Dynamic Simulation Of A Direct Carbonate Fuel Cell Power Plant

John B. Ernest Fluor Daniel, Inc. Irvine, CA 92698 USA

Hossein Ghezel-Ayagh and Ashok K. Rush Fuel Cell Engineering, a subsidiary of Energy Research Corporation Danbury, CT 06813


Fuel Cell Engineering Corporation (FCE) is commercializing a 2.85 MW Direct carbonate Fuel Cell (DFC) power plant. The commercialization sequence has already progressed through construction and operation (Ref. 1) of the first commercial-scale DFC power plant on a U.S. electric utility, the 2 MW Santa Clara Demonstration Project (SCDP), and the completion of the early phases of a Commercial Plant design. A 400 kW fuel cell stack Test Facility is being built at Energy Research Corporation (ERC), FCE's parent company, which will be capable of testing commercial-sized fuel cell stacks in an integrated plant configuration. Fluor Daniel, Inc. provided engineering, procurement, and construction services for SCDP and has jointly developed the Commercial Plant design with FCE, focusing on the balance-of-plant (BOP) equipment outside of the fuel cell modules.

The nature of the next-generation carbonate Fuel Cell power plants is that they will be first-of-a-kind plants, which will follow widely varying loads, with controls automated to run without human intervention. The DFC plant will be highly heat integrated with a very slow thermal response due to the large thermal inertia of the stack itself. These attributes present a compelling case for the use of Dynamic Simulation. Fluor Daniel and FCE have used Dynamic Simulation to address practical issues of process and control design for the past two years.

Dynamic simulation models developed by others have focused on the 2-dimensional temperature details of the stack itself (2, 3), or on the steam reformer (4), or on the separate issue of inverter dynamics, or have taken a theoretical approach to an entire plant based on a simplified BOP (5). Fluor Daniel and FCE have developed a dynamic simulation model of the entire integrated cell stack and BOP for the Test Facility and Commercial Plant and produced results suitable for design. We believe that this effort is broader and has achieved more useful results than all the other dynamic simulations reported in the open literature.

This paper provides a brief orientation to the dynamic simulation technique, its application to FCE's DFC power plant, and the benefits offered by this tool. An illustrative simulation is described and figures show the major plant responses.

The Dynamic Simulation Technique

Dynamic simulation is a process engineering design tool that predicts how a process and its controls respond in time to various process and control changes. In the past, the use of dynamic simulation as a design tool was limited to simple linear modeling techniques for applications such as operator training simulators. Increasingly however, rigorous dynamic simulation with integrated energy and material balances, hydraulics, thermodynamics, and control modeling is being used to design a variety of industrial processes.

Dynamic simulation models must be sufficient to reproduce rigorously the trends in the net performance of the plant equipment and the stability of the control system. This requires much more detail than steady-state models or dynamic models that rely on transfer functions. Fluor Daniel has experience in applying rigorous dynamic simulation to conventional power plants and a variety of other applications from the refining and petrochemical process industries (6, 7, 8).

The dynamic simulation models for the DFC were developed using Aspen Technology, Inc.'s SPEEDUP general purpose dynamic simulation computer program.

DFC Power Plant Dynamic Simulation Modeling Status

Fluor Daniel, in collaboration with FCE, has developed dynamic simulation models for both the Commercial and Test Facility power plants. The Commercial Plant model is based on the data available from the Preliminary design phase. Cases were run which verified this model with two conventional steady-state simulators developed by FCE and Fluor Daniel. Also, a number of simplified transient scenarios were studied. The trends and cause-and-effect relationships predicted by the dynamic simulation model for these simplified scenarios give confidence in the predictions for more realistic cases. The dynamic simulation models will be updated with data from ERC's Test Facility and Commercial Plant design refinements.

DFC Power Plant Modeling Technical Issues

The focus of DFC power plant modeling is on prediction of: control system stability and response; oxidizer operations and cycling; warm-up and cool down time; interactions between the oxidizer, heat exchangers, and stack; the temperature effect on voltage; transient violations of design constraints (none of these can be found by steady-state analysis); and unforeseen problems.

The following characteristics are very important to the dynamic behavior of the DFC power plant: the high degree of heat integration between anode and cathode streams, oxidizer, and DFC stack; the large oxidizer due to startup requirements; the very slow thermal response of the DFC stack; the extensive interactions and constraints on controls; and the extensive control logic, encompassing power generation, standby, startup, and shutdown operations.

