Application of artificial intelligence and using optimal ANN to predict the dynamic viscosity of Hybrid nano-lubricant containing Zinc Oxide in Commercial oil

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

“…Several research teams have modeled the dynamic viscosities or densities of ionic liquids using data, such as temperature, concentrations, and even a structure factor (Billard et al, Valderrama et al, and Kang et al). Dynamic viscosity of hybrid nano lubricants was also correlated by ANN (Afrand et al, Esfe et al). However, to the best of our knowledge, no publication has used the groups’ contributions in an ANN to determine the viscosity of aqueous electrolyte solutions and discuss viscosity sensitivity with different functionalized groups.…”

“…To correctly model a nonlinear property, an activation function must be used to take this nonlinearity into account. The main function used in the literature ,,,,, is the hyperbolic tangent noted as tanh (eq ). Hyperbolic tangent has the necessary property to be continuous over the entire range of relative numbers and it can be derived throughout the interval tanh ( n ) = exp ( n ) − exp ( − n ) exp ( n ) + exp ( − n ) = 2 1 + exp ( − 2 · n ) − 1 …”

“…Therefore, the normalized output is calculated as follows (eq 8) and the definition of each term is included in Table 4 Even if the data is randomly distributed, it needs to be evenly split between training data and validation data (Figure 3). This choice of proportion for data training (70%) is often made by researchers (Longo et al, 38 Esfe et al, 44 and Soroush et al 47 ). However, to determine the best number of neurons, this approach is not fine enough, and the use of cross-validation often appears often relevant.…”

“…Even if the data is randomly distributed, it needs to be evenly split between training data and validation data (Figure ). This choice of proportion for data training (70%) is often made by researchers (Longo et al, Esfe et al, and Soroush et al …”

Accurate prediction of dynamic viscosity of polyalpha-olefin boron nitride nanofluids using machine learning

“…VF 0.9971 The BRT showed slightly better results than ANN and SVM. [ 33 ] MWCNT-MgO-OB ANN, Correlation Temp., VF, SR 0.9990 ANN showed better performance compared to correlation (R 2 correlation = 0.9321) [ 45 ] MWCNT-ZnO hybrid-OB ANN Temp., VF, SR 0.9900 Single configuration was studied [ 46 ] WO 3 -MWCNT-EGB ANN Temp., VF 0.9980 Different training algorithms are evaluated alongside “trainlm” [ 34 ] MWCNT-ZnO-OB Correlation, ANN Temp., VF, SR 0.9921 A single hidden layer with six neurons was studied only. The ANN showed better performance than the correlation [ 35 ] TiO 2 -Al 2 O 3 -EGB Correlation, ANN Temp., VF, Density 0.9982 Slight better performance for the ANN over correlation.…”