Research on Transformer Oil-Paper Insulation State Diagnosis and Life Prediction Method Based on Shuffle-SVM

To monitor the insulation state of the transformer and predict the remaining life, this paper improves the support vector machine (SVM) algorithm called Shuffle-SVM, and carries out the time-domain dielectric response test (RVM) to obtain the test parameters required by the model. Combined with the environmental variables, the prediction of the remaining life of the transformer is completed through the calculation of the degree of polymerization (DP). The moisture absorption aging test platform and recovery voltage test platform of oil-paper insulation are designed and built; The time-domain dielectric response and physicochemical properties of oil-paper insulation samples with different micro water content, different aging degrees, different test temperatures, and different test voltage were studied.