Using different machine learning algorithms to predict the rheological behavior of oil SAE40-based nano-lubricant in the presence of MWCNT and MgO nanoparticles

In the present study, using 15 machine learning algorithms (MLP, SVM, RBF, ELM, ANFIS, D-Tree, MLR, MPR, BPNN, BN, LM, GD, BFGS, XGB and GMDH), the rheological behavior of oil SAE40 based nano-lubricant in the presence of MWCNT and MgO nanoparticles was predicted. According to the review of several criteria and data analysis charts, it can be concluded that the best algorithm for predicting fluid properties in this article, which are & mu;nf and torque, is equal to GMDH and MPR, respectively. Also, it can be seen that the data predicted by the machine learning algorithms were able to predict the experimental data very accurately. According to this correlation and the high accuracy of the algorithms, data analysis can be performed on the equations. After determining the range of input variables and specifying the objectives, optimization can be done by the NSGA-II algorithm. Considering that the problem is multi-objective, it is not possible to find a point where both functions are at their minimum value. For this purpose, optimization provides a set of points to the user to choose the optimal point among them based on the need.