Virtual Simulation Evaluation Technology for Transformer Temperature Rise Based on Sensor Data and Machine Learning
A virtual simulation method for transformer temperature rise based on sensor data and machine learning is proposed to address the low efficiency and accuracy of transformer temperature rise simulation in complex environments.Established transient characteristic variables that reflect load changes,such as load transfer coefficient,load duration,and load rate.An improved artificial fish swarm(IAFSA)optimiza-tion SVR parameter method was proposed to describe the complex multidimensional nonlinear relationship between transformer HST and transi-ent characteristic variables.In the experimental stage,the oil immersed 500 kV distribution transformer was used as the test object to analyze the temperature rise test and simulation of the transformer under different load conditions and wind speeds.The results show that the proposed method can effectively fit the transformer HST,and the comprehensive performance of SSE,MAD,MAPE,MSPE,MSE and other indicators is significantly better than the GS-SVR algorithm.The effectiveness and practicality of the proposed algorithm are experimentally demonstra-ted,and the model has broad application prospects.
transformerhot spot temperaturetemperature rise simulationsupport vector regressionartificial fish swarm algorithm
万方数据
维普
