On the Parametric Study of Lubricating Oil Production using an Artificial Neural Network (ANN) Approach

TL;DR

This paper employs an Artificial Neural Network to analyze how various operational parameters affect lubricant extraction efficiency from heavy petroleum, providing a data-driven modeling approach.

Contribution

It introduces a neural network model trained on industrial data to predict lubricant flow rate based on operational conditions, advancing process understanding.

Findings

ANN accurately predicts lubricant flow rate

Operational parameters significantly influence extraction efficiency

Model demonstrates potential for process optimization

Abstract

In this study, an Artificial Neural Network (ANN) approach is utilized to perform a parametric study on the process of extraction of lubricants from heavy petroleum cuts. To train the model, we used field data collected from an industrial plant. Operational conditions of feed and solvent flow rate, Temperature of streams and mixing rate were considered as the input to the model, whereas the flow rate of the main product was considered as the output of the ANN model. A feed-forward Multi-Layer Perceptron Neural Network was successfully applied to capture the relationship between inputs and output parameters.