Coil current based qualitative classification and recognition of electromagnetic valve degradation failure

The degradation failure qualitative classification and recognition method for intrinsically safe electromagnetic valve used in underground fully-mechanized face control system was investigated. Through analysis of the mechanical kinematical character of electromagnetic actuator in the electromagnetic valve, the electrical character model of moveable core for the electromagnetic valve was obtained and analyzed, so the character of coil current under the normal and fault states was analyzed, and the real current curves of the two types of valves under both normal and fault sates were presented. The accelerated electromagnetic valve tests were developed to collect the coil current information from two types of electromagnetic valves that are run until failure. With the methods of wavelet transform, differential coil current signal norm, and differential coil current signal norm based cloud theory the degradation features were extracted from the collected coil current information. The neural network was used to realize the qualitative classification and recognition. The experimental results show that the method researched can recognize the degradation qualitative states which can be a basis for maintenance of electromagnetic valves.