Machine learning for creation of generalized lumped parameter tank models of low temperature geothermal reservoir systems

• Published in: Geothermics Vol. 70; pp. 62 - 84
• Main Authors: Li, Y., Júlíusson, E., Pálsson, H., Stefánsson, H., Valfells, Á.
• Format: Journal Article
• Language: English
• Published: Oxford Elsevier Ltd 01.11.2017; Elsevier Science Ltd
• Subjects: Algorithms; Artificial intelligence; Cluster analysis; Clustering; Complexity analysis; Complexity reduction algorithm; Geothermal power; Geothermal reservoir modeling; Information criterion; K-means clustering algorithm; Learning algorithms; Low temperature; Lumped-parameter models; Machine learning; Mathematical models; Model accuracy; Model selection; Parameters; Physical properties; Production planning; Recursive methods; Reservoirs; Switch-back method; Tanks; Thermal energy; Vector quantization; Complexity reduction algorithm; Lumped-parameter models; Complexity analysis; Geothermal reservoir modeling; Information criterion; Model selection; Machine learning; K-means clustering algorithm; Switch-back method
• ISSN: 0375-6505; 1879-3576
• DOI: 10.1016/j.geothermics.2017.05.009

•Automating generation of tank models. The subject material should be of interest to reservoir engineers and operators of geothermal systems who are concerned with use of efficient modelling methods, optimal production planning and introduction of machine learning techniques to the geothermal sector.•The automated procedure described in this paper utilizes a complexity reduction algorithm to reduce a fully linked model to a simpler model via a process of merging of tanks and removal of connections between them.•We introduce a switch-back method whereby the parameter values describing the “best” model obtained from the complexity reduction algorithm, is fed back into starting estimates for parameters of a fully linked, T tank model Lumped parameter tank models have gained renewed interest in recent years as an alternative tool for geothermal reservoir analysis and production planning. The models can be structured in various ways regarding the number of tanks, connections between the tanks and the parameters representing the physical properties of the geothermal system. It usually requires a time consuming and difficult process of trials and errors to manually decide the optimal configuration of a tank model. Inspired by recent development in the use of machine learning methods, we propose a method for automatically generating accurate and computationally feasible generalized tank models for isothermal, single phase, reservoirs. This is an extension of earlier work on complexity reduction of generalized tank models (Li et al., 2016). Here, a recursive “switch-back” method is constructed to maximize prediction accuracy of the model. It is also shown how the K-means clustering algorithm can be used to aggregate production wells in generalized tank models. One synthetic example and one field application from t Reykir geothermal fields in Iceland are used to illustrate the effectiveness of these methods.