Optimal corrosion prevention in crude oil refineries with surrogate modeling

Yandık, Yücelen Bahadır
Energy demand in the world is increasing day by day, which makes energy markets extremely competitive. Crude oil refineries have to adapt to this competition like other players in the energy field. Corrosion is a common problem in crude oil refineries. Production may need to be stopped for maintenance to fix problems caused by corrosion. These stops cause businesses to miss their production target and lose their competitive advantage. Today, it is known that the salts in crude oil play an important role in corrosion. Even though there are several methods to remove salts in the crude oil, these methods are not perfect. For this reason, chemicals such as neutralizers, corrosion inhibitors, and caustic soda are used to reduce the corrosive effect of salts. However, using these chemicals in the inappropriate amounts can increase the corrosion or create a coke in the tanks that affects the production and needs to be cleaned. Therefore, the amount of chemicals to inject should be determined carefully. If this is performed manually by field operation staff based on heuristic approaches, it may lead to failures. Carrying out the decision-making process with a data-driven analytical method may provide more successful results and enable optimizations. Yet, developing analytical methods is seen as a costly and challenging way due to the complex nature of crude oil refineries. Using surrogate models instead of theoretical models can reduce costs and make the development process more manageable. To this end, we propose a method that optimizes the amount of chemicals added to prevent corrosion in crude oil refineries using an analytical method that relies on surrogate models. In order to evaluate the applicability and performance of the proposed method, an application was carried out in a refinery, and positive results were observed in the short term.
Citation Formats
Y. B. Yandık, “Optimal corrosion prevention in crude oil refineries with surrogate modeling,” M.S. - Master of Science, Middle East Technical University, 2020.