Multi-objective optimization of double-walled steel cofferdams based on response surface methodology and particle swarm optimization algorithm

Abstract

To improve the safety of steel cofferdams and reduce the cost to the extent possible, a multi-objective optimization method based on response surface methodology (RSM) and particle swarm optimization (PSO) algorithm was proposed to optimize the structure of double-walled steel cofferdams in the deep-water bridge foundation construction. The thickness of the inner and outer wall plate, the height and thickness of the inner-wall brace and the vertical spacing between annular plates were selected as design variables, taking the stress of the inner-wall brace and inner wall plate as the optimization objectives and the cost as the constraint condition, an optimization model of double-wall steel cofferdam was proposed. On the basis of the finite element calculation results, the experimental design and RSM were used to obtain the explicit relationship between the design variables and the structural response of the cofferdam. The optimal values of the design parameters of the cofferdam were obtained using PSO algorithm. The optimal design parameters and the structural response of the cofferdam variation with the cost were investigated. It is shown that the combination of RSM and PSO is able to optimize the cofferdam structure, with a stress optimization rate of 49.8% for the inner-wall brace while keeping the original cost constant. In addition, by adjusting the cost threshold, the reasonable range of cost coefficient K for this cofferdam is 0.119 to 0.162, and the section size of inner-wall brace is the key design parameter for the cofferdam, which provides a design idea for the optimization of deep-water cofferdam structure.