting alternative to current APUs. Already at the moment, there is interest for the technology and a further development of the market ready system is planned. Period covered: 1 November 2011 31 May 2015 Page 1 of 24

Project context and main objectives Context Local and global environment issues as well as the consumption and supply of energy are major challenges for the future and for the European Community. Today EU depends on 50% import of energy and it is expected to further increase in future. Greening road transport is necessary to achieve EU and world targets in emissions reductions. In the EU, 19% of total EU greenhouse gas emissions and 28% of CO2 emissions in 2005 are linked to the transport sector. More than 90% of total EU transport emissions are due to road transport. While total EU emissions declined, transport emissions increased continuously between 1990 and 2005 due to high growth in both passenger (28%) and freight transport (62%). To combat climate change and increase the EU s energy security while strengthening its competitiveness the European Commission has decided to implement the 20 20 20 targets to be met by 2020: A reduction in EU greenhouse gas emissions of at least 20% below 1990 levels 20% of EU energy consumption to come from renewable resources A 20% reduction in primary energy use compared with projected levels, to be achieved by improving energy efficiency. In the Multi Annual Implementation Plan 2008 2013 of the Fuel Cells and Hydrogen Joint Undertaking (FCH JU), Auxiliary Power Units (APU) has been identified as an important way to increase efficiency of onboard power generation and reduce CO2 emissions and local pollution from truck, marine and aeronautic applications. For mobile applications the increasing demand for electrical power when the vehicle or boat stands still has led to an increasing need for an on board electric power generator which operates with high efficiency and very low emissions. Exactly these parameters are the main drivers behind fuel cell (FC) systems that can be used as auxiliary power units (APUs). However, proton exchange membrane of the FC operates efficiently with hydrogen rather than with the ubiquitous hydrocarbon fuels such as diesel found at gas stations. Therefore a fuel cell based auxiliary power unit (APU), with a diesel fuel processor is regarded as one of the most interesting options since it combines high efficiency, low emissions and the use of the same fuel as the main engine. Such onboard generator or Auxiliary Power Unit, APU, should be able to run when the main engine is shut off. EPA has estimated the emission from long haulage trucks in US at idling to; NO x : 180 000 tons per year PM: 5 000 tons per year CO 2 : 11 million ton per year Large numbers go also for yachts in harbours. By developing a fuel cell based APU for truck application in Europe, we can compete also with the strong ongoing initiatives in US supported by Department of Energy, and contribute to: Decarbonisation of transport Ensuring mobility: reliable, safe and secure transport Global competitiveness growth & jobs The FCGEN project was founded as the first of its kind that seeks to integrate PEM FC based diesel fuel powered on board APU systems. The project consortium has been set up consisting of leading OEMs, system integrators, components suppliers and research providers with long experience of collaborative research and development on the key components and subsystem needed for the APU. The consortium has taken a cross thematic approach going from basic, performed by the project partners in previous projects, to applied research through to validation of the technology to development of an APU Proof of concept clearly responding to the topic SP1 JTI FCH.2010.1.5 in the call description. Period covered: 1 November 2011 31 May 2015 Page 2 of 24

Scientific & technical and objectives The overall objective of the FCGEN project is to develop and demonstrate a proof of concept autonomous diesel powered PEM fuel cell based auxiliary unit (FC APU) in the laboratory environment. The initial project goal was to demonstrate real application and on board a truck, but after the truck demonstrating partner (CRF) terminated their involvement, suitable replacement was not found and the goal had to be changed. Another objective is to further develop the APU key components and subsystems that are expensive and/or still behind the required level of maturity and stability. The APU system consisting of a lowtemperature PEM fuel cell, a diesel fuel processor and necessary balance of plant components is targeting close to automotive requirements regarding e.g. size, mechanical tolerances, durability etc. High targets are set for energy efficiency (30%) that in turn leads to significant emissions reductions and greener transport solutions in line with EU targets. The APU technology has been advanced in previous projects (HYTRAN, Profuel, FURIM), where research on key technologies and also complete laboratory APU fuel cell systems have been developed. Although these projects have advanced the technology considerably and proven the feasibility of fuel cell based APU, they have also clearly shown that there are several remaining challenges which need to be solved. The main identified challenges are addressed in the FCGEN project and the solutions are demonstrated in a proof of concept system. To move towards targets, the need of breakthroughs and innovations on the component level leads to the following developments planned within FCGEN: ATR diesel reformer Heat exchanger and radiator customised for the application Micro structured gas purification unit for diesel reforming High efficiency and robust DC/DC converter Electronic control unit handling all required signals Highly efficient compressor A compact and robust integrated APU system It is clear that fuel processor and sub systems are major bottlenecks towards commercial solutions of APUs. FCGEN therefore largely focuses on the development of the fuel processor subsystem and necessary components and sub systems to make them meet the requirements derived from the application. A key point in the project is the development of a fuel processing system that can handle logistic fuels. Such fuel processor consists of autothermal reformer, desulphurization unit, water gas shift reactor, reactor for the preferential oxidation of CO. A proof of concept of the developed fuel processing technology for commercial available diesel has to be demonstrated and tested in a test bench before it is integrated in a complete system for APU demonstrating in laboratory. Since cost and compactness are important issues, innovative solutions are incorporated in the fuel processor to fulfil these targets. Previous work by the partners has shown that the costs and size of the existing catalysts for the clean up system (water gas shift reaction and preferential oxidation of CO) are prohibitive. These issues are both addressed by catalysts developed with reduced platinum group metal loadings, and by presenting these catalysts in such a way as to maximize their volumetric productivity by coating them onto micro channel heat exchangers. The fuel cell stack is also a key component for successful development of the APU. In this project there is however not foreseen a specific fuel cell development, this is covered in other projects. It is important for meeting the cost targets that the fuel cell stack and component has dual or multiple uses with other fuel cell applications. In this project the specific requirements for the application were addressed. The automotive environment is harsh regarding shock and vibrations. The chemical environment also put high demands on the fuel cell with chlorine in the air for marine application and polluted air at truck stops. The Period covered: 1 November 2011 31 May 2015 Page 3 of 24

