Data-driven approaches for predicting wax deposition. (15th February 2023)

• Record Type: Journal Article
• Authors: Ahmadi, Mohammadali
• Is Part Of: Energy. Volume 265(2023)
• Publication Date: 2023-02-15
• Subjects: Wax deposition -- Modeling -- Crude oil -- Fuzzy logic -- Genetic algorithm -- Artificial neural network; GA Genetic Algorithm -- FL Fuzzy Logic -- WDT Wax Disappearance Temperature -- WAT Wax Appearance Temperature -- ANN Artificial Neural Network -- SVM Support Vector Machine -- MF Membership Function -- MSE Mean Square Error -- FIS fuzzy Inference System; Power resources -- Periodicals; Power (Mechanics) -- Periodicals; Energy consumption -- Periodicals; 333.7905
• DOI: 10.1016/j.energy.2022.126296
• Languages: English
• ISSNs: 0360-5442

Abstract

Abstract: Deposition of Wax is a challenging issue in the upstream oil and gas production process. Thus, precise estimating of this deposition/precipitation issue can increase the performance of petroleum products. To solve this issue, enormous research has been done previously, but unfortunately, the outcomes of the addressed studies are not a proper solution for the mentioned problem. In this research, a method of coupling the fuzzy logic and genetic algorithm (GA) has been developed to propose a robust and efficient method for predicting the amount of deposition of Wax. The input variables in the developed model are oil composition, temperature, pressure, and oil-specific gravity, which evolved in conventional approaches to estimate wax deposition amount. Different statistical indexes like mean square error and relative deviation were implemented to determine the results' accuracy and integrity. Based on the addressed statistical criteria, the gained results of the evolved model have high performance and integrity in contrast with other methods or correlations in the determination of the wax deposition values. It is worth noting that the suggested method's advantages are wide ranges of applicability and simple utilization. Outcomes gained from this communication confirmed the applicability of fuzzy logic in wax deposition measurement, which can better understand wax deposition in pipelines, wellbores, and reservoirs. Highlights: Data-driven approaches developed for predicting WAX deposition percent. Handling extensive WAX deposition data in oil reservoirs using fuzzy models. Aconnectionist approach to predicting WAX deposition using a hybrid of ANN, fuzzy logic, and GA algorithms developed.