Decision making of HVAC system using Bayesian Markov chain Monte Carlo method

Abstract

Building simulation has become an indispensable decision making tool since it is capable of capturing dynamic behavior of building systems and predicting impact of energy saving components. However, it has been well acknowledged that simulation prediction is often significantly influenced by treatment of uncertain inputs. This paper presents multi-criteria (construction cost, total energy consumption) decision making of HVAC systems under uncertainty. In this study, a library building was selected and modeled using EnergyPlus 6.0. There were two HVAC candidates: (1) variable air volume (VAV) for interior zone+fan coil unit (FCU) for perimeter zone+gas boiler+electric chiller, (2) VAV+FCU+gas boiler+electric chiller+ice thermal storage system. For uncertainty analysis, unknown inputs were identified based on the literature and the Latin hypercube sampling (LHS) method was employed. Then, Bayesian decision theory was applied to solve stochastic decision making. In particular, the paper includes preferences of building stakeholders (three architects, four simulation experts, three HVAC experts) by using Markov chain Monte Carlo (MCMC). It is shown that such quantitative stochastic appraisal yields more meaningful information than the traditional deterministic approach, and helps to improve confidence in simulation results.