Research On Fault Diagnosis Method Of Water And Thermal Management System Of Fuel Cell Engine

In today’s rapidly developing world of new energy technologies,hydrogen energy is attracting attention for its high efficiency and non-polluting characteristics.The proton exchange membrane fuel cell(PEMFC)engine is a power unit that converts hydrogen energy into electrical energy.The operating principle is complex and includes multidisciplinary knowledge of electrochemistry,thermodynamics and conservation of mass of matter,and is highly susceptible to failure.The water and thermal management system safeguards the water balance and temperature equilibrium within the engine and maintains suitable operating conditions for the engine.Flooding and drying are typical failures of PEMFC engine water and thermal management systems.Flooding tends to cause loss of engine output and drying increase the resistivity of the proton exchange membrane,reducing the power output of the PEMFC engine and even affecting its service life.Therefore,it is important to investigate the diagnosis of flooding and drying as typical faults in the water and thermal management system of PEMFC engine.The traditional diagnostic methods for flooding and drying faults are generally based on the output of tests or models,which are not accurate enough and limited in terms of intelligence development.The diagnostic method for flooding and drying faults in the water and thermal management system of PEMFC engines is aimed at studying the fault mechanism of flooding and drying faults,establishing a semi-empirical model of PEMFC engines and a BP neural network health recognition model,using a combination of model and data-driven methods,and using three judgment indicators of output voltage,cathode voltage drop and ohmic internal resistance of PEMFC engines to A diagnostic method was investigated for water flooding and drying faults.The PEMFC engine is a strongly coupled mechanism with a complex gas-hydraulic structure,and the causes of flooding and drying failures are complex.The mechanism of flooding and drying,which are typical faults of the water and thermal management system of PEMFC engine,are studied in depth and the main influencing factors of flooding and drying are summarised based on the PEMFC electrochemical principle and water transfer mechanism.Combining the advantages of the accuracy of the mechanistic model and the simplicity of the empirical model,a semi-empirical model of the PEMFC engine is developed based on MATLAB-Simulink,mainly for modelling the electric stack part,the model includes cathode runner model,anode runner model,membrane hydration model and voltage model.The reliability of the model was verified by analyzing the polarization curves and polarization voltages of the main calculation results of the model.Based on the PEMFC engine flooding and drying failure mechanism and semiempirical model,a diagnosis method for flooding and drying failure is proposed.A PEMFC engine BP neural network health identification model is established,the residuals between the BP neural network model health data and the semi-empirical model fault data are calculated,and the fault diagnosis of flooding and drying is achieved by residual classification using three judgement indicators: PEMFC engine output voltage,cathode voltage drop and ohmic internal resistance.