Research And Application Of Operation Parameter Prediction Methods For Power Grid Based On Time Series

The stable operation of power system is an indispensable support and guarantee for national economic development,social order stability and national life quality.At present,the management of power system is becoming more and more intelligent.At present,the management of power system is becoming more and more intelligent.In order to meet the increasing power demand of users,the power system aims at optimizing the operation strategy of power grid,timely troubleshooting in the process of power grid operation,and saving human and material resources.A modern artificial intelligence management system has been gradually established.By forecasting the operation parameters of power system,the efficiency and quality of power system are improved obviously.However,the operation parameters of power grid are complex,and are affected by internal and external factors.This increases the difficulty of power grid operation parameter prediction.Up to now,researchers at home and abroad have put forward many methods to predict the operation parameters of power grid,such as time series method,regression analysis method,artificial neural network and support vector machine.These methods have their unique advantages and shortcomings,which can't guarantee the ideal prediction results under any influencing factors.Therefore,it is still a subject worthy of further study to explore the methods of predicting the operation parameters of power grid and to improve the accuracy of the prediction results.In this paper,the active power,reactive power and current values collected during the operation of the power grid are taken as the research object.By analyzing the data characteristics and the correlation between the parameters,the original parameters collected are processed as the time series form required by the prediction model.In this paper,according to the variation of the operation parameters of the power grid in the fault state,a set of criteria for determining the operation state of the power grid for the experimental lines is designed.Aiming at the traditional power network operation parameter prediction,this paper mainly introduces the basic principle and modeling steps of time series prediction method.Based on the experimental analysis of the actual observation data of power network operation,the ARMA model is selected as the time series prediction model,and the power network operation parameter prediction model is established.The active power data in normal state and fault state are imported separately.According to the accuracy of prediction results,the accuracy and feasibility of traditional time series prediction methods for power grid operation parameters prediction are analyzed.According to the shortcomings of time series forecasting methods,the forecasting methods and models which are more suitable for power grid operation parameters forecasting are studied.In order to solve the problem that the operation parameters of power grid are affected by many factors,this paper presents a prediction method of operation parameters of power grid based on LSTM.According to the basic principle and model characteristics of the LSTM and its applicability to time series,the network structure of input layer,hidden layer and output layer and the optimization parameters and algorithms of the model are designed according to the characteristics of the operation parameters of the power network,and the prediction model of the operation parameters of the power network based on the LSTM is established.The experimental results show that the proposed method has high prediction accuracy and is suitable for the prediction of power network operation parameters under multi-factor conditions.Finally,the parameters prediction results obtained by the time series prediction method and the long and short term memory network based power grid operation parameters prediction method are sorted out,and the decision samples are obtained.According to the criteria designed in this paper,the decision-making framework of power grid operation status is built.To verify the effectiveness and practicability of the power grid operation parameter prediction method,the classification decision of each group of samples is made,and the classification effect and accuracy are compared.Experiments show that the proposed method based on long-term and short-term memory network achieves the expected results,and the method is practical and feasible.