Genetic Algorithm And Optimality Criteria Method Based Optimum Design For Tall Buildings

The design of tall buildings not only needs to satisfy the requirements of structural safety and building functions,but also should achieve the reasonable structural system with the minimum material.The present design method of trial-verification-modification always results in the acceptable design project that satisfies the regulations of the design codes, however,it may be an unoptimal design and the design is time-consuming.Structural optimization dsign is an important development of the structural design,whose inherent aim is not only to pursue the minimum cost or material volume,but also the rational distribution of the resourses.The research on structural optimization had a long history,and there were a lot of successful applications in many fields and some available commercial FEM softwares with optimization modules.However,there has been few research and application of the structural optimization for building structures.In general,the optimization problem for building structures is an complex discrete optimization one with multi-load,multi-variable, multi-constraint and multi-objective,and there exsits a lof of uncertainties(loads,materials, analsys models,and etc.).Moreover,the structural system and form and the component type of buildings are quite complex and with numerical regulations of the design codes.These difficulities make it much challenging to establish a comprehensive and practical optimum algorithm.This paper uses the genetic algorithm and the criterion method as a tool for the optimum design of building structures,and focuses on the multi-objective optimization of beam bars, size optimization of the component sections,multi-objective optimization of structural scheme and topology,multi-objective damage-reduction seismic design.Some optimization modules were integrated to the commercial softwares of structural analysis and design.The main contents of this paper are listed as follow:1.Based on the deep understanding of the relevant provisions of the design codes,this paper proposes the template of beam reinforcement,which not only accounts for the feasibility of the design and the construction convenience,but also the design experiences and standard atlas.The optimum reinforcement detailing,which satisfies the strength,serviceability, ductility and other constraints related to the user specified rules and regulations in design codes,is achieved by constructing the data sets containing different available reinforcement bar diameter in a pre-specified pattern.A target evaluation system of beam reinforcement is given.Fuzzy analysis method is used to obtain the relative membership to the optimum design.At last,a genetic algorithm(FMOGA) with fuzzy analytic decision for solving the multi-objective optimization problems is presented.2.Based on the current Chinese design codes,a 2-level hybrid method of optimum design for structural optimum design of reinforced concrete(RC) tall buildings under multi-load cases is proposed.The entire optimum design procedure is divided into two levels:element optimum design and system optimum design,A genetic algorithm is used to perform the discrete optimization of tall buildings under the strength constraints with the results of system optimum design as the lower bounds of member size,then the system optimum design is conducted using optimality criteria method with the results of element optimum design as the lower bounds of member size.This procedure is repeated until convergence is obtained.An adaptive feasible region technique is developed to improve the efficiency of genetic algorithm.The proposed optimum design method is integrated to the large-scale finite-element software ANSYS and the specialized design softwrae for building structures, SATWE.Finally,a practical frame-shear-wall structure(30-story) is optimized to illustrate the effectiveness and practicality of the proposed optimum design method.3 The structural displacement based seismic design is an important approach to realize the performance-based seismic design.For the RC frame structure,the story drift is descided by the deformation of beams and columns,which is influenced by the axial compression ratio, the reinfoecement ratio and the characteristic value of the confining reinforcement.According to the stress-strain relation of confined concrete,the formulation of plastic hinge deformation capacity of beams and columns is established,and this formulation creates the relationship among the axial compression ratio,the characteristic value of the confining reinforcement and the ultimate plastic hinge rotation of beams and columns.The relationship between the story drift and the deformation of beams and column is determined by the virtual work.A 2-1evel hybrid method of optimum design is proposed for performance-based seismic design of RC frame structure,and a flow chart of the proposed criteria method based on the ineleastic analysis is given.4 Based on the literatures,the design experiences and the experts' suggestions,the relation between the influence factors and the structural scheme design is established,and the fuzzy evaluation technology is used to make multi-objective decision.According to the relationship between the structural design and the architectural design,the design template of building structures is proposed with the combination of structural elements,which provide a quantitive description for the scheme optimum design of building structures.Then,the multi-objective topology optimum design of the structural scheme of building is studied. 5.The basic concept,principle and optimization model of the damage-reduction seismic design of the complex structure are discussbed.It is suggest that the arrangement scheme of the damage-reduction elements should be considered as a multi-objective decision problem, and the target evaluation system of the arrangement of the damage-reduction elements is established according to the cost-benefit criterion.The multi-objective optimum design problem of damage-reduction element is studied,and a flow chart of the proposed method is given.