分布式光纤传感系统托辊故障监测技术研究OA

In order to minimize interference with coal mine production systems,a fault diagnosis method for idler rollers based on distributed fiber optic sensing system was studied with the background of fault diagnosis for mining belt conveyors.By improving the traditional isolated forest,a majority voting isolated forest(MV-iForest)framework is proposed.Firstly,two types of fiber optic arrangement are used to collect data from the same faulty roller.Then,the collected time-domain signal is subjected to discrete wavelet transform(DWT)denoising processing.Fast Fourier transform(FFT)and short time Fourier transform(STFT)are employed to obtain frequency domain and time-frequency domain signals of faulty rollers under two different arrangements,and comparative analysis,is conducted to determine the appropriate fiber optic arrangement.Further analyze the time-domain,frequency-domain,and time-frequency domain signals under the fault conditions of roller bearing detachment and cylinder skin fracture,and compare the roller fault signals with normal operating signals.Next,the threshold of the majority voting method was discussed to determine the optimal threshold.Finally,a comparative evaluation is conducted between the majority voting isolated forest(MV-iForest)and the original isolated forest(iForest)and the extended isolated forest(E-iForest),indicating that the improved isolated forest has faster computation time and higher accuracy,and can effectively diagnose roller faults.