Optimized load vector regression for load prediction and improvement using trombe walls in household electrical energy consumption

Abstract

In many countries, residential energy consumption constitutes a significant portion of total energy usage, making it a crucial focus for power systems and urban planners. Addressing energy consumption in buildings involves two primary facets: accurately predicting and optimizing load demand. This research aims to address the challenges of load demand prediction in the context of building energy consumption. It introduces innovative approaches, including optimized support vector regression (SVR), temperature factor consideration, and the integration of Trombe walls (TW), ultimately contributing to more accurate load demand forecasts and enhanced energy efficiency. To enhance the precision of load demand prediction, we introduce a novel variable that quantifies deviation from the ideal temperature; a key factor in energy usage. This innovative temperature factor plays a pivotal role in forecasting the load demand more accurately. Leveraging this novel approach, we employ an optimized (SVR) that considers weather conditions. The parameters of a radial basis function kernel-based support vector regression (RBF-SVR) method are fine-tuned through an improved particle swarm optimization algorithm (IPSO). In addition, installed TW prove highly effective in reducing building energy consumption by harnessing and redistributing heat, consequently improving the load demand profile. We employ mathematical models to analyze the impact of Trombe walls on predicted load demands, demonstrating that our proposed method yields low mean absolute percentage error (MAPE) when applied to sample buildings. The findings reveal that our method significantly enhances the accuracy of energy usage prediction, while the installation of Trombe walls results in a remarkable reduction of load demand – by 19.32% in winter and 16.24% in summer, thereby promoting energy efficiency and sustainability.