ANN based solar thermal energy forecasting model and its heating energy saving effect through thermal storage

Abstract

Among various types of renewable energy sources, solar thermal energy can be economically utilized for space heating and service hot water. However, its performance is significantly influenced by the orientation and location of panels, as well as weather conditions. Therefore, accurate prediction of solar thermal systems is essential to ensure the reliability and stability of the technology. This study aims to predict solar thermal energy production using Artificial Neural Networks (ANN) and to apply the technology integrated with thermal storage for energy savings. A predictive model was developed using field data collected from solar thermal collectors in the community complex from August 1, 2019, to July 31, 2020. The performance of the predictive model was evaluated using Cv(RMSE), NMBE, and R² indexes as recommended by ASHRAE Guideline 14-2014. The accuracy of the hourly data predictive model showed ANN prediction results of Cv(RMSE) = 11.7 %, NMBE = −1.21 %, R² = 0.93. To evaluate the heating energy saving by applying the ANN predictive model to the target buildings, five cases were selected. The Base_case represents the space heating and service hot water load of the building itself. Case_1 applies the predicted solar thermal energy production to the Base_case, and Case_2 applies the measured solar thermal energy production. Case_3 applies a storage tank to Case_1, and Case_4 applies a storage tank to Case_2. It turned out that Case_1 and Case_2 showed about a 14 % energy saving rate compared to Base_case. Case_3 and Case_4 showed about a 43 % saving rate compared to Base_case, and about 34 % compared to Case_1 and Case_2.