High-voltage Circuit Breaker Mechanical Characteristic Monitoring And Pattern Recognition Of Faults

In order to confirm the reliability and the running conditions of HVCB, the research of on line monitoring and faults recognition is a hot spot at home and abroad for the service conditions of HVCB. But the research almost focused on one of the specific electrical or mechanical parameters and was lack of multi-parameters to synthetically monitor and identify. In this paper, based on the fundamental status of HVCB condition diagnosis, the characteristic between the HVCB path and time, the vibration signals when the HVCB switching in and switching out, are detected synchronously. The mechanical conditions of HVCB analyzed in many ways and combined with support vector machine (SVM), the faults pattern recognition of HVCB was done.First, according to the requirement of on-line monitoring the dynamic characteristic of HVCB, the experiment test is set up by choosing the adaptive displacement transducer and acceleration transducer. The velocity of the HVCB's moving contact is deduced from the principle of four linkage mechanism and the real dimensions of HVCB drive-gear. So the all important velocity parameters can be calculated which include the maximum velocity of opening-closing and the average velocity of opening-closing. It is certificated that the monitoring platform can meet the circuit breakers path - time of monitoring by comparing the calculated data and the data tested by equipment.For the lack of vibration signal spectrum analysis by FT, then put forward by the new method of Empirical Mode Decomposition(EMD). In light of the information entropy theory, The main parameters are extracted by the empirical mode decomposition .As the dimension of parameters are a little more, the method of factor analysis is introduced to compress the dimension.The new parameters are invented by combining the parameters of HVCB vibration signals with the parameters of HVCB velocity characteristics. Compare the pattern recognition effect of vibration signals characteristic and the effect of combinated parameters, a recognition model of multi-parameter is established.At last, the SVM is presented to construct the HVCB faults pattern recognition system, it reveals that the method of faults pattern recognition based on multi-parameters is effective.