Research On Load Prediction And Optimal Scheduling Of Community Integrated Energy System

The integrated energy system(IES)integrates various energy sources such as cooling,heating,electricity,and gas in the area,adopts advanced electronic information technology to realize the optimal control of various equipment in the system,and utilizes the complementarity of various energy characteristics to improve the energy utilization efficiency of the system reduces the harm to the environment and is an important carrier of the Energy Internet.Compared with the traditional large power grid,the community integrated energy system(CIES)has large load fluctuation,complex energy coupling and high prediction difficulty.How to improve the prediction accuracy of multi-energy load and more effective dispatching strategy has important research value.Firstly,aiming at the characteristics of large load fluctuations and complex energy coupling in the the community integrated energy system,a new framework based on Wide&Deep and ResNet framework was proposed,which adopted Complete Ensemble Empirical Mode Decomposition with Adaptive Noise(CEEMDAN)and Principal Component Analysis(PCA)combined load forecasting method for integrated energy systems.CEEMDAN was used to decompose the cold,heat and electricity,and principal component analysis was used to extract and sort the main influencing factors of the decomposition results.The proposed model consists of two parts: width and depth;the depth part of the model referred to the idea of ResNet,stacked multiple LSTM sub-layers to build a depth prediction network,and realized the cascade processing of data with different information densities;the width part of the model adopted a simple model and improved the input of the Wide part of the traditional Wide&Deep-LSTM model,which effectively reduced the training difficulty of the model.Through the analysis of practical examples,it can be seen that the proposed model has good prediction accuracy and convergence speed.Compared with conventional models,the proposed model has certain advantages.Secondly,aiming at the nonlinear and high latitude problems of the park-level integrated energy system scheduling,a global intraday integrated energy system optimal scheduling model based on continuous action space reinforcement learning is proposed.Combined with the hourly forecast data of multi-energy load and photovoltaic output in the early stage,the optimization goal is to minimize the operating cost in the whole cycle,and the optimal scheduling model is solved by using Deep Deterministic Policy Gradient(DDPG).Through the analysis of two typical daily scheduling results in summer and winter,and the comparison with discrete action space reinforcement learning algorithm and particle swarm optimization(PSO)algorithm,it is shown that the proposed method has certain flexibility and effectiveness.