POWERCELLS FUEL PROCESSING DEVELOPMENT AND ITS APPLICATION IN APU SYSTEM WITH A PEM FUEL CELL Per Ekdunge, Ida Toftefors, Lisa Kylhammar and Jonas Nordström PowerCell Sweden AB, Gothenburg, SWEDEN Introduction Today, almost all commercially available primary fuels are based on hydrocarbons, mainly from fossil fuels. Fuel cost and supply as well as emission abatement are in the focus of the transportation industry of today. The debate about climate change has emphasized the importance of these issues and increased the search for sustainable technical solutions. The transport sector is often considered the most difficult energy sector when it comes to sustainable development. The number of long hauling trucks used for transportation of goods is rapidly increasing (1). Hence, increases in diesel truck hazardous emissions are expected and huge effort is spent to improve and develop technology for exhaust emission control, in order reduce these emissions (2). These systems are primarily designed and optimized to function when the truck is running, i.e. in the driving mode. Drivers idle engines to power climate control devices (e.g. heaters and air conditioners) and sleeper compartment accessories (e.g. refrigerators, microwave ovens, and televisions) and to avoid start-up problems in cold weather. Idling increases air pollution and energy use as well as wear and tear on engines. A large number of heavy-duty trucks idle a significant amount and several studies have shown that heavy-duty long-haul engines idle about 20-40 % of the total operating time, resulting in a significant increase in HC, CO, NO x and particle emissions (3,4,5). Idling by heavy-duty diesel truck engines could be avoided by the use of a fuel cell based auxiliary power unit (APU). Fuel cells have many attractions as APUs. Not only do they provide the potential to reduce pollution, energy use, and greenhouse gases, they also provide the potential to reduce costs, and increase driver comfort. However, in order to make use of the existing infrastructure, a system for converting a commercially available fuel into hydrogen, which is the fuel for the fuel cell, must be developed. Production of hydrogen from hydrocarbons has been operational in the oil and gas industry for more than a century and the continued development of small scale fuel processors for fuel cell applications has received significant attention in recent years. The bulk of these activities have however focused on simple singular fuels such as alcohols and methane. Diesel is a complex mixture of hydrocarbons, with a wide boiling point range, making both the evaporation and reformation of the fuel a difficult challenge. Development of reformer technology and fuel cell systems for automotive applications started for more than a decade ago within the Volvo Group. PowerCell Sweden AB was spun-out from AB Volvo in 2008 to continue develop and commercialize fuel cell power systems, diesel-fuel reformers and reformate tolerant LT-PEM fuel cells. This contribution aims at describing the demands put on the reformer system by the fuel cell and, in the end, the APU application and the end user. Additionally, a description on how this has impacted the design of the components in the reformer system and some results from our testing will be provided. 1
APU systems The central part of the APU system is the electrical power generation in the fuel cell, where an electrochemical reaction between air and hydrogen produces electrical energy, with water and heat as the only byproducts. For a truck APU application producing the hydrogen on-board, from the vehicle fuel is the preferred option due to the need of high energy density fuel and the lack of hydrogen infrastructure. On-board production requires efficient, compact and low-cost fuel processors especially designed for the application. The complexity of the fuel cell system depends on both the fuel and the type of fuel cell (Table 1). The fuel cell technologies differ in both operation temperature and electrolyte composition. The choice of fuel cell type depends on the application and operating conditions. With diesel reformate as fuel, different fuel cell types can be considered such as the Phosphoric Acid Fuel Cell (PAFC), the Molten Carbonate Fuel Cell (MCFC), the Solid Oxide Fuel Cell (SOFC) or the Proton Exchange Membrane Fuel Cell (PEMFC). Table 1. Comparison of Fuel Cell Technologies. Tolerance to poison Fuel cell class Operating Robustness Startup Power