Skip to content

Learning Aerodynamics Through Data to Improve Optimization Algorithms

Notifications You must be signed in to change notification settings

mid2SUPAERO/PIR_Mario_Leupolt

Repository files navigation

The calculation of the aerodynamic coefficients of an airfoil with established methods is a time consuming and computationally expensive process. This work takes a deep learning approach to determine the aerodynamic coefficients of a given airfoil much faster. It uses an already existing data base and tries three different platforms (Keras and SMT package in python and Monolith AI) to predict the coefficients as accurately as possible. The newly obtained surrogate model can then be used to give scalar and graph predictions for certain Mach numbers and angles of attack or in optimization algorithms to calculate the aerodynamic coefficients in the optimization steps.

About

Learning Aerodynamics Through Data to Improve Optimization Algorithms

Topics

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages