Validation of Methods for Tuning System Charge Predictions in Unitary Equipment

To simulate the performance of unitary air conditioning and heat pumping equipment at design conditions, the amount of subcooling leaving the condenser is typically specified at design conditions and the required refrigerant charge necessary to achieve that subcooling is computed. For off-design conditions, the charge is fixed at the value determined for the design conditions and the amount of subcooling is estimated. Previous studies have shown that existing models do not predict the effect of off-design charge very well, even when tuned at design conditions. Four factors that cause the inaccuracies in charge prediction are identified in this paper, which are the unaccounted liquid volumes in the condenser, refrigerant dissolved in the compressor lubricant, inaccurate void fraction models, and an inaccurate estimate of the subcooled liquid length. Some of these factors lead to constant errors, while the others lead to errors that change with operating conditions. This paper presents a method for tuning parameters associated with a charge correction equation that requires data for two operating points. The approach uses a semi-empirical solubility relationship to predict the refrigerant dissolved in the lubricant, associates the variable charge errors with the subcooled liquid length, and considers all the other errors as a constant offset. The ability of a system model to predict off-design charge effects is significantly improved through the use of this tuning approach. This was confirmed though comparisons with measured results and predictions obtained from existing tuning approaches The tuning approach was tested for multiple units and different operating conditions.