Genetic Algorithm Optimisation of Mathematical Models — An Aircraft Structural Dynamics Case Study

Summary

In this paper, a technique for tackling inverse problems in structural dynamics will be reviewed. The processes applied here involve using the artificial intelligence optimisation tool known as genetic algorithms (GAs) where the optimisation problem to be tackled involves creating a structural dynamic aircraft model that best matches the available experimental frequency response function (FRF) data. It will be demonstrated how the inverse structural problem is difficult for more traditional optimisation approaches and how the stochastic processes inherent in a GA allow the problem to be tractable. Unsteady aerodynamic modelling is then used in conjunction with the structural models for aeroelastic analyses to determine the airspeed at which instabilities (commonly known as flutter) will arise. The implications of such model optimisation on aeroelastic analyses, and the estimates of uncertainty that may then be made, will also be briefly discussed.