将计算机用于机械工况的自动诊断时 ,必须从采集的信号中提取表征机械工况的特征参数 (简称SP) ,但目前尚无一种令人满意的提取最佳 SP的方法。为了克服这个困难并能保证对机械工况进行高精度的诊断 ,本文基于遗传基因算法 (简称 GA)提出了一种称为“特征参数自动生成”的新方法 ,能快速找到最佳 SP,并以变速器诊断为例进行了验证 When the computer is used for the automatic diagnosis of mechanical working conditions, the characteristic parameters (SP) characterizing the mechanical working conditions must be extracted from the collected signals. However, there is not a satisfactory method for extracting the optimal SP at present. In order to overcome this difficulty and ensure the high-precision diagnosis of mechanical condition, this paper proposes a new method called “automatic generation of feature parameters” based on genetic algorithm (GA), which can quickly find the best SP Transmission diagnosis as an example has been verified