Research On Short-term Power Load Forecasting And Optimal Dispatch Of Microgrid

Microgrid load forecasting is to estimate power demand in advance through the analysis and research of historical data.In recent years,my country’s economy has developed rapidly,and people’s lives have gradually become richer.Therefore,the demand for electric energy and the quality of electricity are getting higher and higher,Quick and accurate forecasting of power load provides an important basis for power market decision making and scheduling optimization of power networks.With the gradual increase in energy demand in countries around the world,the pressure on resources and the environment become more and more serious.In order to reduce energy consumption and protect the ecological environment,the proportion of clean energy such as wind power and photovoltaics has increased year by year.Due to the output of distributed power since the power generation time of wind power,photovoltaic power generation and other power generation units is related to the weather,there will be uncertain safety problems when connected to the large power grid without any treatment.Therefore,in order to reduce traditional energy consumption,reduce environmental pollution,and ensure power supply safety,it is very necessary to carry out research on optimized operation of microgrids.This paper conducts in-depth research on short-term power system load forecasting and optimized operation of microgrids.The main research contents are as follows:(1)This paper proposes a short-term power load forecasting method that optimizes the bidirectional long and short-term memory network with the whale algorithm combined with the attention mechanism.This method uses the WOA optimization algorithm to optimize the Bi-LSTM model parameters,and introduces the attention mechanism to calculate the contribution rate of various feature quantities,assigns different weights to the bidirectional LSTM hidden layer,highlights the impact of important features,and completes the load.predict.It solves the problems of the recurrent neural network that it is difficult to effectively extract the potential high-dimensional information in the sequence and the long time sequence causes the information to be easily lost.(2)In order to improve the prediction accuracy of the model proposed above,considering the influence of factors other than historical load data on short-term load forecasting,a load forecasting model based on WD-CNN and SVR is proposed.First,the model analyzes and reconstructs historical load data through the WD decomposition algorithm,which reduces the impact of non-stationarity in the original sequence on the prediction accuracy;then,it constructs the characteristics of weather factors and date types;finally,the processed The load data and weather data are input into the CNN-SVR model and the BP-SVR model respectively,and the final prediction value is obtained through the superposition of the results of the two models.(3)To solve the problems of long-term dispatching time and poor economy of micro-grid,a micro-grid optimization dispatching model that comprehensively considers operating costs and environmental protection costs is proposed.First,the combination method of genetic algorithm and nonlinear programming is used to solve the model,and then the microgrid operation cost and environmental governance cost are minimized target,and the constraints of the actual operation of the microgrid are considered to optimize the microgrid dispatching model.(4)Developed an intelligent micro-grid short-term power load forecasting software.Through the design and development of PyQt5 and Qt Designer software,the interface design,data management,model training and result display of the power load forecasting system have been completed.