基于CEEMDAN-CSO-LSTM-MTL的综合能源系统多元负荷预测OA北大核心

With the continuous development of integrated energy systems and the flexible interaction between load side and source side resources,existing single load forecasting methods are difficult to grasp the coupling characteristics between multiple loads,resulting in insufficient accuracy in the prediction of multiple loads in integrated energy systems.Based on this,a comprehensive energy system short-term load forecasting model is proposed,which combines complete ensemble empirical mode decomposition with adaptive noise(CEEMDAN),cross optimization algorithm(CSO),long short term memory(LSTM)network,and multi task learning(MTL).Firstly,preprocess the collected raw load data and calculate the actual load value considering system energy loss;Secondly,the maximum information coefficient(MIC)is used to analyze the correlation between multiple loads and between multiple loads and weather factors,and to extract strongly correlated variables of multiple loads;Once again,the strongly correlated variables of multiple loads are substituted into CEEMDAN,and the load data is decomposed into stationary subsequences;Then,the feature sequence is substituted into the LSTM-MTL shared layer and the CSO algorithm is used to optimize the prediction model,achieving collaborative prediction of multiple loads;Finally,the performance of the constructed model was validated using a multivariate load dataset from a chemical park in Jilin City,Jilin Province,China.The results show that compared with traditional prediction models,the constructed model can effectively improve the prediction accuracy of multiple loads in the integrated energy system.