科研之友
微信
新浪微博
Facebook
分享链接
摘要
At present, there are few researches on micro-motor fault detection, and the traditional motor diagnosis methods based on single domain features have usually low accuracy. So, a fault detection method for micro motors based on the ensemble empirical mode decomposition (EEMD) and bat algorithm (BA) was proposed. The proposed method includes three steps: constructing sample sets, model training as well as parameters optimizing, and model testing. Firstly, EEMD processing was carried out on the collected micro motor signals, and the main intrinsic mode fuction (IMF) components were selected by virtue of the principle of correlation coefficient. Fusing the calculated time and frequency domain features of the motor signals, a multi-domain feature set was constructed and normalized. Then, these features were divided into a training set and a test set according to a certain proportion. Taking the training set as input and employing, the error rate as fitness, the parameters of the kernel based extreme learning machine ( KELM) model were optimized by means of the BA. Finally, the optimized BA - KELM model was tested by using the test set. The experimental results show that the accuracy of the proposed method is 98.75%, which is higher than other methods. ? 2023 Chinese Vibration Engineering Society.
- 单位
