University of Twente Student Theses


Machine Learning on Geothermal Heat Extraction

Sitvast, Cas (2021) Machine Learning on Geothermal Heat Extraction.

[img] PDF
Abstract:Drilling for oil and water has become a common thing, however still resources can be collected from old wells via the help of radial drilling. The way this drilling is done these days, does not always lead to a success. Therefore this bachelor's assignment will investigate collected data from drillings in order to illuminate the process and create more insight into factors that determine the progress. This is done with the help of an unsupervised machine learning algorithm called mean shift.\\ In this paper, a tutorial will be described based on an online data set. An elaboration on how the mean shift algorithm works will also be given, and it will be applied to so-called jetting events that characterise successful steps in radial drilling data. \\ The drilling data as it was obtained, first needed to undergo a sequence of actions in order to obtain what we call patterns or features. These patterns were scaled to render a more uniform representation and subjected to the mean shift algorithm. This resulted in a separation of patterns, which was considered 90\% accurate by an expert. An investigation was also done applying the algorithm to less uniform data. This was considered to have an accuracy of 94\%. In both cases the algorithm was trained on all available patterns. The same investigations were also done with the algorithm trained on half of the patterns. This resulted in a mean accuracy of 69\% on the completely uniform data and 62\% on the relaxed uniform data. The data used in these results were obtained by one sensor, more research can be conducted in combining these results with signals obtained from other sensors. Research could also be done on whether the combination of multiple sensors could result in inaccurate predictions of these patterns.
Item Type:Essay (Bachelor)
Faculty:EEMCS: Electrical Engineering, Mathematics and Computer Science
Programme:Applied Mathematics BSc (56965)
Link to this item:
Export this item as:BibTeX
HTML Citation
Reference Manager


Repository Staff Only: item control page