Analysis of start-up tests of household refrigeration systems. A potential for on-line predictions of test results

Abstract

The performance of household refrigeration systems is usually evaluated throughout experimental tests carried out at temperature and humidity controlled chambers so as to measure compressor power, discharge and suction pressures and temperatures in several system positions. These tests are expensive and time-demanding, e.g., a single start-up experiment can take more than 12 h to be performed. Although mathematical models have been proposed for decades as alternative to experiments they are not sufficiently reliable to substitute completely an experimental test. On the contrary, the current work puts forward a semi-empirical mathematical model to predict system performance while test is running. The approach is based on energy balance equations whereas the model parameters, such as conductance and capacitances, are fit to match previously measured data. As soon as the parameters are obtained, a simulation is performed to forecast future values of temperature, pressure and compressor power and therefore, to anticipate the final test results. As the test evolves, fittings and simulations can be continuously repeated. Start-up tests carried out in a 300-l single-door vertical freezer designed for the South-American market were employed for model evaluation. The results show that the proposed approach can predict fairly well the results within the calibration period and that the model accuracy increases with the calibration time. Despite the discrepancies between measured and computed values being outside the experimental uncertainty bounds, the model is quite promising as the differences reduce with the extension of the calibration period. However, the evaporating pressure was found to be the model bottleneck so as to increase the model accuracy the prediction of this variable has to be enhanced. Additionally, the model calibrated for a complete start-up test was used to forecast the system performance in different ambient temperatures and also to evaluate the compressor replacement.