Research On The Method Of Information Acquisition And Processing For Aero-engine Sensor

Failure are most likely to occur in sensors in aero-engine control system. Its failure may result in the breakdown of system and catastrophic results will follow it. To establish and perfect detection, analysis, isolation of the aero-engine failure signal can improve the reliability of the engine control system. A signal processing system is designed according to collection and processing method of the sensor signal, and with the use of embedded IPC, which is quick and real-time,and use neural network with good nonlinear characteristics, fault tolerant ability and learning mechanism. The system is used for detection and analysis of engine sensor signals of aero-engine sensors, and online monitoring, isolation of the hardware failure signals of sensors real-timely.Firstly, select diagnosis methods for sensor failures and choose the corresponding engine signal. Then determine the sensor type, and acquire the engine of the signal. Based on output signal form, we choose IPC UNO-1019 which uses ARM9 motherboard as data processing hardware platform and modified WinCE as the OS of hardware platform.When data acquisition is completed, it will be sent through the Ethernet to the data processing module. The network transmission program is developed with c + + and WinSocket socket is used to link both sides of the TCP/IP network and transmit files, and write the data of received files into the database of sqlite3. Then process the states information of sensors.The BP algorithm is applied for sensor information processing. The paper do a brief study of the basic algorithm of BP network. So is the design of BP network. Then establish engine neural network model, and verify the algorithm with MATLAB, and determine the principles hardware detection threshold selection. Using C+ + to transplant the BP algorithm into the embedded system. A computing simulation is made for the sensor signal processing capacity and export graphics.Based on analysis of simulation graphics, when the sensor is hard fault, the software can deteceted it quickly and accurately, and the result is very obvious, the result of the system for the station of sensor is very reliable abd effective.