Application of LS-SVM Classifier to Determine Stability State of Asphaltene in Oilfields by Utilizing SARA Fractions

Abstract

Asphaltene deposition in petroleum industry is a challenge that causes many problems during production and transportation. Some researchers provided useful works but no one can explicitly determine the stability state because of complex nature of aasphaltene and the numerous forces that affect its stability. Determining stability state of crude oil, especially in oilfields, help to have a thorough look on reservoir management. The author proposes a least square support vector machine (LS-SVM) to classify and determine asphaltene stability state of 200 oilfields. The input variables of model consist of saturate, aromatic, resin, and asphaltene fractions. Coupled simulated annealing optimization was adapted to find the optimal values of LS-SVM. Since the prediction largely relies on the selection of training data, 173 training data sets were randomly selected for model prediction and 27 data was used for testing the model. The input divided into two classes: saturate + asphaltene and resin + aromatic and consequently output divided into three modes: severe, mild, and no/minor problems. The result elucidates that the proposed model can be implemented for asphaltene stability determination and is extremely reliable.