Implementing Surrogate Modeling Techniques for Designing Optimal Building Envelops: A Case Study

Buildings are known to have significant environmental impacts. The life cycle approach for the measurement of CO2 emission and the life cycle costs of buildings are getting more important in the building design process. However, due to the complexity of the design process and the computational time of simulations and data processing, such methods are difficult to implement within optimization processes. This paper aims to apply surrogate modeling techniques as a solution to resolve the computational difficulties in the optimization process of building envelopes. The paper will describe the methods applied and will evaluate several aspects of the process, including the impact of the size of the training set on the prediction accuracy as well as the impact of different energy system efficiencies on the final optimum envelope design concerning seven objectives related to the economic and environmental performance of the building. The results showed that the size of the sampling test has a significant effect on the prediction accuracy; however, a balance between increasing the precision and computational time can be maintained by selecting an adequate number of samples. Moreover, it is found that to achieve the lowest total equivalent cost corresponding to the highest economic and environmental performance of the building, the minimum allowed window-to-wall ratio and the maximum permitted wall insulation thickness should be 0.15 and 0.02 m, respectively. The surrogate model was also shown to be efficiently capable of finding the optimum results according to the other objectives, including both economic and pure environmental aspects. Furthermore, the results provide some insights on how the variation of energy systems’ efficiency might affect the optimum solutions in the optimization process.