Abstract: In this paper, a physics-informed neural network (PINN) is proposed to solve 2D magneto-quasi-static (MQS) problems in time domain. Instead of discretization or interpolation in space and ...