Approximation techniques for broad-band acoustic radiated noise design optimization problems

This study investigates the applicability of various approximation methods to broadbandradiated noise design optimization problems. Low-order series approximationsof dynamic response may be used to replace full numerical system solutions to effectsignificant computer cost savings during design iterations. Also, the ease of evaluatingthe approximate functions may be further exploited by using global optimizationsearch methods, such as simulated annealing, at individual design iterations.The combination of approximating radiated noise spectra and evaluating the approximatespectra for all possible design alternatives greatly increases the possibilityof finding a truly optimal design. The effectiveness of the approximations is measuredby considering optimization accuracy, evaluated by the algorithm’s ability to finda global or near-global minimum independent of the initial design; computationalefficiency, based on the number of numerical design analyses required for convergence;and generality, where the method should be relatively independent of theproblem type. Finite element models of three test cases with varying performancegoals and design parameters were used to evaluate the optimization methods. Shellthicknesses, shell loss factors, and rib stiffener locations were varied to minimizestructural weight and manufacturing costs while lowering broad-band radiated noiselevels below a specified goal. First-order Taylor and half-quadratic series approximationoptimization approaches were compared to traditional local minimizationmethods (Modified Method of Feasible Directions and Broydon-Fletcher-Goldfarb-Shanno). For all test cases, the approximation approaches found the global optimumdesign more frequently than the local minimization methods. Also, the half-quadraticmethod converged using fewer design evaluations than the first-order Taylor method for most test cases.