Theoretical And Applied Research On System Reliability Based Optimization Of Passive Structural Control With Viscous Dampers

Structural safety has gained more and more importance under disastrous loads such as strong winds, earthquakes and ocean waves, etc. It is hard for traditional methods, e.g. to increase the stiffness, bearing capacity and material strength, to meet the current requirement. Therefore, the theory and engineering applications of structural control, which can promote the safety, serviceability and/or comfortability of structures, have being increasingly paid attention to by researches and engineers.As we know, the responses of structures can be reduced effectively by energy dissipation devices or dampers. However, the effect of seismic control is very sensitive to the number, space distribution and the physical parameters of the energy dissipation devices or dampers. The easiest strategy is to install energy dissipation devices or dampers in each story. However, such a strategy cannot always obtain the best effect of seismic control, not to mention its high cost. Thus, it has not been widely applied as yet in the field of engineering. Actually, up to now, most researches on the system optimization of passive control devices are still limited to the deterministic objective function. Such methods may find optimal arrangement of the control devices in some sense, but it might not be the best arrangement if more realistic factors are involved. On the other hand, due to the complexity of structures, the limitation of optimization methods and the randomness of earthquakes, optimization of structural control under earthquake becomes a multidisciplinary topic. Therefore, the study on topological optimization of the structures with dampers is of very significance and application value for civil engineering.Firstly, the most widely used theories for structural control and optimization algorithm are discussed in this paper. Then the topological optimization method for arrangement of the viscous dampers installed on frame structures is specially studied. Based on the system reliability of structures, the topological optimization of the viscous dampers is solved. The main research contents and results are shown as follows:(1) The encoding rules, population settings, selection of the fitness function, design of genetic operator and the convergence criterion of the genetic algorithm are further studied in this paper. Especially, some MATLAB toolbox functions are developed for the frame structures installed with viscous dampers. The analysis indicates that the improved genetic algorithm can quickly find the optimal individual without precocity phenomena.(2) Setting the deterministic indices, e.g. the inter-floor drift, as the objective function of the structural control, and using the improved genetic algorithm, the parameters of the optimization algorithm are defined. Then, the program for topological optimization of the viscous dampers installed in the frame structure is developed in MATLAB, in short as FSVDO. To the situations of classification II and III of the site soil, using the FSVDO program, the topological optimization analyses of the structures are carried out under realistic earthquakes and artificial earthquake. The result is consistent to that obtained by the method of exhaustion, which means that the Improved Genetic Algorithm (IGA) is very effective and reliable, and can be used in the optimization analysis and design of similar structures.(3) Even in the same classification of site soil, due to the randomness of earthquake, the optimal topology of the viscous dampers in the fame stru ctures can be different to different earthquake inputs, and such phenomenon can be referred to as "samples conflict". The analysis in this thesis also shows this phenomenon, which means the randomness of earthquakes has obvious effect on the topology of dampers. In other words, topological optimization of dampers with deterministic indices exhibits some defects, whereas the randomness of earthquakes can not be ignored.(4) With the probability density evolution method and the improved genetic algorithm, setting the reliability of structure as the objective function, a novel method which can solve the problem of "samples conflict" for topological optimization of dampers is proposed. The program is developed in MATALB, named as SRBTOVD. For comparison, the reliabilities of the un-controlled structure, the structure installed dampers in every floor and the structure installed topologically optimized dampers are calculated, respectively. The result indicates that, with the proposed IGA, not only the reliability of the structure system can be improved obviously, but the reliability is also very close to the structure installed dampers in every floor, whereas the number of dampers can be reduced to 50% of the traditional arrangement, which shows the advantage on economy and applications.(5) Due to the low efficiency of MATLAB, for the iteration calculation of topological optimization, C++ was employed as the programming language in the computing kernel program. With the C++ program, the system reliability of the structures with dampers can be obtained, and then the topological optimization analysis can be performed in the environment of MATLAB. The results of analysis indicate that, due to its high efficiency and practicability, the proposed method can be widely used in the civil engineering.(6) Using the probability density evolution method, under non-stationary random earthquake excitations, the evolution processes of the probability density of the uncontrolled structure, the controlled structure installed with dampers in each floor and the controlled structure installed with the topological optimized dampers are obtained, respectively. With the analysis on the processes of the probability density, dynamic reliabilities are evaluated and compared, which shows that, in the three cases, the controlled structure installed with dampers via topological optimization has the highest reliability and lowest cost.