基于概率神经网络和声压的柴油机故障诊断  

Diesel Fault Diagnosis Based on Probabilistic Neural Network and Acoustic Pressure

作　　者：程利军[1] 张英堂[1] 齐子元[1] 范红波[1] 

机构地区：[1]军械工程学院,河北石家庄050003

出　　处：《车用发动机》2008年第1期62-64,共3页Vehicle Engine

基　　金：河北省自然科学基金资助项目(E20007001048)

摘　　要：论述了利用柴油机运转缸盖噪声信息来诊断内燃机故障的一种方法,通过FIR数字滤波技术从发动机缸盖噪声信号中分段提取发动机的燃爆和气门开闭信息,采用参数平均法消除了缸盖噪声测量的随机误差,利用概率神经网络的分类功能,准确判断出了各种故障。The method of diagnosing the faults with cylinder head acoustic information of running diesel engine was introduced. The burning explosion information and the opening and closing information of valve were extracted piecewise from cylinder head acoustic information, the random error of acoustic measurement was eliminated with parameter mean, and finally all kinds of faults were classified correctly with probabilistic neural network.

关 键 词：柴油机 故障诊断 声压 概率神经网络 参数平均 FIR滤波 

分 类 号：TK428[动力工程及工程热物理—动力机械及工程] TB535[理学—物理]