Font Size: a A A

Intelligent Algorithm-based Seismic Behavior Optimization Of Structural System With Self-centering Energy Dissipation Brace

Posted on:2019-02-03Degree:MasterType:Thesis
Country:ChinaCandidate:X T YanFull Text:PDF
GTID:2382330545972170Subject:Structural engineering
Abstract/Summary:PDF Full Text Request
Performance-based structural design concept defines three performance levels:life safety,structural damage and functional insurance.Among these levels,structural damage level is more simply and easier to quantify in practical engineering.According to the damage level of overall structure and component performance groups,the elasto-plastic analysis method and the Monte Carlo method are used to simulate the cost of structural reapir after earthquake,which has a certain practical engineering significance.The exact description of structural damage is the key to achieving this design method,and the designers are supposed to improve the seismic performance of the structure significantly without any cost increase.In addition,in order to make the structure have a small residual deformation after earthquakes,structural systems with resilient components have also emerged.The design parameters and reasonable arrangements optimization of these components have become the research hotspots of domestic and foreign scholars.This study applies more accurate damage indices and brace mechanical model to evaluate and optimize the seismic performance of structures,and proposes the seismic behavior optimization method for pre-pressed spring self-centering energy dissipation(PS-SCED)bracing system based on genetic algorithm(GA)and Back Propagation(BP)algorithm,the main research contents are as follows:(1)The boundary convergence and variable discreteness of the Park-Ang dual-parameter damage model of reinforced concrete(RC)columns are modified.Through the theoretical derivation,the influence of P-? effects on the member's lateral displacement is obtained.Based on the test data of seismic performance of multiple groups of RC columns,two modified models are obtained by logarithmic fitting and Bayesian parameter updating method according to the lateral stiffness of components,the spacing and the constraint type of transverse stirrups.The recommended values of the variable parameters are given as well.A method for calculating the unloading stiffness using the residual displacements and skeleton curves of RC shear wall components is proposed.The applicability and accuracy of the stiffness degradation model and the Kunnath bending damage model are compared and verified by the experimental data when describing the damage evolution of RC shear wall components.(2)Combined with the BP algorithm,a new seismic performance optimization method for the RC structure is proposed.Firstly,the local structural damage indices due to earthquakes are solved.Taking the preset component damage combination coeffients as weight,the predicted value of the global damage index is calculated.Secondly,the error between the predicted value and the global damage index calculated by themacroscopical response is back propagated to modify the damage combination coeffients,which are used to define the structural seismic sensitive components.Combined with the performance-based evaluation criterion,a six-story RC frame structure is taken as an example,and the sectional demensions and corresponding reinforcements of the seismic sensitive components are enlarged.Results indicated that the BP-based seismic performance optimization method can reduce the inter-story drift ratio of the structure and the exceeding probability of each limit state without increasing the project cost.Meanwhile,the global damage of structure is mitigated.(3)A multi-objective failure mode optimization method of RC frame-shear wall structures based on GA is proposed.The method takes the sectional dimensions as the optimization variables,the material consumption as the constraint condition,the maximum inter-story drift ratio and the overall structural damage index as the objective function of the algorithm,and the hybridization and mutation of gene sequences are used to achieve the delivery of favorable genes.A five-story RC frame structure is simplified to a centralized quality system to verify the correctness of the optimization algorithm.Taking a 10-story RC frame-shear wall structure as an example,incremental dynamic analysis(IDA)is used to determine its sensitive ground motion and corresponding peak ground acceleration(PGA),which is used as ground motion input during the optimization process.The structure is subjected to static elasto-plastic analysis to obtain the yield and ultimate displacements,which are used to calculate the global damage index.The linear weighted method of multi-objective minimum optimization problem is proposed,and the algorithm convergence of each structural performance index is evaluated.The GA based multi-objective failure mode optimization method of RC frame shear wall structure has a good robustness.After the failure mode optimization by this method,the maximum inter-story drift ratio and the global damage value of structures are significantly reduced.The distributions of the inter-story drift ratio are more uniform.The structural failure mode near the critical collapse state transfers from the bottom story failure to the top story failure.The mean annual frequency of each structural limit state decreases and the collapse margin ratio(CMR)increases as well.(4)An improved mechanical model of PS-SCED brace is proposed,which can identify the brace working stage according to the indication of the relevant state variables.The mechanical behavior tests of a PS-SCED brace are carried out,and the correctness of the secondary development of brace mechanical model is verified in ABAQUS.Based on the BP algorithm,a new instllation strategy of braces is proposed.Taking a 9-story brace-steel frame structure as an example,the PS-SCED braces are arranged next to the seismic sensitive components,and the seismic performance of this braced structure is compared with that of the structure employing a continuous installation strategy along height.The structure with optimal arrangements of the braces based on BP algorithm has more energy dissipation capacity,and the dispersion of brace energy disspation is reduced as well.The local damage is effectivly improved and the seismic performance of structure is enhanced.(5)A linear acceleration method is used to solve the equation of motion of a multi-degree-of-freedom system with braces.In order to optimize the preliminary design scheme,a multi-objective optimization method of brace parameters based on GA is established.The direct integration method(DDM)is employed to derive the sensitivity of structural response to the brace parameters,which is used to correct the change rate of parameters in the GA module.A multi-objevtive GA-based brace prameters optimization is carried out to a 10-story brace-steel frame.Results indicated that parameters that are more sensitive to structural response should have a smaller rate of change in GA.The maximum inter-story drift ratio of structure with optimized brace parameters by this method is decreased,and the distributions of the inter-story drift ratio are more uniform.Morover,the convergence of this method is the best during the optimization.The results of this study will help to improve the systematic design process of the PS-SCED braced structure.
Keywords/Search Tags:RC structure, self-centering energy dissipation bracing structure, BP algorithm, genetic algorithm, seismic performance optimization, performance index, Bayesian method, incremental dynamic analysis
PDF Full Text Request
Related items