Dynamic Simulation of Fuel Cell Systems

The development of dynamic simulation tools for fuel cell systems and fuel cell/gas turbine (FC/GT) systems is required to understand the response of fuel cell systems to both scheduled and unscheduled perturbations. These tools supply the foundation on which control strategies can be developed and tested before they are actually implemented in a physical system. For high temperature fuel cell systems that generally incorporate reformation of hydrocarbon fuels externally and/or internally, knowing how the system can potentially respond to a load perturbation or a fuel flow composition perturbation is extremely important to system performance and reliability. In addition, when developing control strategies for high temperature fuel cell systems, it is important to identify of the hazardous conditions that may lead to degradation in performance or the destruction of the fuel cell and other components of the system. Being able to identify these conditions with a detailed and robust dynamic model can lead to safer and more responsive operation for fuel cell systems. This work provides a dynamic simulation of a load perturbation that could produce undesirable conditions if proper control strategies are not designed for the fuel cell system. In addition, the dynamic simulations provide insights into both component and system dynamic performance parameters that illuminate some less than obvious dynamic responses within representative integrated high temperature FC/GT hybrid systems. Approach Dynamic models for molten carbonate fuel cells (MCFC), heat exchangers, gas turbines (GT) and catalytic oxidizers have been developed in a Simulink platform. The dynamic models are based on first principles. Descriptions of the models are found in previous papers 1, . All the models incorporate the conservation equations for energy, mass, and momentum. The MCFC model incorporates the electrochemistry and internal reformation chemistry reactions as well. The catalytic oxidizer model assumes complete combustion. The gas turbine uses general performance maps. These components are integrated to simulate a complete molten carbonate fuel cell/gas turbine (MCFC/GT) system. In each of the significant component models (e.g., fuel cell, heat exchanger) some degree of geometric resolution is captured, albeit in a simplified (usually one-dimensional) manner. The dynamic equations that govern the concurrent processes of heat, mass, and momentum transfer, chemical reaction and electrochemical reaction are solved, for example along one dimension of the component representing that which varies most significantly and including geometrical features such as actual fuel cell size and shape. This approach to dynamic simulation is significantly more computationally intensive than the traditional bulk component model approach, but it provides significant added value. First, insights into the performance parameters of each component are provided (e.g., temperature profiles) so that one can determine whether or not significant component stress or other conditions of concern may be reached during responses to perturbations. Only if this insight is provided can one determine what needs to be controlled and how to control it. Secondly, accurate predictions of component performance cannot be achieved without this dimensional and geometric resolution. This is certainly the case for reactors such as catalytic oxidizers or fuel cells which rates of chemical reaction are exponentially dependent upon local temperature. This is also true for systems as simple as heat exchangers whose dynamic performance cannot be captured through a bulk effectiveness heat transfer model. As a result, these more complex and computationally intensive, yet simply resolved component models both contribute to increased insight and more accurate predictions of fuel cell system performance. Results A complete dynamic MCFC/GT system model was constructed to simulate a 1MW MCFC/GT system similar to the sub-MW Direct Fuel Cell/Gas Turbine (DFC/GT) hybrid system developed by FuelCell Energy. A diagram of the system configuration is presented in Figure 1. The model was developed by integrating individual robust simulation modules of the type described above, each simulating the dynamic performance of a component shown in Figure 1. The integration is accomplished using the graphical user interface and toolbox utilities of Simulink. To demonstrate the dynamics of this type of hybrid fuel cell system, a load perturbation was simulated on an open loop system of the type presented in Figure 1. For the load perturbation investigated, the MCFC underwent a 3% load drop. The power drop was achieved by changing the external load resistance applied to the MCFC. Figure 2 presents the results for the MCFC, GT and total plant power as the integrated system responds to the load perturbation. The MCFC power initially drops, undershooting by approximately 12% (90 kW)of the final steady state power. This effect is due to the temporary deficit of fuel in the anode gas (at the end of the anode compartment of the fuel cell). The MCFC power quickly recovers to a lower steady state value of power output. The total plant power reflects the dynamics of the MCFC power response. The GT power begins to rise after 5 seconds due to a rise in the turbine inlet temperature (TIT). It takes approximately two minutes for the GT to reach steady state. There is a slight recovery in total power because of the rise in GT power. When the MCFC drops to a lower power level, there is more unspent fuel in the anode stream. This increase in unspent fuel causes the catalytic oxidizer temperature to rise. The rise in catalytic oxidizer temperature is illustrated in Figure 3. The rise in catalytic oxidizer temperature increases the GT TIT, which results in higher GT power. The increase in catalytic oxidizer temperature also increases the cathode inlet temperature, which is presented in Figure 3 as well. Decreasing the MCFC power by 3% caused an 11% increase in GT power and an overall 2% decrease in total plant power. The cathode inlet temperature experienced a 14oC temperature rise, which could potentially overheat the MCFC. Conclusions Open loop results were presented to demonstrate the dynamics of the MCFC/GT system. From the dynamic results it can be seen that great care must be taken to control the MCFC/GT hybrid system. Safe operating temperatures need to be maintained throughout the system to improve the performance and life cycle of the system. To design FC/GT systems with load following capabilities one must include intelligent control strategies that can meet changing load demands while maintaining high performance and safe operation. The integration of various component technologies (e.g., fuel cell, GT, heat exchangers) results in a complex system with dynamic response characteristics that are not easy to elucidate without robust and geometrically resolved models. Prepr. Pap.-Am. Chem. Soc., Div. Fuel Chem. 2004, 49(2), 796 Figure 1. DFC/GT hybrid system diagram 700 750 800 850 900 95