| Shape Memory Alloy, referred to SMA, which has a shape memory effect, pseudoelastic property, high damping, is a kind of smart material with good prospects for development in the field of structural vibration control. In this article, combining the austenitic SMA passive control system at room temperature with neural network and genetic algorithm, the seismic response of a spatial structure is optimization controlled. The main contents of this paper are as follows:(1) According to the characteristics of SMA materials, two kinds of SMA prevent sliding clamps with good performance are developed and 12 groups of 48 austenitic SMA wires' experiments on mechanical properties are performed. The impacts of loading cycles, loading strain amplitude, loading rate, material's diameter on the pseudoelastic property of SMA are studied. The results show that after cycling stability, loading cycles and loading rate have significant impacts on the energy consumption of SMA wires.(2)Considering the impact of loading rate on SMA mechanical properties, two strain rate related constitutive models, based on material experiments of austenitic SMA, are proposed. One type simplified linear constitutive model, the other one is a model based on smart algorithm, called GA optimized BP network constitutive model. Both constitutive models are simulated by MATLAB. The results show that the simplified constitutive model with simple formulas, can reflect both the mechanical performance of SMA and loading rate's impact on SMA stress-strain curve; The GA optimized BP network model, using experimental data as the training data of BP network, whose prediction curves fit experimental stress-strain curve very well and calculation errors are small, is an ideal type of rate related constitutive model.(3)The number of SMA in the passive control system may be changed during the arithmetic process when using traditional GA to optimize the configuration of SMA, which will greatly reduce optimization efficiency. In order to solve the shortcoming, an improved genetic algorithm is proposed in this article. The traditional GA encoding, crossover and mutation operators are improved, and corresponding algorithm programs are written by MATLAB programming language. The configuration of SMA passive control system in the two cross three floors spatial structure is optimized with the improved genetic algorithm, and optimal results are obtained.(4)The dynamic finite element model of the two cross three floors spatial structure under seismic excitation is created by using MATLAB programming language. GA optimized BP network constitutive model, proposed in this article, and optimal scheme of SMA passive control system obtained by the improved genetic algorithm are applied to the spatial structure's seismic response optimization control simulation. The acceleration response and displacement response with optimal allocation of SMA passive control system and without layout SMA passive control system are compared. In order to verify the effectiveness of SMA passive control system's vibration control under earthquake excitation, the shaking table model experiment of spatial structure, designed in this article, is performed. The results of both experiment and MATLTB simulation show that: The optimal allocation austenitic SMA passive control system can significantly reduce the seismic response of structure. In general, compared with uncontrolled structure, the seismic responses of the controlled structure can be reduced by more than 15%, and the damping effect at the bottom of the structure is more obvious than the upper of the structure. The acceleration responses of two floors at the bottom of structure can be reduced by more than 20% and the interlayer displacement responses can be reduced by more than 40%. |
PDF Full Text Request