Monitoring of a geothermal borehole field: Data analysis and calibration of a prediction model

Abstract

Near-surface closed-loop geothermal systems have an important role to play in the decarbonisation of energy. Thanks to their ability to reduce greenhouse gas emissions and their widespread availability throughout the year, it is essential that these technologies be supported by appropriate design studies to ensure their long-term performance and sustainability.

This master thesis aims to analyse the monitoring data from an existing geothermal installation and to calibrate an analytical heat transfer model to reproduce the evolution of underground temperatures. The datasets come from the heat pump system of the "Relais des Ingénieurs" residence in Louvain-la-Neuve. A preliminary filtering and analysis of the raw datasets were required to address interruptions in geothermal system operation and anomalies in data acquisition, allowing for the computation of the thermal load extracted from the ground. In a second step, a considerable portion of the data was excluded due to abnormally high thermal loads, resulting in a reduced analysis period representing approximately 15% of the total monitoring duration. The analytical model developed by Erol et al. (2015), designed to simulate the thermal response of multiple borehole heat exchangers and capable of accounting for discontinuous extractions and groundwater flow, is then calibrated.

Main model parameters, including initial ground temperature, the thermal conductivity of the ground, grout, and pipe, volumetric heat capacity, and specific discharge, are calibrated sequentially to minimise the discrepancy between simulated and monitored fluid temperatures. The model fit is assessed using error metrics such as the root mean square error and mean absolute error. Calibration is conducted for both daily and hourly thermal loads. Results show that the model can accurately reproduce measured temperatures when properly calibrated, highlighting the value of data monitoring in improving geothermal model predictions and system design, especially when daily thermal loads are used.

Despite the limitations of the available data, the study proves robust and could be enhanced through improved instrumentation and longer monitoring campaigns.