Multiresolution Analysis Based Effective Diagnosis of Induction Motors | Science Publications

American Journal of Applied Sciences

Houda B.A. Sethom¹  Mediha A. Ghedamsi¹ 

关键词: Induction motor;    unbalance voltage;    multiresolution analysis;    non linear demising detection;    transient information;    simulation results;    current signal demising;    remove efficiently;    noise filtering;   

DOI  :  10.3844/ajassp.2012.624.632

学科分类：自然科学（综合）

来源: Science Publications

【 摘 要 】

Problem statement: Effective detection and localization of unbalance voltage supply affecting an induction motor may be compromised in presence of additionnal noise. Approach: In order to overcome the non possibility of the default detection and localization in presence of noise, the use of the discrete wavelet transform and especially the MultiResolution Analysis algorithm, to remove efficiently the noise associated to the stator currents is proposed. Results: Simulation results show that the de-noised stator current is a good estimation of the non disturbed one. They show also that the default occurrence instant can be well detected starting from high frequency detail signal. Furthermore, the signal details which characterize the default are not smoothed and still characterize the default occurrence. Experimental results validate the de-noising approach efficiency and the effective unbalance detection considering the MRA technique. Conclusion: In this study, current signal denoising problem is studied in order to perform an effective detection of an unbalance voltage supply induction machine default. It can be deduced that the wavelet transform and particularly the MRA technique is a good and powerful solution for both non linear noise filtering and transient default detection. Both simulation and experimental results show clearly that the stator currents MRA allows not only to detect when the default appears but also helps to separate the useful signal from noise without affecting or suppressing the default transient information.

【 授权许可】

Unknown