Research On Optimal Scheduling Algorithm Of Home Energy Management System Base On PV Electricity And Load Forecasting

Energy problem has always been one of the most important topics about the human's development in the future.It is not only the base of scientific and technological progress but also the lifeblood of our economic development.China is vast in territory and rich in renewable resources.With the rapid development and widely attention of the renewable energy technologies in our country,more and more distributed photovoltaic power generation equipment is used in the average family.Many users have raised the issue of how to effectively use distributed generation equipment in the home to reduce grid pressure and how to improve energy efficiency to reduce electricity bills.In order to solve those problems and use all the energy more efficiently in the home,a kind of home energy management system based on photovoltaic power generation and load forecasting was used.In this thesis,the innovation of home energy management system mainly focuses on two aspects: system control terminal and system algorithm,and the prediction algorithm and scheduling algorithm are included in the algorithm aspect.In order to help users to better plan the load use solution in the home,the control terminal of the home energy management system which based on Qt integrated development environment is designed to manage other functional modules in the home energy management system and dispatch the load in the family,and the user management function be realized in the control terminal.What's more,the energy management module is designed in the system to control the load and acquire electric energy data.In order to make the scheduling result more accurate and reliable,the modified BP neural network optimized by genetic algorithm and k-nearest neighbor algorithm is proposed to predict the power which the distributed photovoltaic electricity generation equipment generate electricity per half hour.Moreover,the online learning long-term and short-term memory(LSTM)neural network is designed to forecast power that used by the uncontrol load in the home.the modified discrete binary particle swarm optimization(MDBPSO)algorithm is proposed in this paper to schedule the load which can be control in the family,furthermore the algorithm and scheduling strategy in this thesis are simulated and verified using Matlab software.according to the results,the merits and drawbacks of the algorithms and the advantages of the scheduling strategy are analyzed.