The Reliability Optimization Decoupling Method For Crane Metal Structure Based On Dynamic Kriging Models

As the skeleton structure of the crane,the metal structure of the crane needs to bear the working load as well as its own weight,so its design has a significant impact on the economic cost and safety of the crane.In the design stage of the crane metal structure(CMS),the use of reliability-based design optimization(RBDO)method has progressively become an necessity demand.In the wake of the modern simulation technique expend,finite element simulation analysis of CMS has already got the extensive application,but for large CMS,RBDO based on simulation is costly,how to reduce the cost of calculation and valid and accurate reliability optimization is badly in need of issue.At present,structural RBDO using the mathematical agent model technology has followed extensive interest from learned man,but in the RBDO field of CMS has just set in motion,and also faces many issues that need to be solved,such as CMS simulation technology model agent model is still a static model,the amount of sampling point demanded to build the model is high,the computational cost is high;for reliability optimization solution of the double-loop method,reliability analysis more times,the computational efficiency is low.In response to the above issues,this article conducted in-depth research,and the specific research content is as follows:(1)the reliability optimization strategy of the decoupling method is adopted.By separating the reliability analysis cycle from the optimization design cycle by means of offset vectors,which declines the frequency of reliability analyses and enhances the productivity of RBDO of CMS.(2)the dynamic Kriging model construction method on the basis of the multi-point addition strategy is proposed.Constructed an initial Kriging proxy model on the basis of a small number of high-fidelity sample points,During the process of iterative optimization,the agent model is updated by adding points through the two multi-point addition strategies proposed in the article,and Model precision improves as optimisation proceeds,and the agent model with sufficient precision is obtained at the end of the optimisation.(3)the construction method of dynamic Kriging agent model and the decoupling method of RBDO are integrated to propose a CMS reliability decoupling optimization method based on the dynamic Kriging model and applied to the RBDO of a bridge CMS,and the proposed method is verified by example under the condition of guaranteeing the reliability requirements,with very little computational cost,The proposed method achieves consistent results with the RBDO of the CMS based on the finite element model.Compared with the RBDO results based on the static agent model,the computational cost is reduced by approximately 76.52% under the same computational results,and the proposed method saves computational cost and improves design productivity while pledging the precision of the compute.