Fault Diagnosis And Toleration Technology Research Of Multi-sensors On Fuel Cell Engine

Along with the quantity of automobile growth in our country, the automobile discharges the pollution seriously day by day, the supply and demand contradictory of traditional automobile fuel is also becoming more prominent. Fuel cell electric automobile is the most realistic way to solute automobile exhaust pollution and fuel question. But the fuel cell engine is a complex system, in which there are many sensors which often work under bad conditions, breaking down is unavoidable. Once breaking down, less causes the automatic system's performance drop, more leads to a disaster. Therefore, it seems especially important to diagnose and remove the fault sensors prompt and accurately and ensure the engine working under a safety and reliable condition. This paper takes the fuel cell electric automobile as research object, develops sensors' fault-detection and fault-tolerant control technology of 25KW PEMFC engine system, and designs a new fault-tolerant system to fuel cell engine sensors. Main research results as follows:According to the performance index which the fuel cell engine must achieve, this paper has studied the overall structure of the fuel cell engine system as well as the sensor system structure, combining to the special working condition of fuel cell engine, we choose some fit sensors and useful circuit which have arrived a perfect result.According to the real situation of sensors on fuel cell engine, this paper has introduced some fault diagnosis and fault-tolerant methods, and has completed the failure detection and the fault-tolerant system structure design of fuel cell engine sensors, this structure is simple and easy to realize and the fault-tolerant effect is good.A suit of sensors' fault diagnosis and fault-tolerant system of fuel cell engine has been designed by kalman filter. In addition, an untied kalman filter fault-tolerant structure has also been completed basing on hardware redundancy. And a part of sensors have been done simulation experiment, the simulation result has indicated that, controller speed of response was quick and robustness was good.Another sensors diagnosis and fault-tolerant system research has been done by neural work, in which RBF neural has been used to designed a fault diagnosis model of sensors. This method has applied successfully in the temperature signal fault diagnosis control. The simulation indicated that this method has a good fault diagnosis and fault-tolerant control effect.This paper has thoroughly studied fault-tolerant control question of fuel cell engine control system in sensors breakdown situation. When sensors break down, this system can make the real-time estimate and feedback the signal from the detection data of disable sensors, which has ensured the stability and the reliability of system.