在设备故障诊断中,正确地提取与选择特征参数对于诊断结果的有效性和准确性具有关键性的意义,在提出评价判据时样本的概率分布往往难以确定,针对模式识别中特征量的选择方法,结合人工神经网络原理,提出了利用人工神经网络进行故障特征量评价与选择的方法,实现了对柴油机特征参数的提取及选择,有效地解决了柴油机状态监测与故障诊断中测试参数多而难以优化的问题。 In the equipment fault diagnosis, it is of crucial significance to correctly extract and select the characteristic parameters for the validity and accuracy of the diagnosis results. The probability distribution of the samples is often difficult to determine when the evaluation criteria are proposed. In view of the selection of the characteristic quantities in the pattern recognition Method and combined with the principle of artificial neural network, a method of using artificial neural network to evaluate and select the fault feature is put forward. The extraction and selection of diesel engine characteristic parameters are realized, and the parameters of diesel engine condition monitoring and fault diagnosis are effectively solved Difficult to optimize the problem.