High-resolution monitoring and simulation of the temperatures in a borehole heat exchanger field over multiple years

Abstract

Installation and operation of ground source heat pump systems build upon an enormous practical experience. As these systems are operated for decades, their techno-economic tenability strongly depends on the long-term in-situ heat transfer efficiency and thus the altered ambient subsurface temperatures. Imbalanced seasonal heat exchange causes thermal changes in the ground, which need to be understood to reliably predict the performance of a system during its lifetime. Despite extensive research, understanding the interaction between subsurface and borehole heat exchangers (BHEs) remains inconclusive and not thorough enough. This is due to the scarcity of detailed monitored sites and the monitoring time needed, which renders model calibration with data over multiple years or decades barely feasible. This study monitors an operational field over six years to better understand subsurface thermal changes induced by BHE operation. The ground source heat pump system supplies heat and cold for a multi-family house in Lausen, Switzerland, and consists of an array of three double-U tubes of 146 m each, along with one observation borehole. Ground temperatures were measured by distributed temperature sensing (DTS) at a temporal resolution of 20 minutes in each BHE as well as in the observation borehole. The thermal conductivity of different subsurface layers was derived through an enhanced thermal response test before the start of operation. The monitored data at the location of the BHEs reveal an annual cooling trend of around 0.55 °C. For the understanding of the controlling factors and processes, these cooling trends and temperature variabilities are reproduced in a numerical model. This is also done to evaluate the predictive capabilities of the model and to ultimately judge the spatial and temporal resolution needed for reliable simulation. Thereby, a three-dimensional U-tube model with heat extraction and free cooling was implemented for the site. Comparison of the simulation against measured data of the three BHEs reveals a fair agreement with a root mean squared error (RMSE) of 0.82 °C, whereas the full variability of observation borehole data cannot be reproduced by the model.