Reliability Analysis of an Energy-Based Form Optimization of Office Buildings under Uncertainties in Envelope and Occupant Parameters
Shuai Lu,Jingyu Li,Borong Lin
DOI: https://doi.org/10.1016/j.enbuild.2019.109707
IF: 7.201
2019-01-01
Energy and Buildings
Abstract:Building performance optimization is effective in finding optimal design solutions and improving the building energy efficiency, but the reliability of its results can be affected by uncertainties in the input parameters. In existing research, the reliability of building form optimization, the influence of uncertainties caused by the design procedures, and the influence of individual uncertain parameters have yet to be thoroughly explored, and thus should be better addressed. In this study, a reliability analysis is conducted on an energy-based form optimization of office buildings under uncertainties in the envelope and occupancy parameters. An optimization process involving Rhinoceros, EnergyPlus, and the genetic algorithm is first implemented. Then, parametric studies of 644 configurations involving 4 cities in different climates and 3 form variables are conducted on a medium-sized office building. The results indicate that the uncertainties in the input parameters can lead to major unreliability of the optimization results, including reductions up to 13% in energy saving achieved by optimization, decreases up to 10% in energy efficiency compared with the results before optimization, and large deviations in the optimized forms. Moreover, it was found that the uncertainty in the visual transmittance of windows is the most significant cause for the unreliability, followed by the U-value of walls, whereas the uncertainties in the occupant density and occupant schedule have a limited influence. It was also found that the impacts vary by locations and the form variables for optimization. Finally, a stochastic optimization method was raised to acquire the overall optimal design under the presence of uncertainties that reduces the risk of getting very poor performance under extreme conditions and decreases the performance dispersion in various scenarios.