Design And Optimization Scheduling Of Home Energy Management System Based On Predictive Methods

In recent years,as traditional fossil energy resources are depleting,the country has begun to focus on and actively develop renewable energy generation,and as a result,the importance of distributed generation technology is becoming increasingly apparent.With the rapid development of China’s economy,distributed power sources are gradually becoming more commonplace,and smart homes are widely used in households with the development of technology.The emergence of home energy management systems can rationally distribute energy to reduce users’ electricity costs and improve their electricity comfort.This article proposes a predictive home energy management system that primarily targets electricity costs,with clean energy utilization and user comfort as secondary targets.The system manages demand-side from the perspective of prediction algorithms,forecasting distributed generation power and rigid load electricity,and monitoring and controlling energy in the home through scheduling algorithms based on predictive information.The specific work is as follows:First,predict the next day’s power generation of the household’s distributed photovoltaic power generation system.This article improves the atomic search algorithm through logistic chaos mapping and applies the improved model to optimize the parameters of the backpropagation neural network.The optimized model predicts the next day’s power generation of the distributed generation equipment,and this article compares the prediction results of the proposed model with those of the original backpropagation neural network,the backpropagation neural network improved by genetic algorithm,and the backpropagation neural network improved by particle swarm optimization algorithm to verify the superior performance of the proposed improved algorithm over traditional algorithms and further improve the accuracy of photovoltaic power generation prediction.Second,predict the electricity consumption of the household’s rigid load for the next day.This article decomposes the original load data by empirical mode decomposition and inputs the decomposed components into the extreme learning machine prediction model.The whale optimization algorithm is used to optimize the prediction model parameters,and a combination model of online prediction and offline optimization is designed,taking into account the issue of changes in user electricity consumption habits.Finally,the combination model is established,and the original extreme learning machine and the extreme learning machine optimized by whale algorithm are used to compare the prediction results to verify the superior performance of the proposed load prediction combination model.Finally,model the flexible load and energy storage devices in the household and study the optimization scheduling strategy for electricity consumption,targeting the lowest user cost and the highest satisfaction with new energy output.Based on whether users have energy storage,the system designs scheduling optimization strategies that include battery storage and those that do not.A whale optimization algorithm based on Levy flight and chaotic opposition learning strategy is established to optimize the model.The system schedules flexible loads based on time-of-use electricity prices and excess solar power,achieving maximum user benefits while ensuring the utilization of clean energy and the comfort of electricity consumption.The effectiveness of the scheduling model is verified through comparative cases.