태양광열-공기열원 하이브리드 히트펌프 축열조 온도 및 에너지 사용량 예측모델 개발

Abstract

Purpose: This study aims to develop a predictive model for optimal control of a hybrid heat pump system using artificial neural network(ANN). The developed prediction model predicts the temperature of the heat storage tank and system energy consumption, and through this, it aims to improve system efficiency. Method: the target building was modeled using TRNSYS 18 simulation program, and data for 1 month for each heating and cooling were acquired to build a dataset for training and performance evaluating of the predictive model. Prediction models consists of three temperature prediction models and two energy consumption prediction models according to the heat sources used during the heating and cooling period. Result: All prediction models for the heating and cooling period showed satisfying performance in accordance to the accuracy criteria presented by ASHRAE. In addition, as a result of analyzing the error between the predicted value and the actual value, the stability of the prediction model demonstrated very low error value. In the future, the prediction model developed through this study will be applied to the optimal control algorithm to conduct demonstrative experiments.