Research On Large Data State Monitoring Of Substation Equipment

With the increasing dependence and demand of various industries on electric power,it is becoming more and more important for the power grid to operate reliably.The normal and stable operation of the power grid is determined by the health of the power equipment,especially the health of the transformer.Therefore,improving the operation stability of the existing substation equipment is the main measure to improve the operation reliability of the power grid.Generally,the management staff maintain the stable operation of the substation equipment by means of regular inspection and maintenance,but the efficiency is relatively low and the equipment fault detection is not timely.Therefore,it's needed to monitor the condition of substation equipment.By analyzing the state data of the substation equipment through the state monitoring,further evaluation the status of substation equipment,improvement of the vigilance about abnormalities,and the timely maintenance the equipment timely and the reduction of losses caused by delay all become probable.Through the analysis of the deep learning method,this paper selected a set of condition detection value of substation equipment,designed the index system of monitoring data of transformers and routing status.A feature reduction method of state monitoring index based on self coding network is put forward.The state value of monitoring data of substation equipment is dimensionality reduction by using the automatic encoder method.Experiments prove that the method of feature reduction using automatic encoder is effective and efficient,which improves efficiency and saves time for the later stage state classification model.Meanwhile,a state monitoring classification model based on adaptive convolution neural network is proposed.The feature vectors generated by the dimension reduction generated by the automatic encoder are used as input of the network model,and the state classification of the substation equipment is realized through the volume integral classifier.The proposed model is compared with the simulation experiment,and the accuracy of the equipment state classification is higher than that of the other classification models.At the same time,transformer and circuit breaker.The experiment of condition monitoring of electrical equipment based on the model,the experimental results prove the correctness and validity of feature reduction method based on auto encoder,also proved the effectiveness of the convolutional neural network classification model based on state monitoring.