reformate gas may also content trace concentrations of hydrocarbons. By extensive laboratory testing the influence of the different pollutant were investigated. Another essential system for enabling the use of APU power, is the power conditioning. Within FCGEN a power conditioning system had to be developed that provides the quality power regardless of the load, perform and control battery charging, prevent system components from electrical shocks from load or BoP, and also provide power for the BoP components. Again the design pursued efficiency, cost and reliability targets. The BoP subsystems need to fulfil the need for the fuel processor, fuel cell stack and subsystems. Further, the BoP subsystems materials need to be compatible with the chemical environment from reformate and stack products. The components need to also withstand the truck environment regarding temperature; vibrations, shocks etc., and have minimum efficiency losses. To be able to meet the aggressive cost target, the BoP work focused on identifying on shelf components that can meet the system requirements as much as possible. Where this was not possible, specific components were developed to deliver that specific functionality. The air compressor is one of the main challenges in such system. An air system that has minimum parasitic losses and very low noises was developed to fulfil the requirements in a truck APU system. Final key point is the development of an efficient and reliable control system and the electronic control unit (ECU) for the APU, compatible with the truck systems. The developed control unit has to handle versatile and numerous sensors and actuators of the prototype APU, while incurring minimal cost, size and weight, as well as targeting automotive standard requirements. Control system has to ensure completely autonomous and fully automatic operation of the developed APU during final demonstration. Period covered: 1 November 2011 31 May 2015 Page 4 of 24

Description of the main S&T results/foregrounds WP1: Integrated system and demonstration The first work package was devoted to the integration of the developed APU system on board a vehicle and to demonstrate the performance of this system under real conditions. At the end of P2 the leading partner CRF terminated its project involvement and despite substantial efforts no replacement has been found. This resulted in removal of truck demonstration goal and related tasks no work within that WP in third period. During the first two years, the vehicle specifications, necessary for the development of the APU unit were defined and communication plus physical signal interface to the vehicle were created. The vehicle application selected for FCGEN project was the European trailer truck. The APU was planned to be mounted in the place of the 760 L fuel tank. This means that the available space is far above the FCGEN APU volume target (500 l). APU, its main components and vehicle integration location The system voltage is defined at 24 V dc, with maximal current of 125 A and maximum power of 3 kw e. The data communication interface necessary for APU controls and operation was made available and the protocol defined: experimental set and operational set (smaller message set oriented to mass market implementation). The APU ECU communication via CAN bus and an additional CAN network is setup for FCGEN APU. Further, guidelines and reference standards for crash, vibrations, mechanical shock, free fall, protection of electrical equipment against foreign objects were defined. With regard to environment conditions, it was decided that the system has to operate in open space and temperature limits in the range 18 C 50 C. A part from the work summarized above, the profit of electrifying the most conventional auxiliaries in terms of fuel saving during stop phases was estimated and the electric auxiliaries necessary to the driver when the truck is parked were selected and the supplied electric powers provided by the APU to these auxiliaries as alternative to engine in idling were defined. WP2: Complete APU system This work package is devoted for the integration of the components and sub systems which are developed within the project as well as to provide balance of plant (BoP) components that ensure a proper supply of feedstock under defined conditions to secure optimal APU performance, controllability and system durability. As the availability of some of the BoP components which fulfills the APU requirements is limited, a lot of work has been devoted to find optimal solutions either internally within the project or together Period covered: 1 November 2011 31 May 2015 Page 5 of 24

