Investigation of Wax Thickness Rate in Pipelines Using ANFIS

One of the ongoing challenges in the oil industry is wax deposits on the oil and gas pipelines. The majority of crude oils produced include a considerable amount of wax. When the temperature is the below of specified value, wax deposits occurs from crude oil and condensates. Therefore, by reducing the oil temperature increases the wax deposition. Wax deposits mainly create problems in pipelines and equipment of oil wells. Also, wax deposits in pipes increases the pressure drop and reduce production and reduces production efficiency. According to the reasons mentioned, forecasting wax thickness requires accurate and secure way. In this study, by using the fuzzy and neural-fuzzy systems a model was evaluated for predicting the wax thickness in single-phase and turbulent flows. Comparing the experimental results with the predicted values by the fuzzy and fuzzy neural systems indicated that the predicted values by these models have a good agreement with the experimental data and it would show the reliability of these models. The average of absolute deviation error in fuzzy and neuro-fuzzy systems is 6.1577 and 6.2148 percent respectively. Results of this forecast show favorable performance of modeling conducted for predicting the rate of wax thickness in single-phase and turbulent flows

Keywords:

fuzzy systems , neural-fuzzy networks , wax deposition , single-phase turbulent flows