Design Of A Smart Surface Pressure Sensor For High-speed Train

Surface pressure test of high-speed trains is an important part of further study of aerodynamic effect and improving the comfort and security of high-speed trains.The piezoresistive pressure sensor used in surface pressure test is sensitive to temperature,which causes terrible zero drift and sensitivity drift.Therefore,compensation technology of smart sensor is widely used to compensate the temperature error.Smart sensors can seldom meet the requirements of high precision,high real-time and large throughput for test system of high-speed trains,therefore,a smart surface pressure sensor system combined with advanced compensation algorithm and high performance processor is designed in this paper.The compensation algorithm used in the system is PSO-BP neural network,which is a hybrid algorithm of particle swarm optimization(PSO)algorithm and BP neural network.The BP neural network has high precision,but it is easy to fall into local minimum value to cause high volatility of accuracy.Characteristics of global searching of particle swarm optimization algorithm can make up the defects of BP neural network.PSO-BP neural network is combined with the advantages of both methods,which has high precision,high stability and obvious effect of optimization.The hardware system of smart sensor uses a TMS320F28335 chip as the core.This series of DSP chips have both of high-speed data processing ability and control ability,which is conducive to realize and update functions of system.System executes AD transform of signal with DSP's own high-speed ADC module and external communication with CAN bus,which realizes the stable and reliable data transmission function.Software system includes data acquisition module,data processing module and CAN bus communication module.The system supports different sampling frequencies between 500Hz and 10kHz and the output data can be reserved 2?5 decimal places.In order to ensure the order and reliability of data transmission,a set of frame format is designed for this system.Hardware function of system is tested by using Code Composer Studio software and a simulation debugger.Functions of reading and writing instructions and data transmission is tested by a CAN to UART circuit and a serial assistant.And,the smart sensor system is used to collect data in an environment with controlled values of temperature and air pressure to test the effect of temperature compensation.Finally,the error sources of the system are analyzed.