with particular BoP component suppliers. Another important target which is covered in this work package is system design and packaging optimization to meet efficiency and size targets and finally to build, commissioning and test the complete APU system. The work performed within this work package and most important results obtained are highlighted below: System design and simulation: A steady state model was developed and used to support system design work. The system design and simulation work was performed by Powercell in collaboration with Jülich and IMM using the Aspen tool. Modelon has developed component library and together with Powercell developing models for steady state and dynamic simulation of the full APU The system model has been used in the project to solve several tasks; investigate dynamic behaviour and test that the design meets all requirements, validate and tune component models based on test data for final system design, control design and hardware in loop (HIL) control verification using real time capable system models. Developed FCGEN APU (left) and fuel processor (right) models in Dymola environment BoP optimization: The system design was frozen and main components selected in the second period of the project. During the first half of the project, Powercell has been focusing on finding on shelf components which fulfil the needs of the fuel processor, fuel cell stack, and the power conditioning subsystems. The demands have been on finding components that fulfil the requirements from size and packaging point of view and demands on ambient conditions that the APU will be exposed to. The system design for which the component selection was made was developed during the spring of 2012 and frozen end of 2102. However, the system design was modified several times because of needs which were difficult to foresee at the time point when the first design was set up. The changes were both in form of relocation of some system compounds as well as adding new ones. These changes were based on the test results on components and subsystems. All commonest has been tested and characterized in a specials developed BoP test rig. Bop component selection, testing and characterization Period covered: 1 November 2011 31 May 2015 Page 6 of 24

System integration and packaging: The work in this task has been on package design and to build and test the two main subsystems and the complete APU system. Based on the CAD design Powercell build two subsystems the Fuel Cell Module (FCM) and the Fuel Processor Module (FPM). FCGEN APU system schematics (left) fuel processor CAD design (right) In the FPM the Hydrogen rich gas of the quality needed for PEM fuel cells is produced. In order to do this the FPM consist of auto thermal reformer (ATR) to produce hydrogen from diesel, a desulfurization reactor to remove sulphur from the gas stream, a water gas shift reactor and preferential oxidation reactor to reduce the CO content in the gas stream to the required level. For these reactors to work properly, a number of support systems are needed: a steam production and management system and air management system for the ATR reactor. Air supply system for the preferential oxidation reactor and for all reactors a heat management system is needed. The main task for the FCM unit is to convert the hydrogen from FPM and produce electricity. It also recover the needed process water by condensates both the anode off gas and the cathode of air. The main components in the FCM unit are: Fuel Cell Stack, Condensers, Water separators, Radiators, Humidifier, Cathode air blower, Hydrogen recirculation pump, water tanks, both cooling water and process water. Final CAD packaging models of fuel processor (left) and fuel cell module (right) Period covered: 1 November 2011 31 May 2015 Page 7 of 24

The FCM and the FPM systems were built and commissioned / tested separately, to ensure proper operation before they were send to JSI. There they were integrated together and the control unit (ECU) and the DC/DC converter were installed, forming the complete APU. Integrated complete FCGEN APU system Complete system testing: To fulfil the final goal of the project, the complete APU tests and operation demonstration were performed in the laboratory of the project partner JSI. Figure below shows the APU installed in the laboratory at JSI. The modern communication technologies were fully exploited by organized online meeting sessions during experiments, allowing all project partners to monitor the operation of the APU remotely and discuss results from their locations via Internet. The following was shared via Internet: APU main HMI (extensive HMI with all parameters of the APU) Voltage monitor, connected to the FCM module. Note that voltage monitor is not a standard part of the APU, but it was used for diagnostic purposes during the tests. By voltage monitor it is possible to monitor the individual voltages of all 55 cells in the FCM stack. FTIR computer to monitor the quality of reformate generated by FPM. FCGEN APU during testing (left) and gas content & commissioning HMI computer with remote access (right) Period covered: 1 November 2011 31 May 2015 Page 8 of 24

The following experiments were performed: test and optimization of FCM with its control functions, test and optimization of interactions between FPM and FCM, test and optimization of electricity generation, operational load range, and load change rate, test and optimization of energy efficiency of the complete APU, demonstration of the autonomous operation using battery powered startup and shutdown. At the beginning, the APU was controlled by industrial controller and first the FCM operation was tested with FC connected directly to the electronic load and this way the FC and FP control loops were verified. The feasible APU load range has been defined and safe load change rates determined (50% > 100% in one minute). Before final testing, the control hardware of the APU was upgraded according to plan. Industrial control system was removed and a much more compact and tailor made ECU, designed specifically for the FCGEN APU by the project partner JSI, was installed. The structure of the control software and all the algorithms were preserved from industrial controller, while individual function blocks were compiled from Ladder diagram to C code. Together with the new ECU, also two other tailor made units were installed: An electronic board for supply, protection (fusing) and sequential connection of BoP components to power supplies. A DC/DC converter unit for conversion raw FC stack voltage (32 50 V) to stabilised voltage 28.8 V suitable to be connected to and to charge the 24 V automotive battery pack Now the APU was connected to the automotive battery pack via its own DCDC converter. Several load changes were performed, even load fall out, and CAB blower control was tuned to ensure safe operation at all times. To optimize the system efficiency and operation robustness the CAB and FC cathode air flow as well as system temperature control loops were tuned. Further measures were taken for maximizing the water recovery. Finally the efficient autonomous operation of the APU was demonstrated using commercially available diesel fuel. The operation reached efficiencies above 25% and was able to power its start up, and shutdown and recharge the battery on its own. During numerous FP and complete APU testing sessions, the necessary measures to significantly improve the following parameters were identified for future work: bring the efficiency above 35%, shorten start up time, durability and serviceability. FCGEN APU during demonstration at final project meeting (left) and APU with final looks (right) Period covered: 1 November 2011 31 May 2015 Page 9 of 24

