Genetic Algorithm And Its Application Research On Optimization Design Of Space Grid Structure

Space grid structure is a sort of preferable roof type for long span buildings, which is widely used in public architectures, such as theaters, gymnasiums, aircraft hangars, gas stations, industrial premises and other places. It has the 3D transfer mode of force, save steel material, large rigidity, high degree of industrialization, beauty architectural sculpt, and so on.The genetic algorithm is a optimization algorithm, which belongs to bionics algorithm simulating biological evolution. Compared with the traditional optimization design method, genetic algorithm has strong adaptivity, global optimization and connotative parallel searching when it's used to solve complicated nonlinear and discontinuous optimization problems in practices.So genetic algorithm is given more attention, for instance, function optimization, artificial intelligence, engineering technology, image recognition, and so on.In view of space grid's wide applicability and genetic algorithm's superiority, the optimization design method based on genetic algorithm is analyzed in order to save materials and reduce cost in this paper. Therefore, main works are as follows:(1) In this paper, basic features, foundations of mathematics, basic ideas and implementation methods of genetic algorithm are researched. Refering to design-related specifications of space grid structure, the lightest weight grid structure as objective function and strength, stiffness, stability and geometries as constraints, etc, a reasonable mathematical model applied to grid structure section optimization design is established.(2) According space grid finite element with enetic algorithm, the dissertation writes the corresponding calculation program in Fortran language, which can realize automatic optimal design of space grid under static load. By comparing the calculated examples and the results of discrete variable optimization methods, the reasonableness of the genetic algorithm and the accuracy of the calculation program are verified. (3) Based on an project overview, with grid structure design-related specifications, using structural design software 3D3S and considering combination value of actions under constant loads, live loads, wind loads, temperature loads and seismic load combinations, the paper analyses the structure internal forces and displacement checking and draw the corresponding drawings. Afterwards, using general-purpose finite element analysis Software Midas/gen, the grid structure model is computed again. Besides, in order to acquire the lightest total weight of the grid structure, defining multiple sets of variable cross-section truss rods, optimum design is done based on Software Midas/gen and full stress criterion.The results of this thesis show that genetic algorithm as a emerging technology in the field of emerging optimization, is an intelligent optimization method. Compared with traditional optimization methods, due to the ability of global optimization and user-friendly computerized characteristic, the genetic algorithm is more suited to solve discrete combinatorial optimization problems. Discrete variables using genetic algorithm to optimize the design of grid structure cross-section has much advantage, such as simple principle, easy to convergence, reliable results and short computing time.Besides, optimum design results can meet the engineering design requirements, so it's valuable for engineering application.