An integrated energy system load prediction study based on deep belief networks and multitasking learning

Load forecasting is an important research content in the field of integrated energy systems, improving the accuracy of load forecast results, and helping to improve the economics of the planning and operation of integrated energy systems. This paper puts forward a comprehensive energy system load prediction method based on deep belief network and multi-task learning, first of all, based on the comprehensive energy system general planning model, using Pearson coefficient quantitative calculation of the correlation between multiple loads, analysis of the common influencing factors in the load prediction process; The results show that the pre-test model presented in this paper has good prediction accuracy by validating the prediction method by study.