WP3: Fuel processor The objective of WP3 was to develop a complete fuel processor system. With respect to the overall goal of FCGEN develop and demonstrate a proof of concept diesel powered fuel cell based auxiliary power unit (APU) in the laboratory environment the task of the fuel processor system was to supply a hydrogen rich gas flow at adequate quality to the anode side of the fuel cell. Adequate quality means that the concentrations of so called non methane hydrocarbons NMHCs (ethene, ethane, propene, benzene etc.) should be ideally zero, the concentration of CO should be below 25 ppmv, the concentration of sulphur should be below 1 ppmv and the maximum molar flow of hydrogen should be high enough for the generation of 3 kw gross electric power in the APU. At the beginning of the project, the load range of the fuel processor was defined to be between 50 and 100% referred to the electric power of the APU. Design aspects such as weight, volume, cost and emissions should be thoughtfully considered. A simplified schematic drawing of the FCGEN fuel processor system with its main components (including the fuel cell stack and its bonding to the fuel processor) is shown in the figure below. The development of this drawing was a joint effort of all partners in FCGEN. FCGEN fuel processor schematics The autothermal reformer (ATR) is the key component of the fuel processor, since it has the function to catalytically convert a mixture of diesel fuel, water and oxygen into the hydrogen rich product gas (reformate) mentioned in the introduction according to these reaction equations: C n H m + n H 2 O n CO + (m/2 + n) H 2 endothermic C n H m + n/2 O 2 n CO + m/2 H 2 exothermic In the course of these reactions, special attention has to be given to the quality of the product gas. It must be free of NMHCs to enable its injection into the downstream components of the fuel processor (H 2 S trap, reactor for the water gas shift reaction WGS, reactor for the preferential oxidation of CO Prox) and the fuel cell itself. Otherwise, the catalysts of these components would be irreversibly damaged. The ATR is operated with cold fuel and air combined with superheated steam. The steam temperature is critical to achieve complete fuel evaporation and a homogeneous mixture of fuel, steam and air in the fuel evaporation chamber of the ATR. The catalytic burner (CAB) has two functions. On the one hand, its task is to completely convert all combustible gaseous components of the anode off gas of the fuel cell (H 2, CO, CH 4 and all residual hydrocarbons) yielding no other emissions than CO 2 and H 2 O. These reactions are Period covered: 1 November 2011 31 May 2015 Page 10 of 24

supported by a particularly developed catalyst. On the other hand, it is designed to transfer the heat from the combustion reactions to a mass stream of water to produce steam for the ATR. Both components, autothermal reformer and catalytic burner, are equipped with integrated heat exchangers to use the process heat for educt evaporation und super heating. The combination of reaction heat and its temperature level from a single reactor is not sufficient to cover the energy required to evaporate and superheat the necessary amount of steam for reforming. Therefore, this task is shared by both reactors. A part of the water required is completely evaporated in the catalytic burner heat exchanger and mixed with the remaining fresh water. The resulting saturated steam is then fed into the reformer heat exchanger to be evaporated and superheated. By varying the ratio of evaporated water to fresh water, the steam production process can be controlled. During intensive experimental tests of the single components and the complete fuel processor system, both reactors ATR and CAB proved to fulfil their above mentioned functions very reliably. Steam production in the CAB was possible at all load points between 50 and 100%, while only CO 2 and H 2 O were detected as emissions. In the ATR, the concentrations of NMHCs were found to be at or under the detection limit of the laboratory gas analytical system. So, the product gas of the ATR showed a sufficient quality for its injection also into the fuel cell. The molar flow of hydrogen at maximum load was equivalent to an electric gross power of 3 kw. To reach these targets the nozzle technology for fuel injection into the ATR fuel evaporation chamber was significantly improved and an additional heat exchanger was introduced into the fuel processor system for air preheating. The figures below show the FCGEN ATR reactor (left) and the FCGEN CAB reactor (right). Autothermal reformer ATR (left) and catalytic after burner CAB (right) Downstream the ATR reactor, the H 2 S trap is arranged. It is fed with the reformate of the ATR and has the function to reduce the sulphur concentration in the reformate to less than 1 ppmv. This concentration range is acceptable for the downstream components of the fuel processor and the complete APU. Especially the catalyst at the anode side of the fuel cell is critical in this respect. The H 2 S trap was equipped with a granulated zinc oxide material for the adsorption of sulphur containing components in the reformate such as hydrogen sulphide. The experiments with the complete fuel processor system proved that the H 2 S trap worked very reliably. No deactivation of downstream catalysts due to sulphur poisoning was observed. In additional laboratory experiments it was found that structured zinc oxide sorbents had an even better capacity for sulphur removal than the granulated material applied in the trap. Downstream the H 2 S trap the WGS and Prox reactors are arranged. Both reactors together were denoted as clean up system in FCGEN. The WGS reactor has the function to reduce the concentration of CO in the reformate from approx. 10 vol% down to approx. 0.5 vol%, while that of the Prox reactor is to further reduce the CO concentration to less than 25 ppmv. This reduction is necessary to protect the catalyst surface in the anode of the fuel cell from intensive CO adsorption. The CO adsorption would strongly reduce the number of accessible catalytically active sites for the electrochemical fuel cell reaction between Period covered: 1 November 2011 31 May 2015 Page 11 of 24

