Fault Diagnosis And Fault Tolerance Control Of SOFC System With Reforming Units

Solid oxide fuel cells(SOFCs)have attracted worldwide attention due to their high efficiency and pollution-free characteristics.At present,the research on SOFC is mainly based on hydrogen-fueled SOFC stacks or systems.However,researches focus on SOFC systems with reforming units,which are more practical,is insufficient.In actual operation,various kinds of faults may occur in the SOFC system especially the SOFC system with reforming units.Therefore,fault diagnosis and fault tolerance control researches are necessary.In this paper,the SOFC system with reforming units is taken as the object,and the system model has been built.Aiming at the steam supply fault,heat exchanger rupture fault,and the composite fault occurring simultaneously,the fault diagnosis and fault tolerance control strategies are formulated.At first,based on the principles of energy conservation,molar fraction conservation,and mass conservation,the sub-components include reformers,burner,stack,the blower,the gas supply subsystem,the heat exchanger are modeled and integrated,which composed the physical model of the SOFC system with reforming units.Moreover,the established model has been validated by experimental data.Secondly,based on the data obtained from the actual experiments,the failure mechanism of steam supply fault and heat exchanger rupture fault,and the composite fault occurring simultaneously have been deeply analyzed,as well as their effect to system.Furthermore,both of these faults and the composite fault have been modeled in simulation environment.Then,the Naive Bayes algorithm,BP neural network algorithm and Bayesian regularization optimized neural network algorithm has been used separately to recognize these faults,experiments have proved the superiority of Bayesian regularization neural network for fault diagnosis of SOFC system with reforming units.Finally,based on the physical model of the SOFC system with reforming units,a fault tolerance control strategy has been proposed for the steam supply fault,heat exchanger rupture fault,and the composite fault occurring simultaneously.Firstly,a fuzzy controller is designed for the steam supply fault,a PI controller is designed for the rupture fault of the heat exchanger,and a fault tolerance control strategy combining both fuzzy controller and PID controller is designed for the compound fault.Then the fault tolerance control strategy combining fuzzy controller and PID controller has been improved.A fault tolerance control strategy based on the combination of compensated fuzzy controller and model predictive controller is proposed for the faults.This thesis has built the model of a SOFC system with reforming units,and analyzed several typical fault characteristics and their causes.Experiments have proved that the proposed fault tolerance control strategy based on compensated fuzzy control and model predictive control is better than traditional control methods,which is conducive to the safe and efficient continuous power generation of the system.The modeling analysis,fault diagnosis and fault tolerance control strategy proposed can also be applied to improve the SOFC systems’ ability to handle and recover from other typical faults.