Optimization and effect of wall conduction on natural convection in a cavity with constant temperature heat source: Using lattice Boltzmann method and neural network algorithm

In this paper, the natural convection (NC) of the Al 2 O 3 –H 2 O nanofluids (NFs) in a cavity with a heat source in its center is numerically investigated. A two-dimensional simulation of the cavity has been performed for this purpose. The D2Q9 LBM was used for the simulation. Then, the largest Nu was obtained using the artificial neural network (ANN). The ANN has been trained in such a way as to be able to automatically modify the input values and to estimate the largest heat transfer. With an increase in the Ra and nanoparticle volume fraction (NPVF), the heat transfer increases, but it decreases with an increase in the Ha number. In addition, the maximum heat transfer rate occurs for parallel magnetic field (MaF) and the heat transfer rate decreases with an increase in the MaF angle. As the size of the isothermal heat transfer and cold walls increase with a rise in the thermal conductivity coefficient, heat transfer also increases. With an increase in the thermal conductivity coefficient, no significant change was observed in the rate of heat transfer after an initial increase in the Nu. The largest Nu occurs at the largest Ra, largest MaF angle, and weakest MaF. The ANN predicted this value to be 7.0703, which is 2.3% different from the exact value resulting from the predicted computations.