Akademik Veri Yönetim Sistemi
24th Congress on Thermal Science and Technology, Ankara, Türkiye, 6 - 08 Eylül 2023, ss.1-7
Physics-informed neural networks (PINNs) have drawn attention in recent years in engineering problems due to their effectiveness and ability to tackle problems without generating complex meshes. PINNs use automatic differentiation to evaluate differential operators in conservation laws and hence do not need a discretization scheme. Using this ability, PINNs satisfy governing laws of physics in the loss function without any training data. This framework uses the ability of PINN to have solutions to thermal convection problems. Remin is a framework of physics-informed neural networks using PyTorch. It has the ability to construct different types of geometric entities using Latin hypercube sampling. PyTorch enables the construction of neural networks with different types of activation functions, optimizers, and schedulers to change the learning rate during training. We implemented our framework to solve steady thermal convection problems. The results are compared with a high-order discontinuous Galerkin solver. Overall, this framework can solve thermal convection problems in different regimes using neural networks.