Using the Principal Component Method to Process Test Results of High-Voltage Circuit Breakers

Introduction. The purpose of the work is to assess the possibility of using the principal component method as an alternative method for processing data from high-voltage circuit breaker test results and classification for predicting their performance. Materials and Methods. The results of tests of 50 high-voltage circuit breakers for 17 parameters were used for the study. The principal component method with preliminary autoscaling and implementation in MATLAB was used to process the experimental data. Results. As a result of the research, a matrix of test input data was formed and its transformation into the principal component space was performed. The matrices of counts and loads were calculated and the proportion of the detected variance was determined. The results were analyzed and visualized in two-dimensional and three-dimensional spaces. Conclusions. It was found that the first six principal components explain 91.9% of the total variance, while the first three components provide 69.1% of the contribution, which is sufficient for further analysis and interpretation. Based on the analysis of the projections of the account matrix, the presence of several groups of objects in the principal component space was revealed.