Effects of using perforated fins on the melting in different type heated trapezoidal cavity equipped with encapsulated PCM during forced convection and modeling with FNN/GNN

In the present study, melting process in a trapezoidal cavity equipped with encapsulated phase change materials (PCMs) under the combined impacts of using perforated fins and different heating types is numerically explored during forced convection. Effects of Re (between 300 and 1000), fin height ($L_f$ between $0.01H$ and $0.14H$), fin number ($N$ between 0 and 4), number of holes in the perforated case ($Nh$ between 1 and 8), and size of the trapezoidal cavity upper wall ($Hp$ between $0.01H$ and $0.75H$) on the flow features and melting dynamics are explored by using finite element method. It is observed that installation of the fins is very effective on the melting process at the highest size and highest number. As compared to non-perforated case, perforated fins produces almost the same performance in the melting process but volume per fin can be reduced up to 42.85%. At Re = 1000, finned configuration improves melting by 21.6%. As compared to one fin case, melting performance improves by 37.8% with four fins. The improvement in the melt fraction is 18.9% when the fin length is increased from $L_f=0.1H$ to $L_f=0.14H$. The melting performance improvement is 8.2% higher when all sides of the inner part are heated as compared to one side heating case for $N=1$. Perforated fins in the trapezoidal cavity with nanofluid improve the melting process by 40.6% when compared to a square cavity without fins and using base fluid. Feed forward networks (FFN) and generalized neural networks (GNN) are used to successfully estimate the dynamics of the melting process.