HAL is a multi-disciplinary open access archive for the deposit and dissemination of scientific research documents, whether they are published or not. The documents may come from teaching and research institutions in France or abroad, or from public or private research centers. L'archive ouverte pluridisciplinaire HAL, est destinée au dépôt et à la diffusion de documents scientifiques de niveau recherche, publiés ou non, émanant des établissements d'enseignement et de recherche français ou étrangers, des laboratoires publics ou privés.

No abstract

Abstract. Performance and accuracy of machine learning techniques to segment rock grains, matrix and pore voxels from a 3-D volume of X-ray tomographic (XCT) grayscale rock images was evaluated. The segmentation and classification capability of unsupervised (k-means, fuzzy c-means, self-organized maps), supervised (artificial neural networks, least-squares support vector machines) and ensemble classifiers (bragging and boosting) were tested using XCT images of andesite volcanic rock, Berea sandstone, Rotliegend sandstone and a synthetic sample. The averaged porosity obtained for andesite (15.8 ± 2.5 %), Berea sandstone (16.3 ± 2.6 %), Rotliegend sandstone (13.4 ± 7.4 %) and the synthetic sample (48.3 ± 13.3 %) is in very good agreement with the respective laboratory measurement data and varies by a factor of 0.2. The k-means algorithm is the fastest of all machine learning algorithms, whereas a least-squares support vector machine is the most computationally expensive. Metrics entropy, purity, mean square root error, receiver operational characteristic curve and 10 K-fold cross-validation were used to determine the accuracy of unsupervised, supervised and ensemble classifier techniques. In general, the accuracy was found to be largely affected by the feature vector selection scheme. As it is always a trade-off between performance and accuracy, it is difficult to isolate one particular machine learning algorithm which is best suited for the complex phase segmentation problem. Therefore, our investigation provides parameters that can help in selecting the appropriate machine learning techniques for phase segmentation.

In this paper a testing device is described that measures the hydraulic conductivity of grout specimens for Borehole Heat Exchangers (BHE). During the operation of closed-loop ground source heat pumps running with antifreeze, freezing of the backfill can occur due to extensive heat extraction. This laboratory device can assess the influence of frost on the hydraulic seals of BHEs. The device is based on a triaxial flexible wall permeameter. A freely selectable number of cyclic freeze-thaw-stresses as well as a confining pressure simulating radial earth pressure (r 2 ¼ r 3) can be applied. Specimens are composed of an annular grout body and a polyethylene pipe simulating the BHE system. The freezing direction is perpendicular to the vertical axis of the BHE from the inside to the outside. Numerical coupled modeling was applied to verify the results of the temperature distribution inside the specimens. It was observed and modeled that the propagation of the frost front and the fabric disintegration processes are correlated. Results of three different grouting materials will be presented. With its relative small dimensions the device can be easily implemented into soil mechanical laboratories and thus can contribute to quality control of grouts.

The amount of research conducted on geothermal energy as a source for heating and cooling demands of buildings, as well as for electrical energy production, has increased substantially in the past decades. The simulation of freezing and thawing is important for geothermal applications involving ground coupled heat pumps. One area of research is the development of grout cements for borehole heat exchangers (BHE). In many cases, BHEs are operated at temperatures below 0° C due to manifold reasons. Hence, the simulation of freezing and thawing cycles (FTC) is important for such geothermal applications, especially in cold regions. Recently, a testing device for measuring and quantifying the influence of FTC stresses on the mechanical integrity and hydraulic properties of BHE grouts was developed (Anbergen, published in 2014). The testing procedure simulates the downhole in situ conditions as confining radial earth pressure, freezing, and thawing directions from the inside to the outside and under saturated conditions. The hydraulic conductivity can be measured in axial flow direction. Thus, statements regarding the susceptibility of grouts against cyclic freezing and thawing stresses can be made. These results differ substantially from earlier findings, as in situ boundary conditions were not simulated sufficiently. For the verification of the procedure’s thermal process, temperature logs were recorded using thermocouples and thermography imaging. The thermal process was simulated using the finite element method (FEM) groundwater, heat, and mass modeling software FEFLOW. FEFLOW is a common software solution for thermohydraulic coupled groundwater applications with mass transport, as well as geothermal applications. However, up until now, the program could not yet simulate phase changes between solid and liquid phases. To enable the program for such simulations, a plug-in was developed. To do this, a C++ code was written and coupled to the simulation routine of the FEM software. The code is based on a modification of the material parameters of fluid and the incorporation of the latent heat effects in the fluid heat capacity. A linear and an exponential approach for the latent heat release were implemented and benchmarked. The code was verified using different analytical solutions and other FEM codes. Finally, the experimental results of the test procedure could be successfully computed using the new plug-in. Thus, it is now possible to compute phase changes with FEFLOW for geothermal applications as well as other applications like permafrost research.