Dynamic Modeling And Control Of Solid Oxide Fuel Cell Systems

Solid oxide fuel cell (SOFC) systems have many advantages given their lowemissions and high efficiency, so they are an attractive alternative for a great many ofpower applications. For the improvement of the safety, lifetime, and efficiency of SOFCsystems, the management and control about the thermal and electrical characteristics ofthe system are greatly important issues. Therefore, this research focuses on the dynamicmodeling and control for a kW scale pure hydrogen planar SOFC stand-alone system inthe perspective of thermal and electrical characteristics.On the one hand, the research is addressed in the perspective of thermalcharacteristics. Firstly, a novel SOFC system is proposed which is helpful for the controlof the stack spatial temperature, and an oriented thermal management dynamic model ofthe system is developed. Secondly, On the basis of the dynamic model, the steady stateanalysis and optimization of the system are explored in detail. The optimal systemefficiencies and operating parameters are obtained for all output power points. The systemefficiencies and stack operating temperatures optimized in the study are better than thosereported in literatures. Thirdly, the open-loop dynamics of the system is also conducted,and it is revealed that the system is a multi-variable, coupling, nonlinear, and delay system.In the context, a multi-variable decoupling control system with feedforward-feedbackcontrol structure is designed based on single neuron adaptive PID control algorithms. Inthe control system, the optimal input-output pairings are selected by using the relative gainarray (RGA) of the system, and three feedforward compensators are developed to rejectthe strong disturbance of stack current. Additionally, the single neuron adaptive PIDcontrollers have perfect robustness for the coupling, nonlinear, and delay behaviors of thesystem. Finally, through simulation, it is demonstrated that the control system canmaximize the system efficiency and ensure temperature safety at any output power point,which is a solid foundation for the control of SOFC electrical characteristics.On the other hand, the research is conducted in the perspective of electricalcharacteristics. During electrochemical impedance spectroscopy (EIS) experiments, a surprising fact was discovered that the electrical double layer in SOFC triple phaseboundary (TPB) has fractional order dynamic behavior. Therefore, a fractional orderdynamic model of SOFC electrical characteristics is proposed based on fractionalderivatives theory. The fractional order model has greater accuracy than conventionalinteger order models, so the fractional order model is an attractive model for the controlsynthesis of SOFC electrical characteristics. However, when factional order model is usedfor control design, the numerical solution of fractional derivatives is a bottleneck as itslarge computation burden. To overcome the bottleneck, an equal weight memory principleis proposed in the research. Compared to short memory and variable memory principles,the equal weight memory principle has higher accuracy and less computation burden,which is satisfying for control design. And then, On the basis of the fractional orderdynamic model, an adaptive generalized predictive control (AGPC) algorithm is designedpaying special attention to the nonlinear behavior of SOFC output power and theconstraints on stack current. The simulation results demonstrate that the dynamicresponses of the AGPC system are quick and smooth, and the stack is temperature safetyfor the change of the current is slow and smooth. By comparison, conventional PID andsingle neuron adaptive PID algorithms can not reach those goals.To sum up, in this research the management and control of the thermal and electricalbehaviors of SOFC systems are well explored, and the efficiency and lifetime of thesystem are greatly improved.