H 2 and O 2 and thus severely reduce the efficiency of the fuel cell. The figures below show the WGS reactor (left) and the Prox reactor (right). Water gas shift reactor WGS (left) and Preferential oxidation reactor PrOx (right) In FCGEN, the clean up system was accomplished in two generations. The first generation clean up system was based on a conventional heat exchanger design applying established fabrication techniques such as wet chemical etching for micro channel fabrication. In the course of the project, both reactors could be redesigned in a way which allows applying metal forming techniques to manufacture the channel sheets. Intensive experimental tests of the single components and the complete fuel processor have shown that both reactors are jointly able to fulfil their functions and to reduce the concentration of CO to the required value of less than 25 ppm. The catalytic start up system is not depicted in the introductory simplified schematic drawing of the fuel processor system. Its function is to provide heat for the start up by catalytically combusting diesel fuel. In the FCGEN project, the start up system consists of a catalytic start burner CSB (see figure below), one steamer and two air heat exchangers (air HEX). In the air HEX the heat in the start burner exhausts is transferred to clean air that heats the reformer and clean up reactors as well as the catalytic burner. The steamer provides the necessary steam for the reforming process before internal steam production is possible. High efficiency start up burner with very low emissions First laboratory testing showed that the start burner system can produce the amount of heated air and steam that is necessary to allow the reformer system to start. Tests have also been performed in order to minimize emissions (CO, diesel and methane) out from the start burner. Unfortunately, during the commissioning phase of the whole fuel processor system the burner did not start although it had earlier showed repeatable start ups during testing and tuning together with the other start up components. At the end, the control board was damaged which also caused burner damage. After laboratory testing and tuning of a second new start up system proved to work properly also during fuel processor commissioning. Heating of the complete fuel processor was demonstrated. Period covered: 1 November 2011 31 May 2015 Page 12 of 24

Fuel pump Fuel flow AIR FUEL x0,75 + ATR temperature control - + + TT REF TREF TT REF TT WGS Temperature (gradient) control TT WGS Prox Temperature (gradient) control TT PROX FT TT TT FT FT FT ATR MODEL AIR Air to carbon ratio adjustment FUEL SUB temperature control ATR SUB SUB combustion control H 2O Water injection CO control T H S TRAP 2 REF EXHAUST Partial Oxydation Air Flow control Steam Air injection AIR WGS TT PROX TT CO CO WGS PROX MODEL MODEL CO 2 CO residue estimation H O x0,25 2 H2 H2 H2 TT REF AIR CO AIR Reformate CO M ANODE CATHODE - + Cathode air flow control STOICH CONTROL ISET H flow estimation 2 TT TSTACK Non - Utilized reformate FC Stack Water injection FC MODEL Stack cooling control FT CAB CONTROL CATALYTIC BURNER STEAM GENERATOR TT TT + + H 2 O - FUEL Steam temperature control TSTEAM Steam to Carbon Ratio Control FUEL EXHAUST P IFC DC/CD Power request POWER CONTROL (Fuel set) Fuel Cell Based Power Generation The final FCGEN APU with fuel cell module on the left and fuel processor on the right The figure above shows in its right part the complete fuel processor system. The fuel processor demonstration phase was divided into two steps. At the beginning of the first phase, it was found that the circuit board for controlling the CSB was malfunctioning. Unfortunately, it was not possible to obtain a stable running CSB, and it was decided to replace the CSB with an electrical heater. High NMHC values were obtained in the first phase as a consequence of damaging the ATR in an early experiment where the reactor was filled with diesel. Before that accident, the NMHC values were low. As a second consequence, the CO concentration at the outlet of the Prox reactor was around 50 ppm, which was above the target of less than 25 ppm. It was decided to replace the ATR for the second demonstration phase. This second phase demonstrated successful operation of the fuel processor module delivering a reformate with PEFC quality as already observed with the ATR, when this reactor was operated in stand alone mode. I could be stated at the end of the second demonstration phase that the FCGEN fuel processor system completely fulfilled its task of supplying a hydrogen rich gas flow at adequate quality to the anode side of the fuel cell. WP4: Control system, Electrical interface and power conditioning The objective of WP4 is to develop the ECU with the control system which is capable of efficient on line control of the APU system and its subsystems and develop the APU power conditioning, including the 5 kw DC/DC converter and BoP power supply board. The work has been split into 5 work tasks. Control system (CS) specification: JSI, leading the WP4, has taken the lead to plan the system control strategy already during the fuel processor development phase and strong input was given to the FP design. In following the conceptual design was defined based on a CS requirement analysis and the necessary functions of the CS were determined and the preliminary set of input and output data was defined. Thereafter, a preliminary design was developed based on further information from project partners while the APU matured. Finally this resulted in detailed CS design document (specifications). Air Air Flow Control Air bleed Battery LOADS Non- utilized H 2 UFC ISET PSET UBAT ICON IBAT "Jožef Stefan" Institute Ljubljana Slovenia High level control automaton (left) and main control scheme (right) Period covered: 1 November 2011 31 May 2015 Page 13 of 24

