An online robust sequencing control strategy for identical chillers using a probabilistic approach concerning flow measurement uncertainties

Abstract

Chiller sequencing control is crucial to the reliable and energy-efficient operation of multiple-chiller plants. It should not only ensure an adequate supply of cooling capacity for buildings, but also make the chillers work efficiently. The measurement uncertainties cannot be avoided and have significant negative effects on the chiller sequencing control. To cope with the challenges and uncertainties, this study proposed an online robust sequencing control strategy using a probabilistic approach for chiller plants under low-quality and uncertain flow measurements. An uncertainty processing model of flow measurements was developed based on the Bayesian inference and Markov chain Monte Carlo methods and an energy balance model. As the core of the proposed control strategy, the uncertainty processing model can quantify the measurement uncertainties of water flow rates accurately. According to the probability analysis, an online decision-making scheme was designed, and the risks in the online decision-making processes were assessed. Compared with the conventional chiller sequencing control strategies, the proposed control strategy could reduce the impacts of flow measurement uncertainties significantly. The results of case studies showed that the root-mean-square error of cooling loads was reduced significantly by about 79%, the total switching number of chillers was reduced by up to 35.71% under the positive flow measurement uncertainties, and the cumulative unmet cooling load was reduced by up to 31.22% under the negative flow measurement uncertainties. The proposed chiller sequencing control strategy is able to tolerate flow measurement uncertainties.