Abstract:Detection and pattern recognition of different types of partial discharge (PD) signals in transformer are of great significance for the safe and stable operation of transformers. This paper demonstrates the application of mi-crofiber coupler sensor (MFCS) to the pattern recognition of PD ultrasonic signals with an optimized support vector machine (BSO-SVM) algorithm. First, a PD ultrasonic detection system based on MFCS is built, compared with the traditional piezoelectric sensor, the MFCS can be directly implanted in the transformer oil to detect the PD signals, avoiding the attenuation of ultrasonic signals as they propagate to the transformer outer wall. Then, four kinds of PD ultrasonic signals are collected by the PD detection system, and feature extraction is carried out for the time domain and frequency domain characteristics of PD ultrasonic signals. Finally, a BSO-SVM classi-fication algorithm is proposed, which uses the beetle antennae search (BAS) and the particle swarm optimization (PSO) algorithm to jointly optimize the support vector machine (SVM) algorithm. The BSO-SVM classifier is used to classify four kinds of PD signals, the classification accuracy is 93%, and the number of convergence is 5. In addition, the BSO-SVM algorithm is compared with the GA-SVM and PSO-SVM algorithms. The results show that the BSO-SVM algorithm has higher recognition accuracy and faster convergence speed. The research results show that it is feasible to use MFCS combined with BSO-SVM to identify PD patterns of transformers, and it has certain significance for transformer fault diagnosis.

Partial Discharge of UHF Pattern Recognition in Transformers Using Adaptive Neuro-Fuzzy Inference System (ANFIS)

Application Of Adaptive Neuro Fuzzy Inference System To The Partial Discharge Pattern Recognition

An Automated Recognition System Of Ultra-High-Frequency Pd In Transformers

Application of ANFIS to Partial Discharge Defect Recognition Based on Statistical Parameters

Recognition of Transformer High Frequency Partial Discharge Based on Time Domain Feature

Application of Statistic Parameters in Recognition of Partial Discharge in Transformers

Application of Several Feature Selection Methods in Partial Discharge Pattern Recognition

Partial Discharge Pattern Recognition Based on an Ensembled Simple Convolutional Neural Network and a Quadratic Support Vector Machine

A Methodology to Diagnose Transformer Faults Based on Principal Components Analysis and Artificial Neural Network

Application of Ultra-high-frequency Partial Discharge Detection in Power Transformers

Partial Discharge Defect Recognition in Power Transformer Using Random Forest

Partial Discharge Patterns Recognition with Deep Convolutional Neural Networks

A Novel Partial Discharge Detection Method for Power Transformers on Site Adopting Its Component As Ultra-High Frequency Sensor

Pattern Recognition of UHF Partial Discharge Signals in GIS Based on Deep Learning

Partial Discharge Ultrasonic Signals Pattern Recognition in Transformer Using BSO-SVM Based on Microfiber Coupler Sensor

Study on a Partial Discharge Detection Method for Transformers Adopting the End Shield As UHF Sensor

Partial discharge localisation in transformers using UHF technique: non-iterative method

Application of Hilbert-Huang Transform in Pattern Recognition for Partial Discharge of Transformers

Research of Feature Extraction and Recognition of Partial Discharge Based on the Probabilistic Neural Network

A Novel Method for Extracting Features of Partial Discharge Pulse Current Signal

Identification of Partial Discharge Sources by Feature Extraction from a Signal Conditioning System