Prediction method of single-phase grounded capacitance current in coal mine power grid based on model migration and SVM
In response to the challenges of low prediction accuracy and difficulties in implementing in-telligent prediction for the capacitive current of the single-phase grounded electrical network in the con-text of intelligent mining,this paper proposes an intelligent prediction method for the capacitive current of single-phase grounded electrical networks.For the problem of incomplete capacitance current data in intelligent prediction,the model transfer is used to expand the sample data of rubber sheathing flexible cable used in mining.Based on Sobol'sensitivity method,the key factors affecting the single-phase grounding capacitance current and their interaction are analyzed,and the input characteristic of the in-telligent prediction model is determined.And the support vector machine based on sparse technology is used to establish the intelligent prediction model of single-phase grounding capacitor current in coal mine power grid,and whale algorithm is introduced to optimize the prediction model's hyperparame-ters,which overcomes the shortage of small sample size of capacitor current data.The proposed method is experimentally tested using test data from manufacturers and field measurements from multiple coal mine electrical networks.The results indicate that the outer diameter and inner diameter of the cable insulation layer are the key factors affecting the single-phase grounding capacitance current.The aver-age error of the proposed method is 2.26％,which is 34.19％,24.91％and 7.40％lower than that of the existing method.The method presented in this paper can effectively improve the prediction accuracy of single-phase grounded capacitance current in coal mine power grid,and provide a new idea for its in-telligent prediction.
coal mine power networkcapacitance currentmodel migrationsupport vector machinewhale algorithm
万方数据
维普
