Study On Continuous Variable Stiffness Control System Based On Genetic Algorithm

Semi-active variable stiffness control system is one of classical semi-active control systems.Semi-active control systems have the advantages of vibration control combined with passiveand active control systems which is the leading subject about vibration control applied in thefield of civil engineering. And attracted extensive attention has been paid in recent years.Control algorithm design is the main part of the study for structure vibration control system.Design control algorithm is the most important aspect for semi-active control systems whichinfluence control effect for the control systems directly. The semi-active control structure modelis highly non-linear performance. Then the general control algorithm is hard to search for theoptimal solution, and this paper is providing a genetic algorithm to solve complex optimizationproblems in common frameworks. In order to solve the problem within the coding space byencoding series next into groups, the optimal solution can improve search efficiency bysimulating biological evolutionary process called survival of the fittest.The design of continuous variable stiffness control algorithm in this paper is based ongenetic algorithm theory and Matlab language. According to continuous variable Stiffnesscontrol system, this paper design a control algorithm including additional stiffness value whichas the "chromosome", the binary coding, performance function, operation genetic factors whichincluding selection, crossover and mutation. Finally, the application of research results andsimulation analysis for a reinforced concrete structure construction and a simple beam bridgehad been made.The main contents of this paper are listed as follows：1）. Deficiencies of traditional semi-active variable stiffness control had analyzed inthis paper, such as its narrow vibration frequency bandwidth and limitedapplication. Analysis also showed that switch control is unfavorable to reduce theacceleration response. Because this bi-state control system between theinstantaneous state and control action is a rough match, full use of efficiency of thecontrol system has been failed to make.2）. Corresponding countermeasures had been proposed in this paper. In order toovercome the limitation of traditional variable stiffness control, this paper designeda control strategies using genetic algorithm. Genetic Algorithm is a new optimal algorithm which used the biologic populations evolve to search the best solution isdifferent from traditional methods. A new index function is presented and how tochoose performance index of the control in the algorithm of active variablestiffness control have been discussed.3）. Matlab language is used to write control algorithm procedures and numericalstructure dynamic response analysis program. Numerical simulations for theseismic response control of a six-storey structure under seismic excitation with thiscontrol strategy have been discussed. The results shown that the floor accelerationof the building could be reduced its displacement and acceleration simultaneously.Better controls effects were obtained simultaneously and broaden range ofstructural periods were received by this new kind of variable stiffness controlalgorithm. This control algorithm in this paper is feasible and effective.4）. Seismic response analysis using a model of the simple bridge as an example hasbeen discussed in no-control, traditional variable stiffness control by Lqr controlalgorithm, and the new intelligent variable stiffness control under the sameconditions.