Impact of cyclic thermal loading on thermal convection along a high-temperature borehole heat exchanger: Laboratory-scale experimental and numerical study

A series of combined laboratory scale experiments and numerical simulations of a high-temperature borehole heat exchanger (BHE) in a geological porous medium was carried out under dynamic cyclic thermal loading. The main objective of this work was to investigate the influence of the transient charging/discharging cycle time scales on thermal convection and heat transfer characteristics. The experimental unit was constructed with a vertical, grouted coaxial BHE installed in the center of a water saturated coarse sand within a 1.4 m3 cylindrical barrel. Five cyclic experiments at 70°C charging temperature and cycle durations of 6, 12 and 24 hours after a pre-charging phase, and 6 and 12 hours without pre-charging were conducted under well controlled conditions. A dense sensor grid of 129 thermocouples was deployed to record the evolution of the temperature distribution inside the unit and across the boundary. A 2D-axisymmetric numerical model of the experimental unit was developed in OpenGeoSys and validated for the coupled thermo-hydraulic processes in the porous medium by comparison to previously performed short-term static charging/discharging experiments. The model then was applied to simulate the experiments presented here. The model-to-data comparison indicated a very good agreement, and thus was further used to analyse the impact of the cyclic operation on the contribution of thermal convection to overall BHE heat transfer. The results indicated a cumulative increase of sand temperature levels during the first cycles in the experiments without pre-charging and a decrease when cyclic operation was started after a stationary pre-charging stage, until a quasi stable behavior was reached between consecutive cycles. In general, the difference in obtained results between both operating modes for a given cycle duration was only observed within the first three to four cycles. An increase of the magnitude of convective flow characterised by increases of heat transfer rates, Rayleigh numbers, buoyant flow velocities, and temperature gradients with the cycle duration from 6, 12 to 24 hours was found. The estimated contribution of thermal convection to heat transfer at the last cycle decreased from 23% at 24 hours cycle duration to 20% and 18% at 12 and 6 hours with pre-charging, and 21% and 16 % without pre-charging, respectively. These results demonstrate that for the given experimental setup and scale, convective flow can no longer fully develop within the shorter dynamic charging cycle durations, and thus is less effective for increasing the BHE-to-sand heat transfer during heat charging.