冯宇韬, 李维滨. 基于响应面模型法的地下综合管廊结构优化方法研究[J]. 云南大学学报(自然科学版), 2022, 44(4): 791-799. doi: 10.7540/j.ynu.20210429
引用本文: 冯宇韬, 李维滨. 基于响应面模型法的地下综合管廊结构优化方法研究[J]. 云南大学学报(自然科学版), 2022, 44(4): 791-799. doi: 10.7540/j.ynu.20210429
FENG Yu-tao, LI Wei-bin. A study on structural optimization of utility tunnel based on response surface model method[J]. Journal of Yunnan University: Natural Sciences Edition, 2022, 44(4): 791-799. DOI: 10.7540/j.ynu.20210429
Citation: FENG Yu-tao, LI Wei-bin. A study on structural optimization of utility tunnel based on response surface model method[J]. Journal of Yunnan University: Natural Sciences Edition, 2022, 44(4): 791-799. DOI: 10.7540/j.ynu.20210429

基于响应面模型法的地下综合管廊结构优化方法研究

A study on structural optimization of utility tunnel based on response surface model method

  • 摘要: 地下综合管廊是城市地下空间开发的重要方式,为了对地下综合管廊的结构进行优化设计,提出了一种基于响应面模型(Response Surface Model, RSM)的结构优化设计方法. 基于有限元模型分析结果,选取适当的结构力学响应作为控制参数. 采用拉丁超立方采样(Latin Hypercube Sampling,LHS)和优化拉丁超立方采样(Optimized Latin Hypercube Sampling,OLHS)方法,建立设计参数与结构力学响应之间的响应面模型,并评估了不同采样方法建立的响应面模型的精度. 结果表明,采用最小样本点数量4倍的OLHS建立响应面模型能兼顾建模效率与准确性. 基于响应面模型建立了考虑配筋等多种条件约束的优化模型,采用非线性规划和遗传算法对优化模型进行了求解,结果表明求解方法不影响优化结果,模型鲁棒性好. 优化后结构造价成本下降27.58%,控制截面弯矩分别下降45.10%、45.72%、24.40%. 采用有限元方法对优化后结构进行了分析,其二维响应面模型与有限元分析的误差小于14%,采用二维响应面模型替代有限元分析过程是合理有效的.

     

    Abstract: Underground utility tunnel is an important way for urban underground space development. In order to optimize the structure of the utility tunnel, this paper proposes a structural optimization design method based on the response surface model. Based on the analysis results of the finite element model, the appropriate structural mechanics response is selected as the control parameter. Using Latin hypercube sampling (LHS) and optimized Latin hypercube sampling methods, the response surface model between design parameters and structural mechanics response has been established, and the accuracy of different response surfaces models has been evaluated. The results show that the optimized Latin hypercube with 4 times of the minimum number of sample points which has been used to establish a response surface model can balance efficiency and accuracy. Based on the response surface model, an optimization model considering various constraints such as reinforcement is established. The optimization model is solved by nonlinear programming and genetic algorithm. The results show that the solution method has less influence on the optimization result and the model has good robustness. After optimization, the cost of the structure was reduced by 27.58%, and the bending moment of the control section was reduced by 45.10%, 45.72%, and 24.40% respectively. The optimized model was analyzed by the finite element method, and the results showed that the error between the response surface model and the finite element analysis was less than 15%. It is reasonable and effective to use the response surface model to replace the finite element analysis process.

     

/

返回文章
返回