Abstract
A proper refrignerant charge amount(RCA) plays an important role in the effective operation of the heat pump systems and contributes positively to reduction of energy consumption. With the commercialization of heat pump systems, predicting RCA is of great necessity for ultimate operation performance. In recent years, data-driven artificial neural network(ANN) techniques have been used for efficient prediction of RCA. In this study, a deep learning-based RCA prediction model was developed using the variables measured under various conditions in a heat pump syste. The developed predictive model has the reliable root mean square error (RMSE) and the coefficient of determination (R2) value of 0.9571. in addition, the predicted values were estimated to be fairly close to the measured values, and statistical significance was also confirmed.