CS development and testing: According to these specifications, JSI has carried out the development of the complete control software comprising function block library with 57 function blocks, complete program with 230 data blocks and prepared final CS documentation. To give even more time to the APU process evolution, the CS has been deployed in two stages. In the first stage it was developed using industrial controller, allowing high degree of flexibility. After the APU and CS reached certain degree of maturity, the complete CS program has then been programmed to C version, which was installed on the APU ECU and used for complete APU testing and demonstration. High level control automaton (left) and main control scheme (right) The APU process with over 60 sensors and 40 actuators represent a small scale factory, packed within few hundred litres space. To successfully handle such complexity, the control system design approach requires modularity, hierarchy standardization and documentation. To cope with these requirements, on the software side there is a multilevel level hierarchic CS structure with Control automaton describing main procedural states on the highest level (4), Control function blocks, (20) PID controllers and diagnostic functions on level (3), Interfaces to sensors and actuators on level (2), and I/O data processing & communication on basic level (1). Documentation and reference library of the FCGEN ECU control program Along with the control system, a comprehensive HMI program with close to 50 screens has been developed. This HMI became a necessary tool for commissioning of the components, of the FPM, and later for optimization of the complete APU. Its functionality has been further extended by combining it with a remote access computer, perfectly serving the dislocated testing sites for shared experiment sessions and operation / performance / issue discussions, joined by project partners from across Europe. Period covered: 1 November 2011 31 May 2015 Page 14 of 24

Fuel Cell Based Power Generation HMI screens for FP, FC modules and for sensors On the hardware level the CS consists of the close to automotive grade APU ECU unit with all required I/Os, communication channels and processing power. It offers the following connections: inputs: 10 digital multipurpose, 16 analog, 5 MFM, 1 lambda, 1 current, 30 thermocouple & 16 Pt 1000 temperature, all with required excitation outputs outputs: 18 high/medium current with pulse frequency capability, 8 CMOS with PWM cap., 8 tacho, 16 analog (6 with ground sense capability) communication: 2x can, 1x Ethernet, 2x programming FCGEN Electronic control unit ECU (left) and power distribution board for BoP power supply (right) Industrial controller Custom designed ECU size: 100 x 40 x 16 cm (64 l) weight:6 kg without power supplies price: very expensive Period covered: 1 November 2011 31 May 2015 size: 45 x 16 x 4 cm (< 3 l) weight: 0.7 kg, price: reasonable in medium production volume Page 15 of 24

Furthermore the power distribution board has been developed to provide power for the sensors and actuators (BoP). It provides separate ECU controlled enabling of 6 power levels (ECU, Sensors, Actuators 1, 2 and 3) to minimize losses when not all levels need to be engaged. It also features pre charge functions to protect the battery from high inrush currents and load dump protection to protect ECU from electric shocks, caused by eventual battery or BoP failures. Final FCGEN control system concept (left) integrated ECU (middle, right) Vehicle interface: CRF, supported by JSI, carried out an assignment to define all functional connections necessary to manage the APU and to make the driver able to operate the APU system, resulting in simple cabin HMI. The working team highlighted the importance of completely disconnected APU from the conventional electric layout during the testing phase to avoid any negative influence of this system on the conventional battery system in the truck. This issue is going to be solved by connecting the APU to a secondary battery system, and providing way to restoring back to single battery after successful proof test. Vehicle interface scheme Hardware in Loop (HIL) testing: To test the CS before application to the real process, the hardware in loop test methodology was used. There the process behaviour is mimicked by the model and control system & unit functionality is verified. The PowerCell provided and installed the HIL test rig. For enabling the required real time model execution the simplified version of the FP process model was developed by Modelon. Jointly the model and CS have been tuned to the desired degree. To cope with dislocated partners the TeamViewer platform was used to access the computers and controller remotely. The figure below shows the complete HIL testing setup. Period covered: 1 November 2011 31 May 2015 Page 16 of 24

