Research On Signal Characteristics And Fault Diagnosis Method Of Gradual Failure For Fiber Optic Current Transformer

Fiber optic current transformers(FOCT)will experience gradual failure when affected by temperature,vibration and aging of the device,which can affect the safe operation of the power system.To improve the reliability of FOCT,a gradual fault feature extraction and fault diagnosis method based on data-driven for FOCT is proposed.Gradual failure mode and the mathematical model of output signal is determined by analyzing the working principle of FOCT.The fault feature vector and the time-domain degradation feature parameter are constructed according to the fault signal feature.The fault diagnosis model and the fault signal prediction model are constructed through the fault feature vector and the degradation parameters.The model can realize the status monitoring and fault diagnosis of FOCT.The main research contents are as follows:(1)The failure mode of FOCT is analyzed: The working principle of FOCT is studied.The failure characteristics of each structure are analyzed through the internal structure principle.According to the characteristics of the fault signal,the mathematical model of the gradual fault output signal is established,which provides theoretical support for the feature extracted and fault diagnosis of FOCT.(2)The fault characteristics of FOCT are extracted: The output signal of the FOCT is studied through time domain and frequency domain methods,which determines the characteristics of the fault signal.The wavelet packet decomposition algorithm is used to decompose the output signal,and the fault signal is extracted based on its frequency information.According to the time domain characteristics of the fault signal,a multi-dimensional feature vector is established.Because the dimension of the feature vector is too high,the PCA method is used to reduce the dimension of the high-dimensional feature vector,which can increase the accuracy and rapidity of equipment fault diagnosis.The gradual fault signal has the characteristics of a large time domain span and a fixed trend of the degradation process.According to the above characteristics,the output signal of the device is sampled across the interval.The fault signal is extracted using wavelet packet decomposition algorithm.The time domain characteristic parameters of the fault signal are evaluated according to relevant evaluation indicators,and the characteristic parameter that best characterize the deterioration trend of the FOCT is obtained.The prediction model of fault signal is constructed according to the time domain characteristic parameters.(3)The gradual fault diagnosis model of FOCT is established: Aiming at the problem that the characteristics of different faults are difficult to distinguish,a gradual fault diagnosis model of FOCT based on SVM is constructed.Multi-grid parameter optimization technology is used to optimize the model parameters,which can improve the accuracy of fault diagnosis.The accuracy of the fault diagnosis model is verified by using test signals,and the validity and accuracy of the model are verified.(4)The prediction model of the gradual fault signal is established: The optimal gradient feature parameters are segmented according to the asynchronous length,and a data set that meets the requirements of the LSTM model is constructed.Fault signal prediction model based on LSTM is build according to the optimal time-domain parameters.The signal prediction model can predict the failure signal in the future,and the equipment failure early warning is realized.The test signal is used to verify the accuracy of the model prediction,which proves the accuracy of the model prediction and the feasibility of fault warning.