Use of genetic algorithms for optimization of thermal energy performance in buildings in early stage design

Author(s)

Felipe da Silva Duarte Lopes

• Document type: Doctoral Thesis
• Press: Campinas, SP.
• Institution: Universidade Estadual de Campinas (UNICAMP). Faculdade de Engenharia Civil, Arquitetura e Urbanismo
• Defense date: 2020-02-20

Examining board members

Lucila Chebel Labaki; Alberto Hernandez Neto; Nathan Mendes; Martin Ordenes Mizgier; Luiz Bueno da Silva

• Advisor: Lucila Chebel Labaki

Abstract

Energy demand in buildings has increased considerably in recent years, and Brazilian buildings already account for half of the country's total electricity consumption. Concerning the environment, building energy regulations have been developed to reduce the impact on climate change. As these standards require quantifying energy use, simulation tools can assist architects in decision-making in the early design stages of high-performance buildings to improve environmental, energy and economic aspects. There are several variables to consider in architectural design, and in this context, optimization methods can generate many alternatives to solve a problem, obtaining the most appropriate solutions based on conflicting criteria, such as energy consumption cost during the building’s life-cycle and thermal discomfort hours. Genetic algorithms, for example, are optimization methods based on the natural selection of living organisms that optimize evaluation functions from diversity operators. This research aimed to evaluate the application of a method with genetic algorithms to find optimal solutions in the early design stages based on energy efficiency, life cycle cost and thermal comfort criteria. The work intends to propose an optimization procedure as an improvement of the current Regulation for Energy Efficiency Labeling of Commercial, Services and Public Buildings (RTQ-C). The thesis methodology is based on an exploratory research applied to case studies. A systematic literature review was developed on genetic algorithms in efficiency and comfort studies, supporting building performance simulation studies with EnergyPlus software. The validation study applied a multi-objective genetic algorithm for a medium-size office building in the city of São Paulo, using passive strategies to minimize the initial construction cost and the life cycle energy cost. 213 cases were simulated and a reduction of 6.7% in construction cost and 5.8% in energy cost were observed when compared to the results of the base case. In the second case, the optimization procedure coupled a non-dominated and crowding distance sorting genetic algorithm (NSGA-II) method with EnergyPlus to reduce the life cycle cost and hours of thermal discomfort in another office building for three cities in different Brazilian bioclimatic regions. The design variables were divided into geometry, envelope and air conditioning system and the optimization was run in jEPlus+EA engine. The results demonstrated a potential reduction of 11% in life cycle cost and up to 37% in discomfort hours. The results of the work suggest that the use of genetic algorithms has great potential to contribute to the Brazilian energy efficiency standards, generating more economical, energy efficient and high environmental quality architectural projects in different bioclimatic regions of the country (AU)

• FAPESP's process: 16/21667-1 - Method for evaluating the thermal energy performance in office buildings with genetic algorithms
• Grantee: Felipe da Silva Duarte Lopes
• Support Opportunities: Scholarships in Brazil - Doctorate