| Abstract |
Thermal response test (TRT) is often used to determine the in-situ thermal conductivity of soil for a geothermal heat pump, which uses the constant temperature of the shallow earth to exchange heat efficiently. A conventional TRT takes about 50 hours or longer to see the straight-line relationship between the temperature and logarithm of elapsed time. This practice is easy to use but takes a long time to test, is costly, and results in less information. Furthermore, this practice requires the input thermal power to be constant, which is often a challenge in reality. We propose a new set of TRT data interpretations by analyzing the whole transient thermal transfer regime to obtain the thermal conductivity and diffusivity of soil and thermal resistance of grout of the test well. In the TRT process, the thermal disturbance propagates away from the borehole as the heated working fluid circulates through the borehole. At an early time, the measured temperature behavior is affected by the grout properties used in cementing the borehole. The thermal conductivity isn’t a constant over the borehole lifespan as it ages. In a relatively long testing period, the measured temperature behavior is mainly affected by the soil thermal diffusivity, which can be characterized by the integral exponential function. As the test duration is getting sufficiently long, a logarithm approximation can be applied in lieu of the integral exponential function. Conventional TRT analysis only uses data in the late stage while discarding the previous measurements. Our method utilizes the whole testing period and uses nonlinear regression to match the measurements and determine the model parameters, which contain grout thermal resistance, thermal diffusivity and conductivity of the soil. In the paper, we first discuss the fundamentals of TRT and its analysis. Then we apply the method to show the procedure of interpretation using the new method. The application of the real case shows that our approach is robust in theory, accurate in modeling, and able to determine more parameters by thoroughly using the measurements. |