A novel algorithm for surface discharge stage recognition of pollution insulators has been proposed in this work based on the detailed analysis of surface discharge characteristics for the purpose of online monitoring and risk prediction of flashover from contaminated insulators. The total recognizing process of the novel algorithm comprises of four steps: first, the discharge stages were recognized based on leakage current (LC) signal through Bayesian-based pattern recognizing algorithm; second, the recognizing result of step 1 was verified immediately through humidity signals; third, if necessary, the stages were re-recognized based on humidity data by applying fuzzy theory and a membership degree was obtained; fourth, the membership degrees of both LC and humidity were input to an arbitration machine and the final recognition decision was made according to the arbitration output. As shown by the results of both intra-group and inter-group verifications, the novel algorithm exhibited good universality, specifically, the algorithm trained using one set of testing data can recognize the discharge stages of other testing data with good accuracy. Moreover, the recognition errors of the novel algorithm were limited and significant errors can be avoided, for incorrectly recognized stages were just adjacent to the correct stages.
[1]
R. Gorur,et al.
Sudden flashover of nonceramic insulators in artificial contamination tests
,
1997
.
[2]
David Christopher Jolly.
Physical processes in the flashover of insulators with contaminated surfaces.
,
1971
.
[3]
Robert Fullér,et al.
Introduction to neuro-fuzzy systems
,
1999,
Advances in soft computing.
[4]
E. Kuffel,et al.
High voltage engineering
,
2006,
2006 Eleventh International Middle East Power Systems Conference.
[5]
F. Amarh.
Electric transmission line flashover prediction system
,
2001
.
[6]
Tm. Et. Al. Romberg.
Signal processing for industrial diagnostics
,
1996
.
[7]
Fu Qiang.
Dynamic Characters Exploring of Flashover Arc for AC Contaminated Insulators
,
2006
.
[8]
Madan M. Gupta,et al.
Fuzzy Logic in Knowledge-Based Systems, Decision and Control
,
1988
.
[9]
F. Amarh,et al.
Signature analysis for leakage current waveforms of polluted insulators
,
1999,
1999 IEEE Transmission and Distribution Conference (Cat. No. 99CH36333).