轴承故障融合诊断方法

针对传统人工神经网络在故障诊断中应用的局限性 ,提出一种基于小波变换、遗传算法与神经网络的融合故障诊断方法。该方法先用小波变换对原始采样信号进行特征提取 ,再用遗传算法优化选择最为重要的特征作为神经网络的输入参数。最后 ,由神经网络进行状态识别和特征分类。这样不仅减少网络训练时间 ,降低网络计算量 ,而且有效提高分类的准确性及故障诊断的可靠性。轴承故障诊断实验结果表明 ,该方法是有效的。

A Fusion Diagnosis Method for Bearing Fault

This paper presents a novel diagnosis method based on a fusion of wavelet transform, genetic algorithm and artificial neural networks (ANNs), to overcome the limitatieus of traditional ANNs for fault diagnosis. By using wavelet transform, the fault features are extracted from original signals sampled. Then genetic algorithm is used to optimize and select the most significant features as inputs for the network. Lastly, recognition and classification is performed using ANNs. The proposed method not only reduces training time and cuts down the number of inputs required, but also increases accuracy of classification and raises reliability of fault diagnosis effectively. The results for applying to experiment of bearing faults indicate that diagnosis method is effective.