【作者】 周海俊；

【导师】 孙利民；

【作者基本信息】 同济大学 ， 桥梁工程， 2005， 博士

【摘要】 作为斜拉桥主要承重构件的斜拉索，由于其质量轻、柔度大、阻尼低，极易在外界激励下产生各种类型的振动。为抑制拉索的大幅振动，在拉索锚固端附近安装阻尼器以提高拉索模态阻尼是目前使用最广泛的减振方法之一。然而，一些斜拉桥的拉索在采用减振措施后仍然发生了大幅振动，表明现有的工程技术对于拉索-阻尼器系统的振动特性和阻尼器参数优化等问题仍然没有能很好的把握并予以解决。本文回顾了拉索风致振动及其控制措施的研究现状，着重讨论了各类阻尼器特点、理论分析方法、半主动控制装置及其控制方法等。进行的研究工作主要包括：①进行了几类拉索减振阻尼器的单体试验。试验发现各类不同的阻尼器的力-速度关系符合不同的简化数学模型，粘性阻尼器较符合分数非线性模型，而油阻尼器(包括MR阻尼器)则符合线性或双线性模型。同时发现几乎所有的阻尼器均带有一定的刚度，且在阻尼器起动时可能具有很小的摩擦力。试验表明阻尼器的实际工作特性比较复杂，建立精确模型来模拟阻尼器的全部工作特性很困难。但在设计阻尼器的参数时，可以采用简化模型，针对拉索减振的要求，通过对安全系数的把握就可以指导减振设计时的阻尼器参数优化工作。目前，许多阻尼减振装置在结构和桥梁中已得到了应用，并有其相对应的简化模型和对应的优化设计方法。这些简化的模型可能不一定完全适合拉索减振的优化设计要求，但通过理论分析和试验验证，可以在修正后予以采用。②进行了附加11种阻尼器的斜拉索实索减振试验，论文详细描述了其中非线性粘性阻尼器、油阻尼器、MR阻尼器和摩擦型阻尼器的实索减振试验结果。试验发现无阻尼器拉索的内阻尼很小，在外界激励下振动难以衰减；而在附加阻尼器后，拉索的阻尼有了很大的提高。试验发现这四类阻尼器提供给拉索的对数衰减率值随拉索波腹点振幅及振动模态阶数的变化而变化；线性阻尼器设计方法无法解释该现象。将非线性粘性阻尼器试验结果与非线性设计方法计算结果比较，发现拉索所能获得的理论最大对数衰减率值远大于试验值，经分析是刚度的影响，该刚度可能是阻尼器的刚度，也有可能是由于阻尼器支架刚度的不足所导致。同时在粘性阻尼器试验中发现由于阻尼器内部构造或连接所导致的摩擦力而使得试验测得的对数衰减率值与理论计算值在小振幅段对应不是很好；在油阻尼器、MR阻尼器的实索减振试验中均发现了该摩擦力对拉索所能获得的对数衰减率在小振幅时的影响。试验发现由于支架刚度或连接的原因导致附加油阻尼器后拉索的对数衰减率均值相较于其它三种阻尼器要低，说明除阻尼器参数外，支架刚度和连接对于拉索减振具有同等的重要性。试验发现附加增强电压的MR阻尼器的拉索的阻尼特性可分为两部分，原因是由于有增强电压后的MR阻尼器力-速度特性符合双线性模型。在摩擦型阻尼器的实索减振试验中观察到了拉索所能获得的理论最大对数衰减率值；同时试验得到的拉索阻尼值与理论预测值在变化趋势上一致，但两者在阻尼器的起动振幅及最大对数衰减率值所对应振幅不一致，原因可能是由于理论计算假定不合理所导致。综上所述，由于各种非理想因素的影响，很难以线性设计方法予以解释，也难以用非线性设计方法完全预测其试验结果；这些因素包括阻尼器的内摩擦、内刚度、非线性粘性阻尼、支架刚度和连接非线性等等。因此，在实际工程中设计优化阻尼器参数时，需针对所采用不同阻尼器的非线性特性采用不同的非线性模型进行分析，同时设计时需保证拉索的支架具有足够的刚度，并在最终通过对安全系数的把握以提供拉索足够的阻尼。③分析了阻尼器的并联刚度对拉索模态阻尼比的影响。研究发现与忽略刚度影响的阻尼器类似，考虑刚度影响后的阻尼器同样存在着通用优化设计曲线。由于刚度的影响，拉索-阻尼器系统所能获得的最大模态阻尼比将显著的减小，而所对应的阻尼器最优阻尼值将要增大。将上述结论推广到非线性分数阻尼器，并将考虑刚度影响后的理论计算结果与粘性阻尼器实索减振试验结果进行了对比分析，发现考虑刚度影响后的理论计算结果与试验结果更接近。但同时也发现考虑刚度影响后的理论计算结果与试验结果在拉索小振幅区域对应仍不是很好，分析是由于阻尼器的摩擦力所导致。由于简谐振动时，阻尼和刚度并联关系可以转化为等效的阻尼和刚度串联关系，通过该换算关系，按模态阻尼比的折减率确定了拉索阻尼器支架的刚度要求，并进行了拉索减振工程应用的计算分析。④对双线性油阻尼器的参数优化进行了了计算分析。分析发现由于双线性阻尼器在小速度时表现出线性及大速度时表现出双线性的两种工作状态，拉索在这两种状态下表现出不同的阻尼特性。且这两种状态在切换时与拉索波腹点振幅是增大还是衰减相关，在设计时需仔细分析在这两个状态过渡段的阻尼特性，取不利值作为设计值。同时将MR阻尼器的实索试验结果与理论分析结果进行了对比，发现两者在变化趋势上吻合，但在小振幅时拉索的阻尼值由于存在着阻尼器的内摩擦而导致两者相差较大。由此进行了实桥拉索减振用双线性油阻尼器的参数优化设计计算。⑤针对阻尼器减振措施的减振效果受阻尼器安装位置的限制，提出在拉索端点设置形状记忆合金(SMA)作动器，通过加热、冷却方式以改变SMA弹模(长度)从而改变拉索内张力的半主动控制方法。进行了国产SMA材料的力学性能试验及模型拉索半主动控制试验。试验发现，采用SMA作动器可以有效的控制拉索张力的变化，SMA作动器的相变切换时滞约为0.3秒；说明采用SMA作为作动器控制拉索的张力变化进行半主动控制值得进一步深入的研究。但由于模型拉索试验中采用自由振动衰减测阻尼的方法评价制振效果，采用等间隔周期利SMA作动器施加扰动的方法制振效果不够理想，数值分析亦验证了试验结论。数值分析同时也表明，如果在拉索振动至不同相位时，通过改变SMA作动器的状态以改变系统刚度的制振方法减振效果明显，但这种控制方式需要SMA作动器具有更高的响应频率。更多还原