Dynamics CO NMHC 1 S temperature time density Molten Carbonate 600-1000 C Low Low Low Low Med High Med (MCFC) Proton Exchange 50-100 C High High High High Low Low Low Membrane (PEMFC) HT-Polymer 150-200 C Med Med Med Med High Med Med Electrolyte (HT-PEFC) Phosphoric acid 175-220 C Med Med Low Med. Med Med Med (PAFC) Solid Oxide (SOFC) 600-1000 C Low Low Med Low High High Med 1 Non-methane hydrocarbons. In applications where frequent starts and stops occur, fast start-up times are needed and where large dynamical load changes are required the PEMFC is the most suitable choice. When employed on a truck, the fuel cell also needs to withstand severe environmental conditions such as chock, vibration and a broad temperature window, areas where the PEMFC is superior over alternative fuel cell technologies. The PEMFC put, however, the highest demand on the fuel processor. Due to the low operating temperature of PEMFC its catalyst is prone to poisoning by impurities in the reformate. The demand on reformate quality is therefore highest with the PEMFC. CO is adsorbed on the fuel cell catalyst and a clean-up system is therefore needed between the reformer and the fuel cell stack in which the CO concentration is reduced from 5-10 % to <25 ppm. Fuel reforming There are three main methods for reforming fuels to a hydrogen rich gas; Steam Reforming (SR), Catalytic Partial Oxidation (CPO), and a combination of the two, Auto-Thermal Reforming (ATR) with the following general formulas (14): SR: C n H m + n H 2 O n CO + (n+m/2) H 2 (1) CPO: C n H m + ½ n O 2 n CO + m/2 H 2 (2) 2
ATR: C n H m + n/2 λ O 2 + n (1- λ) H 2 O n CO + ((m/2) λ + (n+m/2)*(1- λ)) H 2 (3) The choice of most suitable reforming technology depends on the application requirements. The diesel fuel processor developed by PowerCell is based upon the auto-thermal reforming concept for the reforming of the hydrocarbon fuel and both the water gas shift (WGS) and preferential oxidation (PrO X ) reactions (14) for purifying the gas to a quality that is compatible with a PEM fuel cell. The choice of fuel The fuel for the APU should preferably be the same as that used for the main engine of the vehicle. Since the predominant fuel for heavy duty trucks is diesel, the fuel of choice for the APU is standard road diesel. Diesel fuel has a broad variety of characteristics and various specifications are used in different countries, such as SS 155435 in Sweden, EN 590 in Europe and ASTM D975 in USA. An extract of these specifications is shown in Table 2. At PowerCell, diesel specification SD10 is used, which is a diesel quality without seasonal variations, but otherwise comparable to standard road diesel. Table 2. Diesel fuel properties (6,15). Fuel Property Unit SS 155435 EN 590 ASTM D975 SD10 Min Max Min Max Min Max Min Max Density @ 15 Kg/m 3 800 830 820.0 845.0 - - 835 837 C Sulfur µg/g - 10-10 - 15-10 Aromatics Vol% - 5 - - - 35 Typical: 18 PAH (tri+ aromatics =>) - Vol% - 0.02 - - - PAH (di+ Vol% - - - 8.0-11.0 2 6 aromatics =>) Ash Wt% - 0.01-0.01-0.01 0.01 Flash point C, - 55-55 - 52 55 - Carbon residue Wt% - 0.3-0.3-0.35-0.2 (on 10 % distillation residue) Kinematic viscosity cst 1.40 4.00 2,00 4,50 1.9 4.4 2.3 3.3 @ 40 C Cetan number 50 51 40 52 54 Distillation C IBP 180 - - - - - - - 90% 282 338 - - 95% - 320-360 345 350 The properties are specified for the requirements of the internal combustion engine, but have also certain relevance for fuel reforming. The sulfur content makes it necessary to develop a sulfur tolerant catalyst for the reformer and to use a sulfur trap after the reformer to protect the downstream catalysts. The aromatics in the fuel increase the challenge and the ability to get 100 % conversion of the hydrocarbons. The carbon residue is a measurement of the coking tendency of the fuel and the cetan number its ability to ignite. 3
