NeuraFOIL vs ViscousJoukowski

NeuraFOIL brings AI-assisted airfoil analysis to your hand. Meet the first neural network-powered aerodynamic tool designed for engineers, students, and aerospace enthusiasts who want real insights instantly on their iPhone or iPad. Explore NACA 4-digit airfoils by entering the airfoil code, angle of attack, and Reynolds number. NeuraFOIL instantly visualizes the geometry and predicts pressure coefficient, skin-friction coefficient, lift coefficient, and drag coefficient using an on-device Core ML model trained from OpenFOAM CFD simulations. Full data export to CSV for post-processing in Python, MATLAB, or Excel is available. Designed for students, engineers, makers, and aerodynamic enthusiasts, NeuraFOIL makes it easy to compare airfoil behavior, test design ideas, and build intuition about how geometry and flow conditions affect aerodynamic performance. Predictions run locally on your device, so results are fast and available without an internet connection. NeuraFOIL is ideal for learning, early-stage exploration, and quick aerodynamic comparison. It is not intended to replace validated CFD, wind-tunnel testing, or certified engineering analysis.

Are you interested in fluid dynamics? Viscous Joukowski simulates the flow around a Joukowski airfoil by solving the two-dimensional incompressible Navier–Stokes equations directly on your device. For teachers, students, and curious aerodynamicists, it provides an interactive way to explore how viscosity changes lift, drag, pressure, boundary layers, separation, and wake formation. Watch vorticity, speed, pressure, and streamlines evolve around the airfoil. Change the angle of attack, thickness, and camber by dragging directly on the flow field, or use precision controls. Adjust the Reynolds number to explore steady flow, separated wakes, starting vortices, and vortex shedding. Surface-pressure curves and aerodynamic-force estimates update as the simulation develops. The app also includes the classical analytical Joukowski potential-flow solution. Overlay it on the computed viscous pressure distribution to see where the ideal textbook model agrees with—or departs from—the Navier–Stokes result. In the potential solution, circulation is selected using the Kutta condition; in the viscous simulation, circulation and the wake develop dynamically from the no-slip flow. All calculations run locally using Apple MLX and GPU-accelerated Fourier methods. No server, account, or cloud computation is required.