In the process of electrical equipment debugging,the working condition is unstable,and the feature drift is easy to occur in the process of dangerous points,resulting in the difference in the process of dangerous points identification.Therefore,an adaptive identification method of dangerous points in electrical equipment debugging under gradient diffusion is designed.This paper establishes the signal space model during the debugging of electrical equipment,designs the effective value of the e-quipment operation signal in a certain period of time,completes the signal detection of dangerous points,determines the gradi-ent diffusion coefficient,obtains the mathematical model of denoising and filtering to realize signal denoising,prevents the fea-tures of dangerous points from drifting under cross working conditions.This paper also uses ID CNN depth network and deter-mines the convolution kernel size,completes feature extraction,inputs the feature samples into the depth network,and com-pletes the process design of adaptive diagnosis and identification of dangerous points.Setting different dangerous points state for the experiment,and the experimental results show that under cross working conditions,the distribution law of dangerous point samples in different fields is same,which can effectively match the features,and the boundaries between different types of dangerous points are clear.In the process of electrical equipment debugging,the performance of adaptive identification of dan-gerous points is better.
关键词
梯度扩散/电气设备调试/危险点/自适应识别/深度网络/特征漂移/变工况
Key words
gradient diffusion/electrical equipment debugging/dangerous point/adaptive identification/depth network/fea-ture drift/variable working condition
维普
万方数据