A comparative study of mathematical and ANFIS models to determine the effect of ultrasonic waves on the viscosity of crude oil

Abstract

Increasing fuel consumption, industrial development, and population have led to energy supply problems. Hence, the recovery of oil reservoirs has been considered in recent years. One of the most effective methods in crude oil recovery is the use of ultrasonic waves. Viscosity is a significant physical property of crude oil because the crude oil with high viscosity can make serious problems in transportation pipelines. In addition to the effect of ultrasonic waves on viscosity, ultrasonic radiation leads to molecular bond vibrations in the oil and reduces paraffin deposition and clogged pores. In this study, the effect of ultrasonic waves on the viscosity of crude oil was investigated. Three transmission functions such as Inverse Square Root, Natural Log and Square Root, and ANFIS model were used to predict the viscosity of crude oil. The adequacy of the three mathematical models and the ANFIS model were proved by statistical analysis including low P-values (<0.05), high F-values and R 2 values. According to the results, the ANFIS model and Inverse Square Root transmission function showed better agreement with experimental data compared to the other two models.