Determination of Optimum Passive Design Parameters for Industrial Buildings in Different Climate Zones Using an Energy Performance Optimization Model Based on an Artificial Neural Network (ANN) and Particle Swarm Optimization (PSO)

Author:
Open Access
Publisher website
Google Scholar
Abstract
With a focus on reducing building energy consumption, approaches that simultaneously optimize multiple passive design parameters in industrial buildings have received limited attention. Most existing studies tend to examine building geometry or individual design parameters under limited scenarios, underscoring the potential benefits of adopting a comprehensive, multiparameter approach that integrates climate-responsive and sustainable design strategies. This study bridges that gap by systematically optimizing key passive design parameters—building geometry, orientation, window-to-wall ratio (WWR), and glazing type—to minimize energy loads and enhance sustainability across five distinct climate zones. Fifteen different building geometries with equal floor areas and volumes were analyzed, considering fifteen glazing types and multiple orientations varying by 30° increments. DesignBuilder simulations yielded 16,900 results, and due to the inherent challenges in directly optimizing building geometry within simulation environments, the data were restructured to reveal underlying relationships. An Energy Performance Optimization Model, based on a Particle Swarm Optimization (PSO) algorithm integrated with an Artificial Neural Network (ANN), was developed to identify optimal design solutions tailored to specific climatic conditions. The optimization results successfully determined the optimal combinations of building geometry, orientation, WWR, and glazing type to reduce heating and cooling loads, thereby promoting energy efficiency and reducing carbon emissions in industrial buildings. This study offers a practical design solution set and provides architects with actionable recommendations during the early design phase, establishing a machine learning-based framework for achieving sustainable, energy-efficient, and climate-responsive industrial building designs.

This publication has 71 references indexed in Scilit: