基于CatBoost算法的空载变压器励磁涌流预测OA

The excitation inrush generated when the no-load transformer is closed may appear to cause equipment failure,affect the power quality and lead to line overload and other problems,at this stage,phase-controlled opening and closing is an effective means of suppressing the inrush,the circuit breaker closing characteristics and remanent magnetism are the main factors affecting the accuracy of phase control.The influence of switching characteristics of circuit breakers on inrush current was described,and considered the unevenness of the electric field changes in the arc extinguishing chamber during closing,and the pre-breakdown time corresponding to different switching coefficients k was calculated.The influence mechanism of residual flux on inrush current was analyzed.Statistical simulation analysis of inrush current under different residual flux conditions was made by using ATP/EMTP.A method was proposed for predicting excitation inrush current in no-load transformers using Categorical Boosting(CatBoost)algorithm.Based on CatBoost model,the mapping relationship between breaker closing coefficient k,mechanical dispersion 3σ,residual flux Φr and inrush current was established.This model can predict the inrush current under different phase control conditions,and analyze the performance parameters of circuit breakers under different inrush current requirements.Finally,the effectiveness of the CatBoost model was verified through experiments,and by comparing it with other traditional algorithms,it was confirmed the accurate of its predictions.