Scaling remains as the major hurdle in the utilization and development of geothermal energy in addition to corrosion. This is because scaling can occur in almost all locations in geothermal power plants, i.e., in production well, on surface facilities, in reinjection well, and even in the geologic formation around the reinjection well. Despite the vast studies, there are no single and universal mitigation method to date. This is likely due to the unique characteristics of each geothermal water and, consequently, the complexity of scale formation. However, the current understanding suggests that silica scale formation is generally controlled by the interaction between dissolved silicic acid and the metal surface (initiation), and the interaction between silicic acid and the surface of silica scale (growth). This study aims to better understand and quantify silica scaling with the help of artificial intelligence (AI). As training data, the behavior of polymerization of silicic acid in geothermal water and its adsorption on silica gel (model surface material of silica scale) were considered in addition with physicochemical properties of geothermal water. First, geochemical results of various geothermal waters were collected. The data was used to calculate the saturation index of minerals. Both data were used as input parameters. In addition, batch polymerization and adsorption experiments were also conducted onsite each corresponding geothermal power plants at similar temperature condition to quantify the polymerization and adsorption rates of silicic acid. As a preliminary AI calculation, the data were then used as output parameters. Here, we used a supervised type of machine learning, which produce several prediction models. The produced models have percent root mean square error (%RMSE) values ranging from 2.7 to 15.7, suggesting the acceptability and applicability of the models. In addition, factor analysis using the produced model suggest that total concentration of silicic acid contribute the most to the polymerization rate of silicic acid, aligned with the classical understanding. Interestingly, iron concentration of geothermal water gives significantly higher contribution than other species despite its low concentration in geothermal water. This study provides a preliminary result of AI modelling for polymerization and adsorption behavior of silicic acid, as well as proposes its AI modelling procedure. Further, the established AI models need to be examined with more physicochemical data of geothermal water and more experimental data of kinetic behavior of polymerization and adsorption of silicic acid. Finally, we will construct an AI system predicting the deposition rate of silica scales from physicochemical data of geothermal waters (geochemical result, polymerization of silicic acid and adsorption of silicic acid) as input data.

Press the Back button in your browser, or search again.

Copyright 2024, Stanford Geothermal Program: Readers who download papers from this site should honor the copyright of the original authors and may not copy or distribute the work further without the permission of the original publisher.