The fuel also contains different additives such as cetan number improver, antifoam additives, de-icing additives, low-temperature additives and fuel stability additives. The fuel quality and additives has local as well as seasonally variations. The fuel will in the future also contain more biobased components in order to increase the biofuel use. From 2010, diesel fuel in Europe may contain up to 7 vol% fatty acid methyl ester (FAME) to meet biofuel directives. The fuel processor needs to be able to handle all these variations and deliver a reformate to the fuel cell with no NMHC and a CO level below 25 ppm in order to be used with a PEM fuel cell. Experimental Gas emission data were collected with a FTIR spectrometer from MKS (MKS MultiGas Analyzer model 2030) using a liquid nitrogen cooled MCT detector. The spectra were analyzed with the MG2000 software provided by MKS. The data were fitted and compared to gas calibrations developed by MKS, enabling calculation of gas concentrations of many reaction products commonly encountered when reforming diesel. APU design challenges on the reformer system The APU system combines the fuel cell, the reformer and the balance of plant components, which are needed for the heat, air and water management to secure that the reformer and fuel cell have the right working conditions. Figure 1 shows a schematic of the APU system designed to convert diesel to electric power. The reforming process occurs in the ATR reactor where hydrogen, carbon dioxide as well as methane and carbon monoxide are produced. The reformate stream is then transferred to the clean-up reactors. Figure 1. Schematic of the APU system showing the most important components. ATR reactor development The development of the diesel reformer is very challenging, especially considering the complex and poorly defined composition of the fuel. The prime challenge in designing a diesel reformer is creating a perfect mixture of fuel and oxidant before it reaches the catalysts while preventing autoignition of the mixture. In recent years, several concepts for diesel reforming have been proposed to resolve this problem (7-13). Moreover, high efficiency is required in combination with fast startup time, dynamic operation and compactness as well as durability and low cost; and tradeoffs must be done in order to achieve a suitable mix of these components. 4
To achieve high overall system efficiency the ATR s hydrogen production needs to follow the electric load, at the highest possible rate. This means that the ATR continuously operates in quick transients, and the highest possible thermal load change rate is limited by the water evaporation system, which means that there is a tradeoff between efficiency and durability of the ATR since steam is essential to suppress coke formation. In addition, full conversion of the hydrocarbons is essential because hydrocarbons damage the fuel cell stack and the clean-up system. The conversion is enhanced by pressure and temperature, however, a high temperature lowers the efficiency of the ATR and high pressures result in a lowered overall efficiency due to the increased electric power consumption by the balance of plant components in such a system. To achieve full conversion of the fuel and extend the life-time of the catalyst it is important to create a homogenous mixture of the reactants. The mixing process is enhanced by residence time in the mix zone, however, spontaneous ignition in the mix zone can be a problem when overall or local residence time exceeds the reaction delay. Finally, it must also be possible to manufacture the reactor at reasonable cost in available materials that can tolerate the harsh conditions. PowerCell has developed six generations of auto-thermal reformers, where all the issues described above has been taken into account and tradeoffs has been made to gradually improve both the reformer and system characteristics. Table 3 summarizes the performance and operation conditions for the different generations. Table 3. ATR reformer generations at PowerCell. Component G1 G2 G3 G3b G4 G5 G6 1 Hydro carbon slip >1000 <500 <15 <10 <5 <5 <5 (NMHC) (ppm) Operation 925 900 750 750 770 770 800 temperature ( C) Efficiency (%) 75 % 77 % 80 % 80 % 85% 83% 83% Steam-to -carbon 2.5 3 3 3 3.5 3.5 2.8 lambda 0.3 0.34 0.3 0.3 0.27 0.29 0.29 Thermal loads (kw) 8-17 8-17 8-17 8-17 6-12 4.5-12 3-12 1 G6 is currently under development and values are calculated. The main target through generation 1 (G1) to generation 3 (G3) was to ensure a stable operation and full conversion of non-methane hydrocarbons at an acceptable turn down-ratio, i.e. the ratio between the highest and lowest possible thermal loads. Generation 4 aimed to maximize efficiency and turn-down ratio, but problems with spontaneous ignition occurred at low thermal loads. In G5 the efficiency of the ATR was lowered, but the efficiency of the complete APU system was increased and the ATR operation gained stability. The ATR G5 was integrated in a complete, stand alone, self-controlling APU system and proved that it could manage to follow the electric load cycle without getting into unstable operating conditions. G6 is currently under development, and the main target with this generation is to increase turn-down ratio, increase overall APU system efficiency, reduce reactor size and increase manufacturing potential. Clean-up development First in the clean-up stage is the DS, or desulfurization reactor, where the sulphur-containing substances are absorbed, often using a ZnO material. To reduce the CO content in reformate stream 5
