Optimal Anti-seismic Research Of Steel Frames With Semi-rigid Connections Based On Genetic Algorithms

The semi-rigid connections steel frames have adequate response to energy consuming behavior and to earthquake loading. It solve the problem of rigid joint ductility. The semirigid joint can not only meet the concept design requirement of strong column-weak semibeam,but also content the economy level of our country at present. It possesses latent economic benefit. And optimization design is a cross subject, it can not only reduce project cost, but also improve structural performance. It is the important research direction of engineering design. China is a country with frequent earthquakes, it appears more and more important to apply optimization under the dynamic situation at the same time. The work of this paper mainly followed the subject.Structure analyse is the foundation of structure optimization design. Finite element analysis software ANSYS is used to perform the dynamical analysis. Taking the semi-rigid joint as a spring element, semi-rigid frame model is set up to calculate and analyze. Mode and seismic time–history analysis is taken into account of the effects of connection flexihility and geometric nonlinearity. The influence of connection stiffness on natural frequencies, displacement and response and column bottom shearing force of structure is analyzed.All of these above are the foundation for more optimization to the semi-rigid connections steel frames.In this paper ,the basic theory of GA and the improved algorithm based on real number encoding was introduced. The selection, crossover and mutation algorithm of genetic algorithms and other parameters coded were designed. Competitive elitist mode was applied to handle restriction seismic optimization problems.Through APDL language, we achieved the joint of intelligent optimization algorithm GA and finite element structure analyse. The seismic optimization model of the steel frames with semi-rigid connections are put forward, the structure optimized design program is debugged in MATLAB and applied to the example. The results of the optimal designs proved improved algorithm is feasible and applicable.