Short Term Load Forecasting Method for Typical Distribution Stations Based on LSTM+Attention Model
In order to solve the problem of uneven load scheduling caused by the difficulty of short-term load forecasting in dis-tribution areas,a short-term load forecasting method for typical distribution areas based on LSTM+Attention model(short long-term memory network+Attention mechanism model)is studied.It considers the short-term load changes caused by dif-ferent change factors such as holidays,normal work and rest and temperature,determines the relationship between each load and influencing factors in the regional division unit,analyzes the load sampling data,counts the number of influencing factors of load changes according to the contribution rate of different component eigenvectors,and uses the long short-term memory net-work to fully store historical information.Based on the Attention mechanism model,the input vector,the actual state value of the hidden layer in the network and the main parameters of the regional load are determined,and the actual probability density of the normalized power load is calculated to obtain the short term load forecasting results.The experimental results show that the proposed method has fast response speed,high prediction accuracy,and ideal prediction effect and goodness of fit.
STM networkAttention modelshort term loadload forecastingpower distribution information
万方数据
维普
