|
标题：基于噪声分析的电机故障诊断方法研究 |
|
作者：张媛，王飞，王万成 |
|
| 2012年第8期 访问次数：207次 |
|
摘要：本文基于Matlab平台,对电机各种运行状态的噪声进行数字分析和处理,提出了一种基于噪声分析的电机故障类型的诊断方法。首先对电机不同运行状态下的声音信号进行时频域分析和去噪处理；其次分析和提取了信号的能量特征向量。最后利用神经网络分类器对电机故障声音进行了分类识别。此外,为提高智能诊断的准确性，本文在以上一系列过程中均采用了多种方法，并进行了择优。实验表明，本文研究的基于噪声分析的电机故障诊断方法具有较高准确率。
关键词：电机噪声；故障诊断；小波包分解；能量特征向量；神经网络
Abstract: This paper，based on Matlab platform，used the digital signal processing methods to analyze and process motor noise under the different running status, proposed a diagnostic method for the motor failure type. Firstly，this paper carried on analyzing these signal〖WTBZ〗s under the different running status in time domain and frequency domain. Next, analyzed and extracted the signal〖WTBZ〗s energy eigenvector. Finally, classified and distinguished motor fault voice with the neural network sorter. In those series of process, this paper used many kinds of methods and chose the best method to apply in the motor failure type diagnosis to improve the accuracy of the intelligent diagnosis. The experiment indicates that this article has achieved high rate of accuracy in motor failure type recognition which based on the noise analysis.
Key words: motor noise; fault diagnosis; wavelet packet analysis; energy eigenvector; neural network
|
|
| |
|