PC at Powercell (supplied by JSI, with runing Simatic & HMI) Ethernet PLC (APU Control System) CAN com. Real Target machine (runs real-time model) Ethernet Internet Ethernet PC at JSI (remote access from Slovenia) Host computer (Dymola/Simulink plant model) HIL testing assembly scheme Power conditioning: The FCGEN DC/DC power converter was designed especially for the operation with Fuel Cell power systems. It is a 5 kw buck converter that consists of a step down (buck) power stages. It has nominal input and output voltage ranges of 32 75V and 20 30 V, respectively with the conversion frequency ~100 khz and efficiency of 96%. The 5kW output power is achieved by paralleling three basic modules: 80 A + 40 A + 80 A. A control of each basic module is realized by a microcontroller, which is an ARM core based device. Its main features are: inherent short circuit protection, adjustable input and output current or voltage and voltage smoothing on the input and output side. The DCDC converter has been finalized and tested in spring 2014. FCGEN DCDC converter unit (left) and complete APU power conditioning system (right) Cost analysis A cost analysis has been performed based on the final design of the FCGEN system. The analyses was made for different yearly production volume, starting at producing one prototype up to 100 000 units/ year. The partners responsible for developed component have made the analysis for it s respective components: Johnson Matthey: Catalyst for ATR, WGS, PROX, the Catalytic burner catalyst, and sorbent for DS. Forschungszentrum Juelich: ATR reactor, Catalytic burner reactor and fuel nozzle. IMM: PROX reactor, WGS reactor and desulfurization reactor. JSI: DC/DC converter, ECU, Power distribution, Fuse and relays. Powercell: Fuel cell, BoP component, sensors, frame, box and assembly. From the charts below can be seen that in prototype stage the reactors, including the catalyst, in the FPM is the dominating part counting for more than 50% of the total cost. However in higher volume this cost is dramatically reduced, and for high volume production it is in the same order as the fuel cell stack cost. Period covered: 1 November 2011 31 May 2015 Page 17 of 24

257 473 15 337 23% 2% 18% 11% 15% 3% 3% 4% 52% 20% 5% 19% 8% 6% 8% 3% HEXs Reactors Fuel cell stack Actuatos sensors Radiator, filter, etc Cont sys & Electrics APU, material & assambly HEXs Reactors Fuel cell stack Actuatos sensors Radiator, filter, etc Cont sys & Electrics APU, material & assambly Cost breakdown for prototype (left) and for production volume 1 000 units /year (right) 9 944 5 371 16% 13% 18% 18% 16% 10% 24% 16% 7% 7% 3% 14% 14% 8% 4% 12% HEXs Reactors Fuel cell stack Actuatos sensors Radiator, filter, etc Cont sys & Electrics APU, material & assambly HEXs Reactors Fuel cell stack Actuatos sensors Radiator, filter, etc Cont sys & Electrics APU, material & assambly Cost breakdown for production volume 10 000 units/year (left) and 100 000 units /year (right) The cost of the prototype build in the project is very high, 257 473. This is due to the component developed in the project are hand build in prototype workshops, but also several of the BoP components are still on a prototype stage and very expensive. When the APU is series produced the cost will rapidly reduce: when produced in 1000 units/year the cost is down to 15 337. When produced in 10 000 units per year the cost is down to 9 944, which in level with the target in Multi Annual Work Plan for year 2020, see table below. At a highest volume the cost is estimated to 5235, which is only 16% higher than the target for 2023 in the Multi Annual Work Plan. Table 1. Cost target for APU for truck application (3 kw), FCH JU Multi Annual Work Plan 2014 2020 2012 2017 2020 2023 Cost ( /kw) > 10 000 < 5000 < 3000 < 1500 Cost 3 kw unit ( ) > 30 000 < 15000 < 9000 < 4500 When comparing the values is should be remembered that the cost estimation in the FCGEN is based on the current design. It includes many components that can be avoided, replaced and combined. Several of the sensors in the system is only for analysing the system and not needed in a series product. Some of the component has been added to secure the controllability of the system, when the behaviour of the system is better known these components can probably be removed. Period covered: 1 November 2011 31 May 2015 Page 18 of 24

