基于响应面法的无轨伸缩式门式起重机轻量化设计

依赖于传统经验方法的无轨伸缩式门式起重机结构设计往往趋于保守，材料浪费严重，为了充分发挥材料的承载性能，有必要对它进行优化分析。对无轨伸缩式门式起重机结构进行优化设计时，针对非线性接触模型计算不收敛与计算效率低等问题，提出以节点耦合模型代替接触模型来构建响应面近似模型，并将近似模型与多岛遗传算法相结合对结构参数进行优化。以某无轨伸缩式门式起重机各构件截面尺寸为初始设计变量，以结构强度、刚度、自重为模型响应，通过调用Isight平台中最优拉丁超立方设计方法产生设计变量样本点，再由试验设计（design of experiment，DOE）模块调用有限元分析软件ANSYS完成样本点模型仿真以筛选出对响应影响较大的设计变量。利用优选后的设计变量构建响应面近似模型，以结构强度和刚度为约束条件，门式起重机结构质量最轻为优化目标，采用多岛遗传算法对响应面模型进行优化。结果表明：在保证结构性能的前提下，优化后门式起重机质量减轻23.4%，轻量化效果明显。所提出的优化策略可较快获得全局最优解，减少了计算量，为无轨伸缩式门式起重机结构改进提供了理论依据。

Lightweight design of trackless telescopic gantry crane based on response surface method

The structural design of the trackless telescopic gantry crane which depends on the traditional empirical methods tends to be conservative and material waste is serious. In order to take full advantage of the bearing capacity of its material, it is necessary to do the structure optimization analysis. During the optimization design of the trackless telescopic gantry crane, in view of the existing problems such as the non convergence and the low computation efficiency of the nonlinear contact model, a new optimization method which was used the node coupling model to replace the contact model was proposed to construct the response surface approximation model, and the model structural parameters were optimized by combining the approximation model and the multi-island genetic algorithm. Taking the component cross-section size of the trackless telescopic gantry crane as the initial design variables, the structural strength, stiffness, weight as the model response, the design method of optimal latin hypercube in the Isight platform was invoked to generate the design variable sample points, and then the finite element analysis software ANSYS was invoked by the DOE module to complete the sample point model simulation to screen out the design variables which had a high effect on the structural response. On this basis, the optimal design variables were used to construct the response surface approximation model, with the structural strength and structural stiffness as the constraints, the lightest structure weight of gantry crane as the objective, and the multi-island genetic algorithm was used to optimize the response surface model. The results of optimization indicated that the structure weight of gantry crane could reduce 23.4% in the premise of ensuring the structure performance of gantry crane, and the lightweight effect was obvious. The proposed optimization strategy can quickly obtain the global optimal solution and greatly reduce the computational quantities, which provides the theoretical basis for improvement of the trackless telescopic gantry crane structure.