Stability of penetrative convection in a Darcy–Brinkman porous medium: an artificial neural network approach
Аннотация
The onset of penetrative convection with the impact of anisotropic thermal diffusivity in a porous medium saturated with fluid is analyzed via internal heating and a quadratic density–temperature law. Linear and nonlinear stability analyses are carried out, with the flow in porous media described by the Darcy–Brinkman model. The eigenvalue problems obtained from both linear and nonlinear stability analysis are solved using the Chebyshev–Tau–QZ method, and the effects of various dimensionless parameters, such as the Rayleigh number Ra , internal heat source/sink parameter Q , Darcy number Da , upper-layer temperature $$T_\textrm{U}$$, and anisotropic thermal diffusivity parameter $$\xi $$, are investigated. It is found that the system is destabilized by increasing the internal heat sink parameter, whereas an increase in the Darcy number delays the onset of penetrative convection, and the potential region of subcritical instability expands as the Darcy number increases. The onset of penetrative convection is also delayed as the upper-layer temperature and the anisotropic thermal diffusivity parameter increase. Additionally, an artificial neural network (ANN) model is trained to estimate the critical thermal Rayleigh number, demonstrating excellent agreement between the numerical and ANN predictions. The suitable number of neurons in the hidden layer is determined by the coefficient of determination $$R_{*}$$ and mean squared error.