Optimization design of unitized panels with stiffeners in different formats using the evolutionary strategy with covariance matrix adaptation

Abstract

Curvilinear stiffener concept has been introduced to aircraft panel structures most recently for possible further weight reduction during optimization design. However, due to enlarged design space and high design complexity, more computational time is needed to optimize curvilinearly stiffened panels. Considering the requirement for both lighter structure and less design time, optimization designs of a unitized panel with stiffeners in three different formats, i.e. curvilinear, oblique, and evenly distributed straight, are conducted under various loading conditions to figure out which stiffener format should be selected in a real design environment. Four single loading cases with different compression-shear ratios and a set of multiple load cases are considered. Evolution strategy with covariance matrix adaption is combined with sequential quadratic programming to seek out the optimum structure with minimum mass taking into account buckling and strength constraints. Comparative analysis of the optimization results indicates that evenly distributed straight format is suitable for compression-dominant loading conditions whereas curvilinear format become superior in case that shear loading is considerable or multiple biaxial compression and shear loads are applied. Besides, curvilinear format adds more design flexibilities to the stiffeners, which enables them to play a more important role in the optimized structures. Furthermore, evolution strategy with covariance matrix adaption is found to be more efficient than particle swarm optimization for this optimization problem. This study can provide a useful guidance for future optimization design of aircraft structures.

A new model based on belief rule base and membership function (BRB-MF) for health state prediction in sensor

“…The CMA-ES algorithm is able to solve nonlinear and non-convex real continuous optimization problems. [24][25][26] This model contains four steps, whose specific calculation process is shown in literature. 27…”

Health State Prediction of Aero-Engine Gas Path System Considering Multiple Working Conditions Based on Time Domain Analysis and Belief Rule Base

“…The initial parameters of the proposed model are given by experts while there is a certain degree of subjectivity in expert knowledge, which will lead to low accuracy of health state prediction results. Therefore, this paper uses the CMA-ES optimization algorithm to optimize the initial parameters given by experts to further improve the accuracy of the gas path system health state prediction model [34,35].…”