Research On Hierarchical Optimization Method Of Large Span Spatial Structure Based On Discrete Variable Optimization And Topology Optimization

With the increasing demand for national social and cultural activities,China’s demand for large-span spatial structures is increasing.The traditional cross-section design scheme of spatial structure often depends on the experience and level of designers.The material utilization rate is not high,and the structural efficiency is low.It also does not fully play the broad optimization design space brought by the change advantage of large-span spatial structure in the spatial gradient.Limited by the requirements of construction,manufacturing and specification,the structural design needs secondary processing.In other words,in the past,the optimization design of spatial structure often stayed at the design level of the section size of macro members,and the optimization design of meso nodes was not realized by using advanced optimization methods.In addition,in spatial structure,the nodes were often in different stress service states,and it was particularly difficult to obtain the exact boundary conditions of node stress.It was difficult to achieve the matching design of lightweight macro structure and maximization of meso node stiffness.Therefore,this paper proposes an improved intelligent optimization algorithm and an optimization strategy based on multi-scale model for the hierarchical optimization of macro-structure-micro-nodes of large-span spatial,and studies the method of multi-platform optimization system to realize the size optimization of spatial structure members and the method of node topology optimization design for additive manufacturing constraints.The research work and conclusions of this paper are as follows:(1)Aiming at the shortcomings of the traditional intelligent optimization algorithm,this paper obtains the adaptive alternating variant genetic algorithm(AVOGA)suitable for discrete variable optimization through the improvement of the full operation chain.Three algorithm test functions and ten-bar truss optimization problems are used to compare the performance indexes of AVOGA algorithm and traditional intelligent optimization algorithm.The research shows that AVOGA algorithm has faster convergence speed,better algorithm stability and is not easy to fall into the trap of local optimal solution.(2)Based on the improved AVOGA algorithm proposed in this paper,a multi-platform data interactive collaborative optimization design system is established,and a multi-constraint and discrete optimization design model that meets the specification and conceptual design is established.The cross-section size of a single-layer reticulated shell under single and multiple working conditions is optimized.The research shows that reasonable member grouping and section value can effectively reduce the steel consumption,and give full play to the bearing capacity of member materials.AVOGA algorithm is suitable for complex structure optimization,and the algorithm is stable,while meeting the engineering time and economic cost requirements.(3)The rationality of node shape finding is based on the extraction of accurate stress state,and the traditional quadratic analysis method often has errors in extracting the internal force of the node.Based on the parallel uniform multi-scale principle,considering the double nonlinear buckling of initial geometric imperfections,multi-point constraint method is used to establish multi-scale models with different classifications.Compared with the full-beam element model,the multi-scale model takes into account the calculation accuracy and efficiency.The influence of node stiffness and plastic distribution on the structural calculation results can be investigated,and the internal force boundary transition from the macro structure to the meso node can be realized,which verifies the accuracy and applicability of the multi-scale model.(4)Hollow sphere node is one of the common node forms in spatial structure,and the innovative node on this basis has better spatial mechanical properties,but its manufacturability is limited by the singularity of the shape.In this paper,based on the section size optimization results of multi-scale models with different node classifications,the boundary conditions of the internal force of the node to be optimized are obtained by using the multi-scale model,and the optimization design domain of the node core based on the hollow sphere node is constructed.The additive manufacturing constraints are incorporated into the optimization algorithm of the variable density method.The additive manufacturing is realized by the deposition melting technology.Under the premise of ensuring lighter quality,the node optimization results have reasonable mechanical properties and more visual aesthetic effects.It is a work of art combined with force and beauty.