Study On Fault Diagnosis Of The Water-Cooled PEMFC Cooling System

In the day when traditional energy sources are becoming scarce,fuel cells,especially proton exchange membrane fuel cells,have received extensive attention due to their low noise,no pollution,high energy conversion efficiency,and low operating temperature.The temperature is an important factor for the output performance and safety of the stack.Low temperature will cause the internal ohmic resistance of the stack to increase,the polarization phenomenon will increase,and the output performance will be reduced.Excessive temperature may cause the reactor.Excessively high temperatures may lead to poor balance of temperature in the stack,high local temperatures,and the possibility of dehydration or even rupture of the membrane.Therefore,the research on the fault diagnosis of water-cooled proton exchange membrane fuel cell cooling system is studied.This paper first analyzes the operating mechanism and system composition of proton exchange membrane fuel cells,reviews the development history of fault diagnosis theory,summarizes the research status of fuel cell fault diagnosis technology at home and abroad,and independently builds a set of water-cooled proton exchange membrane fuel cell cooling system fault diagnosis platform.By analyzing the influence of different faults on the system-related parameters,a fault diagnosis strategy based on the mutual correction of multi-sensor data is proposed.The main research results are as follows:The alarm conditions for sensor acquisition data are established under steady-state conditions.When the data deviates from its non-alarm interval for a while,the fault phenomenon is considered to occur.Based on the alarm condition,this paper analyzes the characteristics of the data at the time of fault occurrence to formulate different diagnostic strategies.When the detected data deviates from the alarm threshold,the system can effectively diagnose the faults such as pipe clogging,leakage,bursting,water pump damage,and surge due to multi-sensor data correction while being able to alarm.Considering that the change of stack temperature is small when the cooling system is in normal operation,and the large jump of target speed is unfavorable to the maintenance of the centrifugal pump core,this paper develops a fixed-step speed regulation strategy for the pump.This strategy can avoid false alarms caused by the hysteresis of the flow under dynamic conditions.In order to improve the accuracy of the diagnosis,the effects of the step size and the refresh period on the alarm threshold were studied.By selecting appropriate parameters,the diagnostic system can eliminate false alarms of data under conditions that satisfy the dynamic response.The experimental results show that under the improved speed regulation strategy,each group of diagnosis strategies developed under steady-state conditions still has strong applicability under dynamic conditions.For the cooling system installed with the shunt device,this paper proposes a data-based fault diagnosis strategy for the electronic control valve.This strategy can effectively diagnose the two kinds of faults that the valve body is stuck and the feedback angle deviates from the real value.Finally,the diagnosis order is summarized as three layers: valve,multiple sensor and single sensor.The results show that when the data is abnormal,the system can alarm rapidly and in turn troubleshoot the cause of the data anomaly,and finally diagnose the faulty device or its properties.The diagnostic system has great significance for extending the life of the stack,improving the output performance and ensuring the safety of the system.