作　　者： ; ; ; ; ;

机构地区： 郑州大学机械工程学院振动工程研究所

出　　处： 《机床与液压》 2007年第5期208-210,196,共4页

摘　　要： 结合小波包分析和概率神经网络技术,提出了一种基于小波包分解-概率神经网络的机械故障自适应报警方法。该方法利用小波包获取振动信号各有效频带的能量作为报警参数,用概率神经网络构建设备运行状态模型,根据历史数据确定报警值并设置报警线。实验结果表明该方法是有效的,它克服了设备状态监测中报警线的设置与设备运行情况变化无关的缺陷,该方法在机械设备报警处理系统中有良好的应用前景。 Combined wavelet package （WP） with probabilistic neural network （PNN）, an adaptive alarm method named WPPNN was proposed. This proposed method is that the energy information in every frequency band obtained by the wavelet package decomposition is used as the alarm parameters, and the improved probabilistic neural network is used to construct the model of real running condition, the alarm line is determined according to history data from the running machine. Experimental results show that this method is very effective. The proposed method overcomes the deficiency in the setting of the alarm parameters in the machine fault monitoring, i. e. the alarm parameters are nearly independent of the various running condition. The method has very good application prospect in the alarm processing system of machinery.

关 键 词： 小波包分解 概率神经网络 自适应报警 故障诊断

领　　域： [自动化与计算机技术] [自动化与计算机技术]