JAUE2020-063: Optimization of BCHP using two-stage stochastic programming considering energy demand uncertainty described by Monte-Carlo Building Performance Simulation

Abstract

The uncertainty of energy demand is a key issue in the design of Combined Cooling, Heating and Power (CCHP) systems. The design of CCHP under uncertainty is generally solved by stochastic programming, in which the energy demand uncertainty is described by probabilistic scenarios. In this paper, considering the uncertainty of energy demand, an optimization model of CCHP, a two-stage stochastic programming model, is presented. In this model, the uncertainty of energy demand is characterized using scenarios generated by Monte-Carlo Building Performance Simulation (MCBPS), which can maintain the temporal autocorrelation of energy demand. Moreover, a feature-based k-Medoids clustering algorithm is applied for scenario reduction. A case study of an office building in Shanghai is investigated, and both deterministic and stochastic models are applied for comparison. Results indicate that the annual total cost of the stochastic model is higher than that of the deterministic model. The optimal CCHP sizes of the stochastic model are often higher than those of the deterministic model, especially the capacity of the auxiliary boiler (AB) and electrical chiller (EC). The deterministic optimization model may underestimate the capacity of each device in the CCHP system and make an optimistic estimation of the system design, which will introduce potential risks to the system operation.