Sensor Fault Diagnosis Study Of Fuel Cell Engine Based On Wavelet Analysis

Fuel cell engine system is a new energy engine system. It is a new engine system composed by fuel cell stack, hydrogen circulation system, air input system, cooling water system and data control monitoring system. In the possess of the fuel cell engine system's operation Voltage, current, pressure, flow and other dates needed to be collected and monitored can ensure the fuel cell engine system to be in a normal operation. Sensors are applied to this system to collect these data. However, these sensors if there are problems not only affect the normal acquisition of dates, If the problem is serious, but also will damage the fuel cell's performance, and bring significant losses to the entire fuel cell engine system. So, Identifying sensor's faults and reasons in a time is very necessary to the healthy and orderly operation of fuel cell engine. In this paper, sensor system of 60kw fuel cell engine is studied, sensors are diagnosed through wavelet analysis. The main research contents are as follows:First, the fuel cell engine system architecture is introduced, especially the sensors applied in these structures and the signal acquisition system composed of the sensors. On this basis, the collected signals are processed with by butterworth low-pass active filter, and this paper analyzes the detailed circuit structure and selects the parameters.This paper makes a description of sensor fault types and performances in the fuel cell engine system. By comparing the methods of the sensor fault diagnosis, the shortcomings of existing diagnostic methods are pointed out. According to the causes of the sensor fault, sensor faults are classified.Through the wavelet time-frequency analysis, the time-frequency localization capability of wavelet transform is reflected. Through wavelet transform, Signals are decomposed and reconstructed and can be analyzed accurately。This paper proposes sensor fault diagnosis of the fuel cell engine based on wavelet analysis simulates each fault type with matlab wavelet analysis toolbox and gets simulation maps. Sensor fault diagnosis study of the fuel cell engine based on wavelet analysis shows that wavelet analysis has the ability of rapid, efficient, high rate and overcoming the noise characteristics strongly of fault diagnosis. So fault detection can enhance the accuracy, validity and reliability.