Short-Term Forecasting of Heat Demand of Buildings for Efficient and Optimal Energy Management Based on Integrated Machine Learning Models

摘要

The increasing growth in the energy demand calls for robust actions to design and optimize energy-related assets for efficient and economic energy supply and demand within a smart grid setup. This article proposes a novel integrated machine learning (ML) technique to forecast the heat demand of buildings in a district heating system. The proposed short-term (24h-ahead) heat demand forecasting model is based on the integration of empirical mode decomposition (EMD), imperialistic competitive algorithm (ICA), and support vector machine (SVM). The proposed model also embeds an ML-based feature selection (FS) technique combining binary genetic algorithm and Gaussian process regression to obtain the most important and nonredundant variables that can constitute the input predictor subset to the forecasting model. The model is developed using a two-year (2015-2016) hourly dataset of actual district heat demand obtained from various buildings in the Otaniemi area of Espoo, Finland. Several variables from different domains such as seasonality (calendar), weather, occupancy, and heat demand are used to construct the initial feature space for FS process. Short-term forecasting models are also implemented using the Persistence approach as a reference and other eight ML approaches: artificial neural network (ANN), genetic algorithm combined with ANN (GA-ANN), ICA-ANN, SVM, GA-SVM, ICA-SVM, EMD-GA-ANN, and EMD-ICA-ANN. The performance of the proposed EMD-ICA-SVM-based forecasting model is tested using an out-of-sample one-year (2017) hourly dataset of district heat consumption of various building types. Comparative analysis of the forecasting performance of the models was performed. The obtained results demonstrate that the devised model forecasts the heat demand with improved performance evaluated using various accuracy metrics. Moreover, the devised model achieves outperformed forecasting accuracy enhancement, compared to the other nine evaluated models.