Most of the process must be simulated rigorously due to process interactions. For example, to predict accurately the cathode inlet temperature, the dynamic simulation model must predict the heat transfer with the insulated pipes downstream of the oxidizer. However, some physical phenomena can be ignored, such as the fuel treatment system's pressure gradient. The model of the DFC stack includes the temperature, electrical, and gas composition changes due to reforming, shift, anode, and cathode reactions; heat exchange between the cell components and the reformer, anode, and cathode streams; and the current density-voltage-power generation characteristic. The model of the BOP is equally rigorous. Its fuel treatment system includes the natural gas and water/steam flow control valves, preconverter action, and heat exchange. The oxidizer system includes the air supply blower, duct head loss, air flow control valves, thermal burner, catalytic oxidizer, and piping heat exchange and pressure loss. The plant controls include the oxidizer controls, DFC power generation and stack temperature, and master logic.

Figure 1 presents a Block Flow Diagram of the Commercial Plant. Controls and process are of equal prominence.

Simulated Plant Performance

Plant performance is predicted for a wide range of cases: dc power generation rate ramps; control setpoint and mode changes; upsets in gas, steam, air, and auxiliary electrical feeds and thermal burner firing; emergency scenarios; and startup and shutdown.

Figures 2 and 3 present typical simulation results. In this case, the power is being increased from 25% to 100% of rated power, while the oxidizer outlet temperature controller setpoint is raised 100°F at the beginning of the power ramp, then lowered 150°F at the end of the power ramp. The gas flow is set to maintain a fixed fuel utilization. The steam flow is set to maintain a fixed fuel steam/carbon ratio until the flow reaches its maximum.

Figure 2 shows the temperature profile. There are large transients in the oxidizer outlet temperature as the temperature controller follows its setpoint. The temperature is much different at the cathode inlet due to the heat exchange with the intervening pipe. The cathode outlet temperature is still changing long after the power and oxidizer temperature setpoint ramps are complete. Figure 3 shows the stack electrical performance. The major features of Figure 3, the voltage drop and current density increase as the power is increased, could be determined from a sequence of steady-state runs. However, the voltage undershoot and fluctuations are due to the maximum steam flow constraint and controller transients.


Fluor Daniel and FCE have developed dynamic simulation models for both a near-term 400 kW fuel cell stack Test Facility and the preliminary design of FCE's DFC Commercial Plant. These models are already in use and soon will be validated against Test Facility experimental data. The Test Facility dynamic simulation model is able to predict the thermal behavior and control response of the entire power plant. The use of dynamic simulation models in engineering design allows evaluation of system dynamics, equipment design parameters, and control strategies.


The design effort described in this paper was supported by the U.S. Department of Energy (Contract No. DE-FC21-95MC31184) and FCE.


1. A.J. Leo, A.K. Kush, and M. Farooque, "Development and Demonstration of Direct Fuel Cell System at Energy Research Corporation," IECEC, August 12, 1996, Washington D.C.

2. K. Bolwin, S. Hauff, and W. Schnurnberger, "Dynamic Simulation of Heat Dissipation and Electrochemical Conversion Rates," Fuel Cell Seminar Program and Abstracts, pp. 315-318, Nov./Dec. 1994, San Diego, CA.

3. W. He, R. W. J. Kouffeld, A. Korving, and J. G. M. Becht, "The Dynamic Performance of a Molten Carbonate Fuel Cell in Power Generation Systems," op. cit., pp. 602-605.

4. G. L. Ohl, G. E. Smith, and J. L. Stein, "Dynamic Models of a Methanol to Hydrogen Steam Reformer for Transportation Applications," op. cit., pp. 491-494.

5. W. He, "Dynamic Modeling and Control of Molten Carbonate Fuel Cell Systems," Technische Universiteit Delft, Rept. EV-1754,10 Jun. 1994.

6. W. E. Kirchmeier, F. M. Faubert, and W. T. Reid m, "Furnace Implosion Study Verified by Trip Test," Instrumentation in the Power Industry, Vol. 20, Proceedings of the Intl. ISA Power Inst. Symposium, pp. 75-91, May 1977.

7. J. B. Ernest and C. A. Depew, "Use dynamic simulation to model HPU reactor depressuring," Hydrocarbon Processing, Vol. 74, No. 1, pp. 72-79, Jan. 1995.

8. C. A. Depew and R. B. Nielsen, "Dynamic simulation for process design," Hydrocarbon Processing, Vol. 75, No. 7, pp. 67-75, Jul. 1996.

voltage & current density^

temperature (arbitrary units)



0 0

Post a comment