Efficient global optimization and modal strain energy coefficient-based algorithm for fast prediction of dynamic aeroelastic loads

Design optimization of airframes is conventionally conducted employing a deterministic set of critical load cases defined through performing a sheer number of aeroelastic analyses spanning the aircrafts’ design space. Due to its large computational cost, the latter process cannot be integrated within the structural optimization cycle, hindering the development of a multidisciplinary design optimization (MDO) framework. In this paper, we present an algorithm for the efficient estimation of critical dynamic aeroelastic loads, as a first step towards the development of an integrated MDO platform. The method is based on the Kriging metamodeling technique along with the Latin hypercube scheme for initial sampling, and the expected improvement function for subsequent selection of sample points, known together as the efficient global optimization (EGO) algorithm. Furthermore, the use of the element modal strain energy coefficient (MSEC) is investigated as an inexpensive indicator to determine if a substantial change in the loads is expected after structural modification, thus triggering the necessity for the re-exploration of the loads’ design space. A case study is presented to evaluate the performance of the proposed methodology versus a stand-alone Kriging interpolation; a reduction of over 40% was achieved in the total time of execution employing the proposed algorithm.