Health Management Of PEM Fuel Cell Systems

Due to the global energy crisis and environment pollution,proton exchange membrane fuel cells(PEMFCs)are gaining more and more attention.PEMFCs are considered to be the ideal form of new energy for vehicles.However,due to their economy and durability issues,fuel cells cannot be commercialized on a large scale.The complex reaction mechanism of PEMFCs,coupled with different sensors and actuators,leads to a larger chance of PEMFC system fault.If a PEMFC system or stack failure is not diagnosed in time,it may cause a decrease in reaction efficiency,irreversible performance damage to the PEMFC stack and key components,or even serious safety problems.Moreover,the failures of the PEMFC thermal management system and water management system are thought to be the main cause of the stack failure.The research related to thermal&water management subsystem is the focus and difficulty of the PEMFC research.At present,there are many methods to improve the durability of PEMFCs.One is to improve the performance of PEMFC stacks from the perspective of key materials and components.It is often achieved by developing new proton exchange membranes and catalysts,and developing new three-dimensional structures of fuel cells;In addition,the reliability and durability of the PEMFC system should be improved through the optimization control and health management.PEMFC health management usually consists of fault diagnosis and fault tolerant control.Based on the different methods of fault diagnosis,PEMFC health management is often classified into two categories:model-based methods and non-model-based methods.Model-based approaches generally have higher robustness and better fault interpretability,but require more complex models;On the contrary,non-model-based approaches are less difficult to implement,but less portable.In this paper,an in-depth health management study will be conducted on PEMFC temperature management system,water management system,and fuel cell stack,respectively.The main contributions of this paper are summarized as follows.1.The key components and core process modeling of PEMFC system are developed.The thermal management model,water management model and stack voltage model of PEMFC system are introduced and developed in detail.Moreover,semi-empirical models are also established for the core components of the PEMFC system,such as water pump,exchanger,humidifier and other components,which improves the reliability and accuracy of the modeling process.Theses models provide a solid foundation for the subsequent model-based fault diagnosis and fault-tolerant control.2.A model-based fault-tolerant control method for PEMFC temperature management system is proposed.Firstly,based on the proposed thermal management model,the structural analysis approach and DM decomposition are used to analyze the detectability and isolability of sensor faults.A sensor fault detection algorithm is designed based on residual.On this basis,an active fault-tolerant controller is designed based on sliding mode method.Furthermore,considering that the traditional sliding mode controller will have a large oscillation effect on the actuator during the steady state,a smooth switching function is adopted,and the stability of the closed-loop system is proved.Finally,based on the existing experimental platform,the performance of the proposed fault diagnosis algorithm and active fault-tolerant controller is verified online.The proposed algorithm has high control accuracy in the fault-free state.Moreover,it can maintain the control accuracy of ±0.5℃ under the sensor fault,which significantly improves the reliability of the temperature control system.3.A model-based fault-tolerant control method is proposed for fuel cell water management system.First,the water saturation model of fuel cell is equivalent to four sub processes,and the internal water saturation index of PEMFC is established.The accurate response of the designed index to the water flooding state is demonstrated by EIS testing.Based on this,a fault tolerant control algorithm based on AFTC framework is designed.Using the methods of piecewise linearization and linear programming.Thanks to a sliding mode controller,accurate tracking of the water saturation index against the set value is achieved,with an error of less than 1%.Finally,the performance of the proposed fault indicators and active fault tolerant control algorithm is experimented on a real PEMFC system.The experimental results show that without fault tolerant control,the PEMFC stack is prone to flooding under high current conditions(larger than 90A).On the contrary,the proposed controller is able to eliminate the flooding fault and maintain the stack in a healthy state,which improves the system durability.4.A PEMFC health management method is proposed based on fast EIS technology.First,in order to detect the performance of PEMFC stack,a PEMFC condition monitoring algorithm based on voltage is proposed.By comparing the healthy state voltage with the real-time voltage measurement,it can judge whether the stack is in faulty state,and the average condition monitoring error is 0.38%.When the residual exceeds the threshold,the fast EIS measurement is applied,and then the fast electrochemical impedance spectrum is analyzed by using the relaxation time distribution technology for fault diagnosis.Based on this,the system performance can be recovered by adjusting the control strategy.Finally,the proposed multi-fault diagnosis algorithm and fault tolerant control strategy are verified by experiments.The results show that the designed fault diagnosis method can reduce the EIS measurement time from tens of minutes to 15 seconds.The maximum relative error of fast EIS does not exceed 1.5%,and has a very high diagnostic rate.The health management method is also the first attempt to apply DRT analysis to fast EIS measurement,and provides online performance recovery for flooding,drying and air starvation faults,providing a viable idea for the application of EIS technology to online processes.