there are two water gas shift reactors, the first working at high temperature and the other one working at slightly lower temperature. These two reactors convert carbon monoxide and steam to carbon dioxide and hydrogen gas. However, the concentration of carbon monoxide is in general still too high after the water gas shift reactors and the reformate cannot be sent directly to the fuel cell stack. Instead, to decrease the CO concentration even further, there are two reactors where preferential oxidation is performed. Here, the last of the CO is reacted with O 2 to produce CO 2. Throughout the clean-up system there is a number of heat exchangers to control the temperature of the reformate stream as well as providing heat for steam generation to the reforming process, see the schematic in figure 1. After-burner development After the clean-up procedure the reformate is feed to the fuel cell stack where the electrical power is produced. Exhaust gas from the anode side of the fuel cell is lead to an after-burner reactor where it is burnt to produce steam for the reformate process. The amount of steam produced is directly affecting system efficiency because energy is used in this process. However, a certain steam level is needed to ensure CO conversion in the clean-up and to avoid coke formation in the ATR. To be able to follow quick load changes and a shifting steam demand it is important to design a steam/water system with low water accumulation as well as a with a low weight so that steam production can be ramped up and down as fast as possible since this is the limiting factor for thermal load change rates in the PowerCell APU. Results and discussion The fuel processor system has been extensively tested before being integrated into the APU system to secure proper operation and reformate composition. Gas concentrations out from the ATR and clean-up system are shown for different power levels in figure 2. Depending on thermal load the ratio between different gases is somewhat different. Between 26 mole% and 30 mole% of the mixture is hydrogen gas. The methane concentration is increased at higher thermal load, from about 200 ppm at 6.0 kw to around 700 ppm at the highest load, 12.0 kw. The carbon monoxide concentration is well below the tolerable level for the PEM fuel cell developed by PowerCell. The non-methane hydrocarbons concentrations from the reformer system are very low and below the detection limit of our instruments (below 5 ppm). 6
Figure 2. Gas concentrations after the reformer and clean-up system at different loads when running at steady state. In the truck APU application the fuel processor system must be able to be load following and give an acceptable gas quality also under transient operations. Figure 3 shows the CH 4, CO and NMHC concentration when the reformer system is running in steady state and during step changes in thermal load, between 6 and 12 kw. The results show that the CO level is below 10 ppm also under transient operation and that the NMHC remain under the detection limit (below 5 ppm). Figure 3. Concentrations of CH 4, CO and NMHC after the reformer system at load changes. To start the system the reformer needs to be heated to about 400 C in order to start the ATR reaction. The reactors in the clean-up system also need to be heated up to their operation temperature in order to have the right gas quality for the fuel cell. To accomplish this, a startup system was developed based on a catalytic burner. The heat is produced by burning diesel in the catalytic burner and transferred to the reformer system. The resulting temperature increase in the reactors is shown in figure 4. It takes about 30 minutes until the last reactor, the second preferential oxidation reactor, has reached its operation temperature. Further optimization of this system is needed before our required start time is reached. 7
