Optimizing the thermostat setting points of residential and insulated buildings in the direction of economic efficiency and thermal comfort through advanced multi-purpose technique

Abstract

The present research work develops a new approach for the optimization of thermostat setting and insulation designs in residential buildings located in various Iranian climates, including hot-humid, arid, temperate, and cool regions. The objective functions are set to minimize the construction cost, consumed electricity cost, and PPD to improve thermal comfort. Advanced computational techniques are integrated in a structured way to achieve the mentioned objectives. Numerical modeling is done through the simulation of building energy performance and thermal comfort using EnergyPlus. The exact mathematical relations between design variables and objective functions, which were heating setpoint and cooling setpoint, insulation thickness, and thermal conductivity, were identified using Multi-Polynomial Regression. MPR model has been validated respect to a wide set of statistical measures that included but were not limited to R², RMSE, and MAE for its high predictive accuracy. Then, multi-objective optimization is performed through NSGA-II, a well-known multi-objective optimization algorithm, which provides a Pareto front of optimal solutions balancing energy efficiency, cost, and comfort. Shannon's entropy method assigns weights to the Pareto-optimal solutions, whereas the Technique for Order of Preference by Similarity to the Ideal Solution (TOPSIS) selects the most suitable configurations for each city. Calculations show a great reduction in energy consumption to up to 82.66% at Bandar Abbas, with very important improvements in comfort, where the PPD is reduced between 31.1% to 56.3%. The predictive capacity of the MPR model was confirmed by this study, from the value of R², close to 1. The cost-effectiveness of the proposed solutions is underlined by minimizing construction and energy costs while preserving occupant comfort. This innovative approach adapts optimization strategies to regional climatic characteristics, providing practical solutions for sustainable and cost-effective building designs. The integration of advanced machine learning and genetic algorithms offers a scalable framework for future energy-efficient construction practices worldwide, contributing to reduced carbon footprints and enhanced occupant well-being. By addressing the limitations of previous studies and introducing a clear, structured methodology, this research provides valuable insights and practical tools for optimizing residential building performance in diverse climates.