Abstract
In order to realize the accurate positioning and recognition effectively of the analog circuit, the feature extraction of fault information is an extremely important port. This arrival based on the experimental circuit which is designed as a failure mode to pick-up the fault sample set. We have chosen two methods, one is the combination of wavelet transform and principal component analysis, the other is the factorial analysis for the fault data's feature extraction, and we also use the extreme learning machine to train and diagnose the data, to compare the performance of these two methods through the accuracy of the diagnosis. The results of the experiment shows that the data which we get from the experimental circuit, after dealing with these two methods can quickly get the fault location.
Originalsprog | Engelsk |
---|---|
Titel | Proceedings of the 8th International Conference on Signal Processing Systems, ICSPS 2016 |
Antal sider | 5 |
Forlag | Association for Computing Machinery |
Publikationsdato | 21 nov. 2016 |
Sider | 173-177 |
ISBN (Elektronisk) | 9781450347907 |
DOI | |
Status | Udgivet - 21 nov. 2016 |
Udgivet eksternt | Ja |
Begivenhed | 8th International Conference on Signal Processing Systems, ICSPS 2016 - Auckland, New Zealand Varighed: 21 nov. 2016 → 24 nov. 2016 |
Konference
Konference | 8th International Conference on Signal Processing Systems, ICSPS 2016 |
---|---|
Land/Område | New Zealand |
By | Auckland |
Periode | 21/11/2016 → 24/11/2016 |
Navn | ACM International Conference Proceeding Series |
---|
Bibliografisk note
Publisher Copyright:© 2016 ACM.