Machine learning enhancement of thermal response tests for geothermal potential evaluations at site and regional scales

• Machine learning offers a powerful means to account for the testing experience. prediction offer satisfactory substitute on-site testing. Digitalization and pooling of data could decrease need This study explores validates machine approach practical, effective, precise thermo-physical characteristics that are essential analysis design shallow geothermal systems, including borehole heat exchangers: (i) undisturbed ground temperature, (ii) effective thermal conductivity, (iii) resistance. Benefiting from 174 response tests central western Switzerland, algorithm is used provide accurate site-specific as well regional-scale predictions investigated characteristics, which in turn can serve preliminary yet representative evaluations potential even very broad areas.