Figure 4. Temperatures in the fuel processor system during heating up. Before the reformate can be feed to the fuel cell it must be secured that the correct gas quality is established in order to not poison the fuel cell. The reformate gas quality during start-up is shown in figure 5. As indicated in the figure the fuel cell is bypassed during the first 10 minutes to protect the fuel cell. This time can possibly be shortened without endangering the fuel cell stack. Figure 5. Reformate gas quality during start-up. Complete APU system The fuel processor and fuel cell stack were integrated in a complete APU system, which has 8
been developed to demonstrate a compact packed APU that is autonomous and has good load following capability. Figure 6 shows electric power supplied by the APU, and the thermal power in the fuel consumed, for a typical load cycle during a demonstration at PowerCell. The electrical load was typical electric equipment used in a truck during stand still, such as air conditioners, refrigerators, microwave ovens and coffee machines, which were randomly turned on and off by visitors. A standard truck battery was connected in parallel to the APU and was used as a buffer if high loads were applied. The results show a very good load following capability of the APU which automatically adopted to the power demand of the loads and only small amounts of energy was drawn from the battery during fast transients while the battery was charged during quick power drops, which resulted in a close to constant state of charge, see right panel in figure 6. Figure 6. Typical load cycle with a base load of an air conditioner and a refrigerator. A microwave and two coffee machines are turned on and off by visitors at a demonstration at PowerCell. State of charge (SOC) is controlled to a constant value. The efficiency of the APU is shown figure 7. The net efficiency, diesel to electric power out from the fuel cell, reaches 29 % at 10 kw thermal load, whereas the total efficiency, including energy consumption by internal balance of plant components and losses in the power electronics, varies between 19 % and 23.5 %. In the next generation of APU, which is under development, the efficiency will be increased to 30 %, mainly by reducing the hydrogen needed for steam generation. 9
Figure 7. Efficiency and losses for the APU system at different loads. To achieve high overall efficiency it is important to minimize the energy consumption of the balance plant components and create a more efficient steam generation system. Conclusions Auto-thermal diesel reforming combined with PEM fuel cell has shown to be suitable technologies for APU application. The ATR reformer produce reformate of quality suitable for PEMFC both during steady state and transient operation. For the APU application reformer cannot only be optimized for efficiency and durability, several other parameters need to be taken into account the make an efficient and stable APU system. By combining the ATR with a PEM fuel cell an efficient and load following APU system can be developed, which is able to make frequent starts and stops. References 1. M. Contestabile, Energy Policy 38(2010) 5320-5334 2. R. M. Heck, R.J. Farrauto, S-T. Gulati, Catalytic air pollution control: commercial technology, 3rd ed,, Wiley, Hoboken, N-J., 2009 3. P. Agnolucci, Int, J. Hydrogen Energy, 32(2007), 4306-4318. 4. C. J. Brodrick, T. E. Lipman, M. Farshchi, N. P. Lutsey a, H. A. Dwyer, D. Sperling, S. W. Gouse, D. B Harris, F. G. King Jr, Transportation Research Part D 7 (2002) 303 315 5. N. P. Lutsey, C. J. Brodrick, T. E. Lipman, Energy32 (2007) 2428-2438 6. D. Danielsson, L. Erlandsson Comparing Exhaust Emission From Heavy Duty Diesel Engines Using EN 590 VS. MK1 Diesel Fuel, AVL MTC 0015, October 2010 7. Krummenacher, J. J. and Schmidt, L. D., J. Catal. 222(2004) 429 438. 8. Krummenacher, J. J., West, K. N. and Schmidt, L. D., J. Catal. 215(2003) 332 343. 9. Shekhawat, D., Gardner, T. H., Berry, D. A., Salazar, M., Haynes, D. J. and Spivey, J. J., Appl.Catal. A Gen. 311(2006) 8 16. 10
10. Subramanian, R., Panuccio, G. J., Krummenacher, J. J., Lee, I. C. and Schmidt, L. D., Chem.Eng. Sci. 59(2004) 5501 5507. 11. Thormann, J., Pfeifer, P., Schubert, K.,Kunz, U., Chem. Eng. J. 135(2008) 74 81. 12. Thormann, J., Maier, L., Pfeifer, P., Kunz, U., Deutschmann, O., Schubert, K., Int. J. Hydrogen Energy 34(2009) 5108 5120. 13. Hartmann, M., Kaltschmitt, T. and Deutschmann, O., Catal. Today 147(2009) 204 209. 14. Kolb, G (2008), Fuel processing for fuel cells, Wiley-VCH. 15. Product Specification Diesel SD10, version 5, Preem AB. 11