Vertical ground heat exchanger parameter characterization through a compound design of experiments

Abstract

Vertical ground heat exchangers (VGHE) are the most extended elements in geothermal utilization systems. For design and optimization purposes, the precise determination of several thermal properties is of utmost importance. While soil properties can be accurately obtained from simple mathematical models and short identification procedures, the grout characterization and additional uncertainties are usually identified from time-consuming and complex models. In order to optimize and simplify this process, without loss of accuracy, the present work proposes the use of design of experiments (DoEs) for that estimation. Thus, a multi-physic 3D model that replicates the geothermal heat exchanger (GHE) environment has been employed to develop a fractional factorial design of experiments, obtaining a set of output errors by comparison with experimental data. A response surface methodology (RSM) has been used to conduct second-order linear regression fittings, that approximate the error functions to obtain the parameter values through the minima of such surfaces. Overall mean errors lower than 0.01 °C are obtained, which provide excellent agreement in modelling the experimental test. Finally, an additional full-scale experiment has been simulated in order to validate the numerical DoE approach. Consequently, the proposed methodology has been demonstrated to accurately predict the grout thermal parameters.