| Reactor is one of the important equipment in power system,and mainly plays the role of a current limit,filter and compensation.It usually consists of reactor entity and post insulator,and has structure characteristics of high center of gravity,large slenderness ratio and large head quality.It is very bad for seismic,and damaging phenomenon for connection parts between post insulator and reactor components often occurs.And along with the increase in power grid capacity and the continuous improvement of voltage grade,reactor voltage level and capacity were also constantly improved,which made the overall height of equipment increase significantly,so reactor was more likely to be brought safe hidden trouble for bending and seismic intensity.In this paper,for the disadvantages of shape memory alloy(SMA)damper and piezoelectric friction damping device,considering characteristics of reactor equipment and requirements for structural damping control;a new type of SMA and piezoelectric friction hybrid damping device was developed,and combined with the intelligent control technologies,such as artificial immune algorithm,BP neural network and fuzzy control algorithm;and the theoretical analysis and experimental research based on SMA-piezoelectric composite damping system for reactor structure were done.The main research content is as follows:(1)For the problem that genetic algorithm is easy to fall into premature convergence and has the poor population diversity,and based on the clonal selection,immune memory and immune self-adjusting mechanism in biological Immune system,the adaptive immune memory clonal algorithm(AIMCA)was put forward.The modal controllable degree was taken as the influence factor of the optimization goal criteria,and improved genetic algorithm(IGA)and AIMCA were respectively used to discuss the optimization configuration of dampers in a space net shell structure which has 85 nodes and 288 bars.Results show that for the optimization problem of dampers in the complex structure,AIMCA has a stronger optimization ability and better population diversity,and can achieve greater performance index and better damping effect.(2)The mechanical properties test for the developed hybrid damping device was carried out;the influences of excitation voltage and displacement amplitude on the lap energy dissipation,equivalent damping ratio and equivalent secant stiffness were analyzed.Results show that the hybrid damping device can have the force in two-way and full hysteretic curve of the basic symmetry.It also has better energy dissipation capacity and stable performance.Load frequency has little impact on its mechanical property;with the increase of voltage,the absolute maximum of control force of damping device linearly increases.And at 12 mm displacement amplitude and 120 v voltage,the consumed energy can increase by 138.23%,the equivalent damping ratio can increase by 94.23%;the semi-active piezoelectric unit has a better energy dissipation ability.(3)Based on the mechanical properties tests of SMA and SMA-piezoelectric friction hybrid damping device,BP neural network prediction models using two neurons input strategy separately were established.The weight and threshold values of neural network model for damper were optimized by using AIMCA.The results showed that compared with SMA network model which taken the stress and strain of former moment and former-former moment as input neurons,due to reduced neuron input parameters,the damper model has a relatively large fluctuation and its prediction accuracy is reduced.The optimized BP network improves the precision and stability of the forecasting model.BP neural network model can synthetically consider various factors to predict the output force of hybrid damping device,and convenient for implementation in the MATLAB simulation,which provides a new way for established and application of constitutive model hybrid for SMA-based damping devices.(4)The continuous Bouc-Wen model was chosen to simulate the nonlinear restoring force of the structure;BP neural network model after optimization was used to determine the control force of damping device;the output voltage was obtained via fuzzy control;the semi-active hybrid control simulation analysis for the seismic response of the structure was come true.The result shows that according to the characteristics of damping device,and combining with the artificial immune clonal algorithm,BP neural network and fuzzy control technology;the established hybrid control system can really-time adjust friction resistance of piezoelectric actuators according to the dynamic responses of structure,and realize the hybrid semi-active control for the structure.(5)A 10 kV dry-type air-core reactor model whose similar ratio is 1:2 was designed and made;the shaking table test was finished under no control,passive control and hybrid control;the dynamic characteristics and seismic response for model structure at the different seismic waves was analyzed.The results show that the developed SMAPZT friction hybrid damping device can effectively inhibit the dynamic response of the structure and realize seismic control for damping reactor equipment.Generally,the shock absorption rate for displacement and acceleration can reach 40% at passive control,and up to 50% at hybrid control.In addition,after test no weak parts of the reactor structure appear in earthquake damage,which show that the damping device can improve the aseismic reliability of the reactor structure. |
PDF Full Text Request