Physics-informed neural networks for inverse problems in supersonic flows
Ameya D. Jagtap,
Zhiping Mao,
Nikolaus Adams
et al.
Abstract:Accurate solutions to inverse supersonic compressible flow problems are often required for designing specialized aerospace vehicles. In particular, we consider the problem where we have data available for density gradients from Schlieren photography as well as data at the inflow and part of wall boundaries. These inverse problems are notoriously difficult and traditional methods may not be adequate to solve such ill-posed inverse problems. To this end, we employ the physics-informed neural networks (PINNs) and… Show more
“…Jagtap et al [15] propose cPINN that splits the computing domain into several small subdomains with different NNs to solve Burgers equation and Euler equations. Jagtap et al [16] especially study inverse problems in supersonic flows. The above literature shows that PINN is really effective in handle inverse problems with prior information about the development of the flow structures, such as the density gradient.…”
In this paper, we proposed a method to improve the discontinuities (especially shock waves) capturing ability of physics-informed-neural-networks (PINN) in simulating hyperbolic equations. The main idea of the method is to weaken the influence of the points inside discontinuities which can not be expressed directly by the differential equations theoretically, and may let the trainings fall into a confrontation with the physics compressible effect. In this work, we add a weight to each point which is related to the gradient locally, then the network can focus on training the ‘differential equations expressed points’(smooth points). Automatically by the physical compressible effect, all the nearby points will move out of the discontinuously regions, and gain a sharp and exact result automatically with the physical process inside the training
“…Jagtap et al [15] propose cPINN that splits the computing domain into several small subdomains with different NNs to solve Burgers equation and Euler equations. Jagtap et al [16] especially study inverse problems in supersonic flows. The above literature shows that PINN is really effective in handle inverse problems with prior information about the development of the flow structures, such as the density gradient.…”
In this paper, we proposed a method to improve the discontinuities (especially shock waves) capturing ability of physics-informed-neural-networks (PINN) in simulating hyperbolic equations. The main idea of the method is to weaken the influence of the points inside discontinuities which can not be expressed directly by the differential equations theoretically, and may let the trainings fall into a confrontation with the physics compressible effect. In this work, we add a weight to each point which is related to the gradient locally, then the network can focus on training the ‘differential equations expressed points’(smooth points). Automatically by the physical compressible effect, all the nearby points will move out of the discontinuously regions, and gain a sharp and exact result automatically with the physical process inside the training
“…The idea is to combine traditional scientific computational modeling with a data-driven ML framework to embed scientific knowledge into neural networks (NNs) to improve the performance of learning algorithms (Lagaris et al, 1998;Raissi and Karniadakis, 2018;Karniadakis et al, 2021). The Physics Informed Neural Networks (PINNs) (Lagaris et al, 1998;Raissi et al, 2019Raissi et al, , 2020 were developed for the solution and discovery of nonlinear PDEs leveraging the capabilities of deep neural networks (DNNs) as universal function approximators achieving considerable success in solving forward and inverse problems in different physical problems such as fluid flows (Sun et al, 2020;Jin et al, 2021), multi-scale flows (Lou et al, 2021), heat transfer (Cai et al, 2021;Zhu et al, 2021), poroelasticity (Haghighat et al, 2022), material identification (Shukla et al, 2021), geophysics (bin Waheed et al, 2021, 2022, supersonic flows (Jagtap et al, 2022), and various other applications (Waheed et al, 2020;Bekar et al, 2022). Contrary to traditional DL approaches, PINNs force the underlying PDEs and the boundary conditions in the solution domain ensuring the correct representation of governing physics of the problem.…”
The paper presents an efficient and robust data-driven deep learning (DL) computational framework developed for linear continuum elasticity problems. The methodology is based on the fundamentals of the Physics Informed Neural Networks (PINNs). For an accurate representation of the field variables, a multi-objective loss function is proposed. It consists of terms corresponding to the residual of the governing partial differential equations (PDE), constitutive relations derived from the governing physics, various boundary conditions, and data-driven physical knowledge fitting terms across randomly selected collocation points in the problem domain. To this end, multiple densely connected independent artificial neural networks (ANNs), each approximating a field variable, are trained to obtain accurate solutions. Several benchmark problems including the Airy solution to elasticity and the Kirchhoff-Love plate problem are solved. Performance in terms of accuracy and robustness illustrates the superiority of the
“…Since then, there has been an explosive growth in designing and applying PINNs for a variety of applications involving PDEs. A very incomplete list of references includes [36,28,33,45,12,13,14,16,29,30,31,2,40,15,11,41] and references therein.…”
We prove rigorous bounds on the errors resulting from the approximation of the incompressible Navier-Stokes equations with (extended) physics informed neural networks. We show that the underlying PDE residual can be made arbitrarily small for tanh neural networks with two hidden layers. Moreover, the total error can be estimated in terms of the training error, network size and number of quadrature points. The theory is illustrated with numerical experiments.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.