Application of artificial neural network for prediction of 10 crude oil properties

Abstract

This study aims to develop an industrially reliable and accurate method to estimate crude oil properties from their Fourier transform infrared spectroscopy (FTIR) spectra. We used the complete FTIR spectral data of selected crude oil samples from seven different Canadian oil fields to predict 10 important crude oil properties using artificial neural networks (ANNs). The predicted properties include specific gravity, kinematic viscosity, total acid number, micro carbon content, and production of light and heavy naphtha, Kero, and distillate in oil refineries. The 107 different (65 light oil and 42 heavy/medium oil samples) crude oil samples used in this study came from seven oil fields and reservoirs across Canada. In line with standard practice, we used 80% of the dataset for training the ANN models and used the remaining 20% of the crude oil samples to test the models. In the ANN analysis, the mean squared error (MSE) was used as the loss function in models, and the mean absolute prediction error (MAPE) was used as a reference to compare the performance of different neural networks constructed with different numbers of layers. This work demonstrates that FTIR spectroscopy is a promising technique that provides rapid and accurate estimates for the oil properties of interest to the industry. A comparison of the values predicted by the validated ANN models and their corresponding measured (actual) values showed excellent prediction with the acceptable range of error (below 15%) aimed for by our industry partner for all properties except viscosity, for which building models based on the natural logarithmic values of measured viscosities significantly improved the results.

Introduction to the K. Nandakumar Festschrift Issue section ofCJCE

“…They perform CFD simulations and develop machine learning models for predicting the hydrodynamic and mass transfer characteristics of the rotating packed bed system. Alizadeh et al [ 7 ] share their work that uses artificial neural networks to adequately predict 10 important properties of crude oil collected from seven different Canadian oil fields using only their Fourier transform infrared spectroscopy (FTIR) spectra. In their contribution, Basu and Mahajan [ 8 ] employ 3D direct numerical simulation to evaluate four specific flow field designs and make recommendations for a reverse electrodialysis cell that can be used to produce ‘blue energy’.…”

Prediction of Dynamic Viscosity and Sensitivity Study of Potassium Amino-Acid Salt Aqueous Solutions by an Artificial Neural Network According to the Structure

“…However,•to our best knowledge, amino acid anions have not been studied in the framework of the theory of Glasstone et al 27 Another approach is to take advantage of machine learning. In the last two decades, several studies have sought to use ANN to predict properties such as solubility, density, heat capacity, or viscosity of mixtures of molecules (Hamzehie and Najibi, 34 Pouryousefi et al, 35 Alabi and Williamson, 36 and Alizadeh et al 37 ). Longo et al 38 built an ANN to estimate the dynamic viscosity of H 2 O/KCOOH solutions with mass fraction and temperature as factors.…”