How to correct the ambient temperature influence on the thermal response test results

• A TRT evaluation method is proposed to reduce its ambient temperature influence.
• The method reduced the time dependent ground thermal conductivity variation.
• Enhanced insulation reduced the ambient influence on the mean fluid temperature.
• The ambient to mean fluid temperature influence occurred with about a 3 h delay.

Due to global warming and to the increasing energy demand, it is necessary to improve energy efficiency on buildings. In this context, Ground-Coupled Heat Pumps (GCHP) have proved to be the most efficient heating and cooling system. The main parameters to define a ground heat exchanger are obtained via an in situ test called Thermal Response Test (TRT). However, ambient air influence on this test is remarkable due to the exposition of the testing machine, and even the ground undisturbed temperature varies with the ambient temperature oscillations. Therefore, despite the fact that the influence of ambient conditions on the TRT results is an important topic in order to define a ground heat exchanger, there is yet a limited literature on new theoretical methods to correct the ambient temperature influence on the predicted ground thermal conductivity measured via TRT. This paper presents a new methodology to analyse and mitigate the influence of the ambient conditions on the TRT results, with the main advantage that it is not necessary to know its physical origin previously. The method is focused on reducing the mean fluid temperature oscillations caused by the ambient temperature, by analysing the influence of the chosen time interval to fit the data to the infinite line source theory formulae that finally predicts the ground thermal conductivity. With these purpose, results of two different TRTs were analysed, each of them with a different equipment and ambient exposition. Results using the proposed method showed that thermal conductivity oscillations were reduced in both tests. For the first test, the uncertainty associated to the chosen time interval for the estimation was diminished by 33%, reducing significantly its predicted value and thus avoiding the future installation possible under-designing. However, because of the equipment insulation improvements and the smoother ambient temperature variations, the method obtained similar results for the predicted thermal conductivity and for its uncertainty for the second test.