Pattern Recognition for Partial Discharge Using Multi-Feature Combination Adaptive Boost Classification Model

Rui Yao,Jun Li,Meng Hui,Lin Bai,Qisheng Wu

DOI: https://doi.org/10.1109/access.2021.3067009

IF: 3.9

2021-01-01

IEEE Access

Abstract:This paper proposes a multi-feature combination adaptive boost classification model, considering the difference and complementarity of the three different single feature sets for partial discharge pattern recognition. First, eight types of physical models are designed. Then, an UHF measurement system is used to collect partial discharge data. Second, three kinds of single feature sets extracted from the Phase Resolved Pulse Sequence (PRPS) data are combined with pairs and three to construct new feature sets. The final optimal feature set is selected from the single feature set and the combined feature set as the input of the classification model. Finally, using the boosting algorithm in combination learning to process the training data set, taking the support vector machine as the base classifier, and measuring the inconsistency between one base classifier and other base classifiers by using the "unpaired" diversity index based on information entropy. By this method, a series of various SVM-based classifiers with moderate accuracy are obtained, and finally an adaptive boost classification model based on the multi-feature combination method is obtained. For each defect, 25 samples were obtained at the same test voltage level, and a total of 150 samples were obtained at 6 voltage levels through multiple experiments. The proposed method was compared with the traditional methods using these data sets. The results revealed that the proposed method successfully identified the types of partial discharge insulation defects.

computer science, information systems,telecommunications,engineering, electrical & electronic

What problem does this paper attempt to address?

The paper attempts to address the issue of insufficient classification accuracy in partial discharge pattern recognition. Specifically, existing partial discharge classification methods mainly rely on a single feature set, which leads to a low recognition rate for different defect types. To solve this problem, the paper proposes a multi-feature combination adaptive boost classification model, which improves the recognition rate of partial discharge defect types by combining multiple feature sets and ensemble learning methods.

### Main Issues:
1. **Limitations of Single Feature Set**: Existing methods usually use a single feature set for partial discharge pattern recognition, but a single feature set can only describe partial discharge characteristics from one perspective and cannot fully reflect all information.
2. **Balance Between Classifier Diversity and Accuracy**: In ensemble learning, it is a challenge to improve overall classification accuracy while maintaining classifier diversity.

### Solutions:
1. **Multi-Feature Set Combination**: The paper designs 8 physical models and uses a UHF measurement system to collect partial discharge data. Three single feature sets are extracted from phase-resolved pulse sequence (PRPS) data, and then these feature sets are combined in pairs or threes to construct new feature sets.
2. **Adaptive Boosting Algorithm**: Support Vector Machine (SVM) is used as the basic classifier, and the training dataset is processed through the Adaptive Boosting (AdaBoost) algorithm. By adjusting the parameter σ of the Gaussian kernel function, it ensures that the generated SVM classifiers have moderate accuracy and diversity.
3. **Feature Selection and Dimensionality Reduction**: Principal Component Analysis (PCA) is used to reduce the dimensionality of single feature sets and combined feature sets to reduce feature dimensions and improve computational efficiency.

### Experimental Verification:
1. **Dataset Acquisition**: 150 sample data were obtained through experiments at 6 voltage levels.
2. **Model Training and Testing**: Comparative experiments were conducted using SVM, BP neural network, RBF neural network, K-Nearest Neighbors (KNN), and Classification and Regression Tree (CART) methods to verify the effectiveness of the proposed method.
3. **Result Analysis**: The experimental results show that the proposed multi-feature combination adaptive boost classification model can successfully identify different types of partial discharge insulation defects, and its classification performance is better than traditional methods.

In conclusion, the paper effectively improves the accuracy and robustness of partial discharge pattern recognition by combining multiple feature sets and ensemble learning methods.