Uncertainty and sensitivity analysis of cooling and heating loads for building energy planning

Abstract It has become crucial to investigate the uncertainty and sensitivity of building loads (peak cooling load, peak heating load, annual cooling demand and annual heating demand) for meeting the risk assessment of building energy planning. Therefore, a new Monte Carlo (MC) method based on building performance simulation (BPS) is proposed to solve the problem of building loads forecasting at planning phase. Furthermore, the sensitivity of building loads is examined using two global sensitivity analysis (GSA) methods, including meta modeling method based on tree Gaussian process (TGP) and regression method based on standard regression coefficient (SRC). Finally, a case study of office building is conducted. The results show that the MC method constructed by the combination of R language platform and EnergyPlus software can generate models rapidly and simulate accurately building loads. Note that it is necessary to assess the stability of results as a function of sample size from uncertainty analysis in applying the MC method into building loads assessment. The TGP-based GSA method is applicable to identify and analyze key variables affecting building loads. It is recommended that at least two inherently different GSA methods should be applied to provide robust sensitivity results. Moreover, this study also provides insight on building energy planning and energy conservation design according to the results of uncertainty and sensitivity analysis for case study.