Stochastic Optimization of Aeroelastic Response (SOAR)

Using MSI?s Stochastic Optimization of Aeroelastic Response (SOAR) framework, time-savings of several-fold can be achieved during the airframe conceptual design stage. The availability of fast and inexpensive computers along with higher-fidelity design models has provided the opportunity to optimize aeroelastic behavior and consider uncertainty during the conceptual design phase. However, computational time becomes prohibitive in conceptual airframe development due to the many iterations that must be performed to optimize weight and/or lift/drag ratios, while ensuring that aeroelastic flutter is avoided, particularly when uncertainty effects are included. Automated Multi-Disciplinary Optimization (MDO) programs have attempted to address this shortcoming, but recent attempts have had only limited success, due to problems with slow convergence and local minima and due to the computational expense of performing uncertainty analysis. To provide a step improvement, MSI is developing an automated, robust and reliable multidisciplinary optimization method which uses a stochastic approach to avoid local minima, and a compromise response surface method to determine system performance and constraint probability density functions due to both controllable (design variables, operating conditions) and uncontrollable (material defects, wind gusts) random effects. The new SOAR framework combines uncertainty analysis with optimization techniques to obtain an integrated and powerful approach to optimal design.