Model based building energy optimization using meta-heuristics

Thesis Type: Postgraduate

Institution Of The Thesis: Orta Doğu Teknik Üniversitesi, Faculty of Engineering, Department of Civil Engineering, Turkey

Approval Date: 2015




Energy efficiency plays a key role in minimizing energy usage cost and its environmental impacts. Life cycle thinking guides decision makers to develop energy-efficient solutions in building early design stage; however, in practice, energy analysis is done according to technical specifications’ limits due to inefficient tools and lack of methodologies to response frequent changes in design. Therefore, alternative design solutions with different objectives cannot be generated. In this study, two energy optimization models are developed to solve existing energy analysis problems in practise. In the first model, a graphical user interface called EnrOpt that can be fast and flexible enough to be applied to multiple multi-objective problems and any building types is developed by strengthening weaknesses of practically applied TS 825 Turkish Thermal Standard. The metaheuristics with different position update strategies such as Differential Evolution, Particle Swarm Optimizer and Modified Cross Entropy Method are used to provide a flexible model. In the second model, Dynamo based BIM integrated energy simulation optimization model is proposed. This model offers effective communication between stakeholders to avoid possible problems encountered in early design while providing efficient energy analysis by updating frequent changes in design. Performance of energy optimization models are tested by case studies and Pareto optimal results are obtained. Parametric analysis of design parameters that affect energy model or optimization model on EnrOpt are performed. Results indicates that elaboration in climate and geometric data and energy use scheduling influences building energy estimation significantly. These two models can be applied to different building types by analyzing a vast of alternative designs using different meta-heuristics.