Spatial Load Situation Awareness And Situation Orientation Based On DAE-SSA-LSTM

Spatial load situation awareness is the basic work for the lean planning of urban distribution networks.The result is not only the size of the future cell load,but also the spatial distribution of the future cell load.The spatial load situation awareness results can provide basis for the new construction and transformation of power equipment in the distribution network.Therefore,how to improve the spatial load situation awareness accuracy has important guiding value for the lean planning of urban distribution networks.This article first introduces the research background of spatial load situation awareness and the current research status of situation awareness in power systems at home and abroad,and briefly describes and summarizes the existing spatial load forecasting methods;The characteristics of spatial load are analyzed on the basis of the historical cellular load data;To establish the power geographic information system for spatial load situation awareness.Secondly,in order to make full use of historical cell load data to obtain accurate spatial load situation awareness results,a spatial load situation awareness method based on denoising autoencoder,singular spectrum analysis and long-short term memory neural networks(DAE-SSA-LSTM)is proposed.First,in the situation perception stage,the noise reduction autoencoder is used to encode the measured data of each type I cell load to extract their main load change characteristics,and reconstruct the historical cell load data according to the characteristics to reducing noise caused by measurement,communication and other reasons;Then in the situation comprehension stage,the singular spectrum analysis method is used to decompose the cell load data after situation perception to obtain a strong periodic low-frequency component sequence and a strong random high-frequency component sequence;Finally,in the situation forecasts stage,different long-short term memory neural networks models are used to predict the low-frequency component and the high-frequency component respectively,and the two prediction results are superimposed to obtain the predicted value of the class I cells load in the target year.On this basis the spatial load grid technology is used to obtain the predicted value of the spatial load based on the class Ⅱ cells.The engineering example verifies the effectiveness and accuracy of the method.Finally,in order to consider the impact of demand response on spatial load situation awareness results in the future power market,a spatial load situation guidance method based on demand response is proposed.According to the classification of flexible loads,a price-based flexible load responsiveness model based on consumer psychology principles is established.At the same time,fuzzy parameters are used to express the uncertainty of user participation in demand response;Use the Logit discrete choice model to simulate and determine the demand response participation of residential users and commercial users in the region under different electricity price mechanisms;Based on the DAE-SSA-LSTM spatial load situation awareness results,considering that different types of load users have different degrees of participation in demand response measures and the geographical location distribution of users who participate in demand response measures,the spatial load situation awareness results after users are affected by demand response measures in the power market are obtained,so as to reduce the power equipment capacity needed for distribution network planning.The engineering example verifies the effectiveness and practicability of the method.