Empirical modeling of the impacts of faults on water-cooled chiller power consumption for use in building simulation programs

Abstract

Empirical models of four chiller faults that can be applied within existing building models to study overall impacts are developed in this paper. The faults include overcharging, excess oil, non-condensables in refrigerant and water-side condenser fouling. A single generalized model structured was developed for these faults that forces predicted fault impacts to be zero with no fault and increase with increasing fault level. The models were trained and tested using available laboratory data for a water-cooled centrifugal chiller where all four faults were artificially introduced. The fault model behavior was studied and then they were integrated in hospital models from DOE commercial reference building models (Deru et al., 2011) and simulations were performed in different climates. The simulation results showed maximum increases of building electricity consumption, electricity peak demand and water consumption of the hospitals due to faults of 4.7%, 7.8% and 1.8% respectively. The fault impacts were found to be more severe in hotter and more humid climates.