A retrofit decision support approach for improving energy efficiency and indoor environmental quality in buildings

Abstract

Retrofitting of existing buildings offers significant opportunities for reducing global energy consumption and greenhouse gas emissions. This is being considered as one of the main approaches to achieve sustainability in the built environment at relatively low cost and high uptake rates. Although there are a wide range of retrofit technologies readily available, methods to identify the most suitable set of retrofit actions for particular projects are still a major technical challenge. Such methods can be categorized into two main approaches; models in which alternative retrofit actions are explicitly know a priori and models in which alternative retrofit actions are implicitly defined in the setting of an optimization model. This thesis focuses on using modeling and optimization techniques to assess technology choices in the built environment. Firstly two multi-objective optimization models using a classical optimization technique, namely Tchebycheff technique are developed. The functionality of the proposed models is discussed through the application on a residential building. The results verify the practicability of the approaches and highlight potential problems that may arise. Afterward a multi-objective optimization model based on the Genetic Algorithm Integrating Neural Network (GAINN) approach is developed. The benefits of this approach with respect to the classical optimization models are its rapidity and computational efficiency. This model is used for the optimization of the energy consumption, retrofit cost and thermal comfort in a school building. The results from the optimization show the impact of each objective function on the building’s overall performance after retrofit and more importantly illustrate the trade-off between different objectives. Finally, the proposed methodology highlights the improvements added to the GAINN methodology by use of a multi-objective genetic algorithm.