The cost does however not include any scrap of depreciation on tools for the APU assembly. For components from external suppliers some of the cost estimates are from the supplier and other has been estimated based on cost engineering principals. Conclusion As expected for the development projects, also FCGEN faced numerous challenges, technical and management ones. The first demonstrated in the difference of component behaviour compared to manufacturer characteristics when interacting with each other, or compared to their models due to prototype components provided by suppliers as well as hidden details related to operation in new environment. The management challenges mostly sourced from the pursuit for solutions of the complex technical issues. The technical development work in the project has been carried out jointly and with the involvement of all participating partners. The challenges have been discussed in frequent design team meetings, where valuable inputs to the issues and possible solutions were collected. This is stressed here, as only this way achieving final result was possible. Moreover, this way many important lessons were learned, that can be sumarized in three lines: reformer/fc tecnology specific, required for the emerging tecnology to successfully hit the market, partner specific, enhancing and broadening area of expertize, improving workflow, etc., an invaluable expeirience of remote experimentation with dislocated partners being able to attend testing in real time and share their knowledge & experience in the light of project progress By strict management, the work ethics of bringing the problems out, and strong incentives to find sollutions and finish the project successfully, all problems were overcome in the best possible way, albeit some with compromises (6 months project time extension and somewhat compromised weight, size and efficiency of the prototype). In the light of encountered delays, the truck demonstrating partner (CRF) leaving project, the consortium has undertaken substantial efforts and achieved main goal with one of the first demonstrations of autonomous PEM FC based APU operation and electricity production using commercially available diesel fuel. Finally, the ATR reformer technolgy lived up to high expectations concerning operability and efficiency. Despite its complexity, with certain technological solutuins stil left to be improved, there is significant potential to achieve and surpass future MAIP goals. During numerous testing sessions, the necessary measures to significantly improve the following parameters were identified for future work: bring the efficiency above 35%, further reduce weight and volume, radically shorten start up time, and improve durability and serviceability. Period covered: 1 November 2011 31 May 2015 Page 19 of 24

Potential impact, main dissemination activities and exploitation Project potential impact The mission of the FCGEN project was to move the FC based APU systems a major step towards industrialization. The project was targeting several issues which make this possible among others: system cost, improved system design for better performance, better efficiency and durability, reduced system size and weight. Efficient, durable and cost effective FC based APU systems provides clean electricity (less CO2 and extremely low NOx and HC emissions) and less noise to the driver cabin during stand still conditions compared to the condition when electricity is generated by engine idling. The FCGEN APU system provides around 80% fuel saving when it is used as electrify source to the truck cabinet under stop phases compared to electricity provided via Internal Combustion Engine (ICE) idling. The reduced consumption results also in same level reduction of CO2 emissions. Compared to diesel driven APUs a 40% consumption reduction can be expected. The reported idling time of trucks in US in is in the range of 1500 to 2500 h per year, results in fuel costs of 4000 7500 per year. Even higher value confirmed also by internal project study addressing EU situation makes opportunity for a good business case. Within the FCGEN project, cheaper fabrication techniques have been developed, such as embossing, to reduce the production costs of a future micro channel coated heat exchanger reactors compared to the currently employed fabrication methods. Further cost reduction is expected via the reduction of precious metal loadings in the fuel processor catalysts and the selection of some cost effective BoP components. The in project undertaken cost analysis has shown that in high volume production the cost of current design is estimated to 5235, only 16% higher than the target for 2023. Moreover, the current design includes many components that can be avoided, replaced and combined. Several of the sensors in the system are only for analysing the system and not needed in a series product. Some of the component has been added to secure the controllability of the system, when the behaviour of the system is better known these components can probably be removed. With new generation designs costs can be further reduced. Due to the advanced catalytic technologies used in the FCGEN for system heating at start up, fuel reforming, reformate purification and anode off gas combustion, the level of emissions NOx, non methane hydrocarbon, CO and SOx are < 1ppm, which are significantly lower than the corresponding emissions produced during ICE idling. As the FCGEN is one of the pioneering projects which focuses on integration of FC based APU systems targeted for on board a vehicle and tested under close to real conditions with logistic fuels, the project provides valuable data and findings to vehicle OEMs with respect to various aspects of APU integration on board the vehicle: mechanical, electrical, communication, safety. This is a powerful step for the commercialization of FC based APU systems for on board power generation and may open the path for additional utilization areas for these systems (e.g. electrifications of auxiliaries, H2 supply for other application, etc.). Project partners have already detected commercial interest for developed APU from OEMs in various transport fields, namely maritime and road recreational vehicles, and are pursuing ways to jointly further increase the TRL level with gained experience: to reduce cost, significantly increase the efficiency, improve reliability, ensure serviceability and durability. Furthermore, with the achieved development stage of technology, gained experience and noted possible improvements, also other mobile as well as stationary application fields become attractive. Period covered: 1 November 2011 31 May 2015 Page 20 of 24