The Reliability Optimization Of Bridge Crane Girder Based On RBF-NN

Cranes are widely used in all walks of life.With the rapid development of the social economy,the performance of cranes is also required to be higher.The crane metal structure is considered to be the skeleton of the crane,and its quality accounts for 60% to 80% of the mass of the whole machine.Lightweight design is particularly important;in addition,the metal structure is subject to complex external loads in addition to its own weight.The faults occur frequently and are prone to accidents.In the annual mechanical equipment accidents,the accidents caused by the cranes are always in the first few places,which will seriously endanger life and have a negative impact on China’s economic development.Therefore,the safety of the crane is urgent.solve.Therefore,the metal structure must consider both economic and reliability requirements at the design stage.The reliability-based design optimization method is a new optimization design method developed on the basis of the conventional optimization design method.It combines the reliability analysis theory with the optimization theory.The application of reliability optimization to the design of crane metal structures not only ensures the safety of the crane but also meets the economic requirements.With the maturity of various computer simulation software,the current engineering structure design and analysis is increasingly dependent on computer modeling.Whether it is structural optimization design or structural reliability analysis,most of them need to be realized by means of finite element simulation tools.This process often requires repeated calls to the finite element model during the iterative process to find the most reasonable design.With the improvement of various modeling complexity and the improvement of the analysis accuracy requirements of designers,the efficiency improvement caused by computer simulation analysis is very limited.This paper is aimed at the problem of high number of calls and high computational cost in reliability optimization based on finite element simulation technology.The Surrogate Model is constructed based on the finite element model of strength,horizontal static stiffness,vertical static stiffness and vertical dynamic stiffness of the main girder of bridge crane.,using the established Surrogate Model for reliability optimization.The Surrogate Model uses the radial basis neural network algorithm(RBF-NN)in the field of artificial intelligence.After reliability optimization,the cross-sectional area of the main girder is reduced by 8.0% compared with the initial design,and the reliability of each performance value is above 0.99.The whole process of this paper takes 3281 s,which is much smaller than the time used for optimization based on the finite element model.The results show that the proposed method greatly reduces the computational cost and improves the reliability optimization efficiency while ensuring the economic and safety of the design.