Modelo para predição de resultados de ensaios de sistemas de refrigeração em tempo real

Abstract

The performance of household refrigeration systems are usually evaluated through experimental tests carried at in temperature and humidity controlled chambers. During the tests, the discharge and suction pressures, the temperature in several system positions, and the compressor power are measured. These tests are expensive and time-demanding, e.g., a single pull-down test can take more than 24 hours to be performed. Although the mathematical models have been proposed for decades as an alternative to the experiments, they are not sufficiently reliable to substitute completely the tests. Therefore, the current work proposes a semiempirical mathematical model to predict the system performance with the purpose of reducing the test time instead of replacing it. The model is based on the mass and energy conservation equations in which the constant parameters, such as conductance and capacitances, are calibrated from previous measured values of temperature and pressure. 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 end of the test. Calibrations and simulations can be continuously performed as the test evolves. Preliminary results show that steadystate values of discharge and suction pressures can be predicted within error bands of 5 and 10%, respectively, after only two hours of a pull-down test being performed.