Comparison and Analysis of kW-Class SOFC-CHP Systems with Different Fuels

SOFC is an ideal device for developing distributed combined heat and power (CHP) systems. In this paper, kW-class SOFC system models based on semi-empirical SOFC stack model were established to optimize the systems with the aim of maximizing the system electrical eﬀiciency. Four types of SOFC-CHP systems using different fuel including natural gas (NG), hydrogen, methanol, syngas were compared, and the relationships between key parameters and system energy eﬀiciency were analyzed. Simlulation results show that decreasing the air inlet temperature (AIT) and increasing stack temperature will promote electrical eﬀiciency most significantly. In addition, the operational costs of four different systems were also compared, and among which the methanol-fueled system was the lowest.


Introduction
The distributed system combined heat and power (CHP) is a user-oriented energy consumption method, which directly provides users with various kinds of energy, can effectively improve energy efficiency by minimizing loss during transport and making full use of the waste heat. Solid oxide fuel cell (SOFC) can be used in a distributed energy system to improve energy efficiency effectively. Currently, SOFC-based distributed CHP system (SOFC-CHP) is mainly adpoted by domestic consumers, which ranges from hundreds watts to several kilowatts. The commercial Type-S from Ene-farm in Japan is a SOFC-CHP system used for households with a rated power of 700 W [1]. SOFC can run on different fuels, hydrogen, hydrocarbon fuels such as natural gas and methanol. Although the hydrogen-fueled SOFC-CHP system may achieve higher performance and longer lifetime than hydrocarbon fuels systems, in remote areas, liquid methanol fuel is a good alternative because it is much easier to store and transport. The syngas from the water-gas reaction can also be used by SOFC directly, which is an clean and efficiency approach to gain power from coal. Since lots of cycles can be applied for SOFC-CHP system to achieve different purpose and performance. For this reason, system simulation is a useful tool to investigate the influence of key operating parameters on different systems. In this work, several system configurations with different fuels will be selected for comparison and sensitivity analysis of critical parameters on energy efficiency will be presented. Based on the above results, parameter optimization and economic analysis will also be investigate.

Configuration 1: NG-fueled System
There are three material flows in the NG-fueled system, as shown in Fig. 1. Natural gas is firstly reformed and then flows into the anode. Water is divided into two streams, one of them supplies hot water and the other is used for steam reforming. An air compressor is used to provide preheated air to the cathode. Hydrogen separation is adopted to improve stack performance and the H 2rich gas separated from reformed gas flows into the anode. The remaining gas is combustible and burned with cathode off-gas to provide heat at the same time. In this configuration, the anode off-gas are all recirculated back to the reformer.

Configuration 2: hydrogen-fueled System
As shown in Fig. 2, a hydrogen-fueled system with two serie-connected SOFC stacks is used to promote efficiency. The number of Cells in 2nd-stage is the half of 1st-stage. Assumming that the stack temperature and fuel utilization of these two stacks are the same. The anode off-gas of the first stack is firstly cooled down, and then further supplied to the second stack. The anode off-gas of the second stack is burned to provide heat. The cathode inlet of the second stack is the cathode off-gas of the first stage mixed with fresh air.

Configuration 3: methanol-fueled System
The methanol-fueled system is shown in Fig. 3. In this configuration, methanol is passed into stack after steam reforming, and which is similar to the NGfueled system. The main difference between these two systems is the different carbon content in methane and methanol. The latter one needs less steam because of lower carbon content, and steam content in recirculated off-gas is enough for the methanol reforming, which can be a great advantage when the system simplification is considered.

Configuration 4: syngas-fueled System
Since solid fuels like coal can be transfered into feasible fuel for SOFC after water gas shift reaction, syngas-fueled system is also studied in this work. As shown in Fig. 4, it has a simpler system configuration compared to NG-fueled system. Syngas containing carbon monoxide and hydrogen directly enters the anode after preheating, and off-gas from both anode and cathode is burned in combuster to provide heat for the system.  In this paper, a semi-empirical SOFC stack model was established based on the previous work [2]. Four correlation formulas for data fitting was summarized and four constants were adjusted according to the experiment data of a 3cell stack experiment. The results were listed in Table 1. The validation of the semi-empirical model is validated by comparing to the experiment data, as shown in Fig. 5. The I-V-P curves calculated by the model(solid lines) is perfectly matched with the experimental result(solid dots). Fig. 6 shows parameter sensitivity of electricity efficiency and a relative increment of about 12% is realized by a variety of 50 • C in stack temperature and AIT respectively. The temperature of reformer has a big impact on the system, a decline of 50 • C can result in a reduction of 24.3%, since the maximum current is 11.6A which will lead to a shortage of hydrogen supply. The result of system configuration 2 is shown in Fig. 7. The fuel inlet temperature of the two stacks is set to 700 • C while the AIT of the 2nd stage is determined by the amount of fresh air. The electrical efficiency can be improved by increasing the stack temperature or decreasing the fuel flow and AIT of the 1st stage. As shown in Fig. 8, a decrease of 164.7% is caused by an 50 • Cincrease in AIT, which means external power is required for system normal operation. And it is mainly due to the exothermic reforming reaction of methanol, large amount of power is needed for extra air. Aside from AIT, an increase of 26.1% is achieved if the current is lower to 18A. The electrical efficiency of system configuration 4 is shown in Fig. 9, a negative value of net power is obtained at an AIT of 700 • C which is the same as methanol-fueled system. This is not only due to the heat released from electrochemical reaction but also casued by exothermic shift reaction in system configuration 4. Therefor, when raise the CO content from 50% to 70% will result in a decline of 42.3% and the maximum current is around 15.5A to generate less electricity.

Optimization and Economic Analysis
Parameter optimization is done based on the above sensitivity analysis, and the result is shown in Table 2. For each system, high electrical efficiency can be achieved at the low fuel flow rate, a current close to the optimal current, low AIT, and high stack temperature.
The operational modes fixing power based on heat was adopted in this work [3]. Assuming that electricity is purchased from grid and heat is obtained by NG combustion. As shown in Fig. 10, the lowest cost of 0.08$/h is obtained   Cost of other systems except for system configuration 2 can all achieve a decrease of about 25% compared to conventional mode. In winter time, the cost of system configuration 3 and 4 is the lowest because of more heat demand, while only a decrease of 6% is achieved by system configuration 1 with part of electricity wasted. The cost of the hydrogen-fueled system is higher than the conventional for high fuel cost. Table 3 lists some of the indexes to evaluate the performance of each system. It indicates that methanol-fueled system has relatively lower complexity and operational cost while lower H2 content results in a high command for the stack.  For the NG-fueled system, although the system with gas separation is the most complicated, a good performance is achieved by pure H 2 supplied to the anode. Thus a low cost can be obtained when electricity is mainly required since the highest electricity efficiency and the well operating condition is realized in this system. In syngas-fueled system, a system that uses syngas as a fuel, the operational cost is low while it is applied to users with a larger amount of heat.

Conclusions
The highest electrical efficiency of 53.4% among the four CHP systems was achieved in the system configuration 1. The methanol-fueled system was supplied to reach the electrical efficiency of 42%. The economic analysis revealed that the usage of low-cost methanol can realize a decrease of 40% approximately compared to the conventional mode. The NG-fueled system can be operated at a low cost while electricity is mainly required. The hydrogenfueled system with series stacks had the highest operational cost for its high cost of fuel.