Magneto‐Thermal Convection of Tri‐Hybrid Nanofluids in a Rectangular Cavity Using the Finite Element Method and Artificial Neural Networks

ABSTRACT

This study proposes a thermal predictive analysis of a rectangular cavity containing ternary nanofluids. The cavity is embedded with mixed convection, magnetohydrodynamics, and a heat source. The mathematical model is developed with the Navier–Stokes and energy equations and then simulated using the finite element method. The employment of the SIMPLE technique and grid sensitivity tests are employed for different mesh sizes to obtain a refined mesh. We examine the effects of Hartmann number , Reynolds number , volume fraction coefficients , and Richardson number . The artificial intelligence model of a neural network is used to make predictions on the Nusselt number in a rectangular heat source. Regression coefficients and the mean squared error are used to determine performance. This result will be close to predicting ANN predictions to the experimental data of the average Nusselt number, which proves that the model is properly trained and can strongly predict heat being transferred in this system. It was noted that a rise in the Richardson number leads to a major increase in the horizontal and vertical velocity components, and hence the implication is increased convective motion. In addition to this, increases in Richardson are accompanied by a significant decline in the Nusselt number adjacent to the heated wall. Ultimately, the hybrid ANN–LMS approach's resilience and dependability were confirmed by the ANN forecasts' magnificent accuracy and tight alignment with numerical simulations.