RBDO Method Research of Complex Structure Based SVRM

 

Li Yugang

1: Deepwater Engineering Research Center, Dalian University of Technology, Dalian, 116024 China
2: State Key Laboratory of Coastal and Offshore Engineering, Dalian University of Technology, Dalian, 116024, China

Ren Nianxin

1: Deepwater Engineering Research Center, Dalian University of Technology, Dalian, 116024 China
2: State Key Laboratory of Coastal and Offshore Engineering, Dalian University of Technology, Dalian, 116024, China

 

 ABSTRACT

Uncertainties and optimization are two major considerations in modern structural design, reliability- based design optimization (RBDO) is necessary. Traditional RBDO requires a double-loop iteration process, solving such nested optimization problems is extremely expensive for complex structure. For handling the difficulties associated with the problem, a new approach of RBDO is presented. In this research, a two-tier response surface approximation strategy is carried out based on support vector regression machine (SVRM) and converts the nested optimization problem to single-loop optimization problem, the first-tier of response surface is analysis response surface (ARS), which is fitted to limit state functions in terms of both design variables and random variables at various Latin hypercube samples. The second tier is design response surface (DRS) that is fitted to probability of failure as a function of design variables. Following reliability analysis, the optimization problem is solved by a global algorithm- particle swarm optimization (PSO). The overall performance of the technique is addressed referring to a trestle example.

Keywords: Reliability based design optimization; Support vector regression machine; Analysis response surface; Design response surface; Latin hypercube sampling; Trestle

Get the entire paper (pdf)    Go back to the TOC