Classification Method of Cable Defect Types Based on Broadband Impedance Spectrum and Bi-LSTM

We implement a cable defect classification method based on Bi-LSTM (Bi-directional Long Short-Term Memory). First of all, according to the different transmission characteristics caused by different defect types of cables, the cable defect types are roughly divided into five defect types: transition resistance grounding, open circuit fault, short circuit fault, and impedance increase and impedance reduction. Secondly, the simulation cable is used in the sample localization spectrum database of the five defect types. Then, the data of the two defect types of impedance increase and impedance reduction are combined into the impedance change class. Establishing the Bi-LSTM model, and the recognition effect of the four defect types was verified. Finally, if the model defect type is identified as impedance variation, the deviation direction is observed by comparing the impedance spectrum and the remaining three categories are the directly realized defect type. The experiments show that the above method can effectively realize the identification of the five most common types of cable defects, and the average accuracy of the model identification of the defect type is 96.35%. This algorithm combined with deep learning algorithm Bi-LSTM, establishes a complete set of cable defect type identification system. It solves the problem of inefficient manual classification of cable defects, reduces the requirements for workers' technology and experience.