Comparative study of the thermal performance of an earth air heat exchanger and seasonal storage systems: Experimental validation of Artificial Neural Networks model

Abstract

This article presents a comparative study between two different geothermal systems: an earth air heat exchanger (EAHE) and seasonal storage systems. The energy efficiency of a combined geothermal system composed of two series of connected underground tanks coupled with a cooling floor and a PVC earth-air heat exchanger coupled to an experimental cell is presented through an experimental and numerical analysis. A measurement system is installed to study the thermal performance of both systems under the same conditions. The climatic conditions, the temperature at the outlet of the two systems were measured. This experimental protocol allows to evaluate and understand the interest of each system for cooling needs. Experimental results show that the energy performance of two systems is different. The blowing temperature of the EAHE varies between 19 and 24 °C, however the outlet temperature of the underground tank varies between 16 °C and 20 °C. Based on experimental data, a Long-Short-Term-Memory (LSTM) neural network algorithm is then used to forecast the underground tank and EAHE temperatures evolutions. The forecasting accuracy of the proposed model was evaluated using different statistical measures. The determination coefficient (R2), the Root Means Square Error (RMSE), and the Mean Squared Error (MSE) for the EAHE outlet temperature are 0.953, 0.093 and 0.306, respectively. Furthermore, the R2, MSE and RMSE for the underground tank temperature are 0.984, 0.165 and 0.406 respectively.