A fault diagnosis and degradation failure prediction method for the electromagnetic valves which are used in hydraulic supports electro-hydraulic control system of the underground combined mining face is proposed.Through analyzing the mechanical kinetic characteristic of electromagnetic valves

the parametric relationship among the driving voltage

length of magnetic circuit working air gap and driving current is determined and the steady state and transient performance of driving end current is also obtained.According to the transient performance

the energy feature and K-L transforming feature based on template matching is extracted in the movement interval of valve spool

and according to the steady state performance

the inductive reactance feature of the drive coil is also obtained in the static interval of valve spool.24

000 sets of data are gained by 2.4 million switching experiments in 143 days.The five stages of electromagnetic valves which are normal operation

beginning of failure

a little of failure

some failure and damage is accurately identified between two types of electromagnetic valve by using the neural network analytical approach

and the rates of identification are respectively 98.1% and 99%.The study provides basis for the fault prediction and maintenance of electromagnetic valves

and the productive efficiency of coals and the productive safety are improved.This prediction method can be widely applied to the degradation failure detection of many other types of electromagnetic valves.