Regular Article

A machine learning model for the computation of thermophysical properties of WCO biodiesel mixed with multiwalled carbon nanotubes

¹ Department of Mechanical Engineering, SECAB Institute of Engineering and Technology, Vijaypur, Karnataka 586109, India
² Department of Mechanical Engineering, Maulana Mukhtar Ahmad Nadvi Technical Campus, Malegaon, Maharashtra 423203, India
³ Department of Mechanical Engineering, Malla Reddy Engineering College, Maisammaguda (H), Medchal Mandal, Telangana 500100, India

^* Corresponding author: syed.abbasali86@gmail.com

Received: 31 August 2024
Accepted: 6 May 2025

Abstract

A Machine Learning (ML) model has been developed to compute the thermophysical properties of Waste Cooking Oil (WCO) biodiesel dispersed with MultiWalled Carbon NanoTubes (MWCNTs). The thermophysical properties when incorporating multiwalled CNTs into biodiesel are critical in improving performance, combustion, and emissions in internal combustion engines because of the high thermal conductivity and mechanical strength. Firstly, MWCNTs are mixed with WCO biodiesel for dosages of 30 ppm, 40 ppm, and 50 ppm. After it, each of the properties, including calorific value, density, viscosity, flash point, and fire point, is evaluated. Further, the MultiLayer Neural Network (MLNN) is a ML model that employs a back-propagation algorithm for mapping the input-output parameters. The parameters that constitute the input are WCO biodiesel blends and MWCNTs ppm. The parameters that are output include the calorific value, density, viscosity, flash point, and fire point. The optimum model’s results indicate that six hidden neurons (2_6H_5) can accurately predict the aforementioned properties under various operating conditions. Then, the Multivariable Regression (MVR) model has been devised to compare with the MLNN model. Subsequently, a comparison between the MLNN and MVR models has been carried out. The properties predicted by MLNN in comparison to the MVR model are seen as close to experimental values with good accuracy, as they depict the good “R” values as 0.98209, 0.97921, 0.99261, 0.9558, and 0.99139 for calorific value, density, viscosity, flash point, and fire point, respectively, and also give the average relative error (RE) for calorific value as 0.803%, density as 0.322%, viscosity as 3.036%, flash point as 5.104%, and fire point as 3.266%. Furthermore, the developed MLNN model is suitable for predicting the calorific value, density, viscosity, flash point, and fire point of WCO biodiesel that has been infused with MWCNTs. This saves time, money, and effort required.