【Abstract】 Stay cables are very sensitive to environmental excitations due to low internal damping. The wide exhibited vibrations may induce fatigue or broken up of stay cables. To mitigate cable vibration, mechanical viscous dampers have already been installed to stay cables of some cable-stayed bridges in practice. However, some stay cables still vibrate after damper installation and the damage of damper devices are reported. It is regarded as due to that the dynamics of cable-damper-system is still not fully understood yet. On the other hand, the span of cable stayed bridges become longer and longer, passive dampers may not be able to suppress these super long stay-cable vibrations effectively. The research and practice for semi-active control is needed.To further investigate the dynamics of cable-damper-system, the mechanical behaviour of different dampers are tested. It is found that different dampers own different mechanical behaviour. The force velocity relation of different damper may be near to fractional model, linear or bilinear model. It is also found that nearly all dampers own stiffness.To testify the damping effect of different types of dampers, full scale experiments are conducted by using two 215 meters long stay cables with eleven sets of dampers attached to. It is confirmed that the internal damping of free cable is very low, smaller than 0.5% in logarithmic decrement for most of cases, however, the damping of cable attached with damper is much larger and greater than 0.03 in logarithmic decrement in average. It is also found that the damping observed is changed by amplitude and modes number, which is totally different to the results predicted from the linear design method, which assume damper is linear and thus damping of cable will be rely on cable frequency only.The comparison of the test results of the viscous fluid damper to the analysis is made. It is found that the tested changing trends of the damping to amplitude are similar to the analysis; however, the tested maximum damping is far less than the maximum damping predicted from analysis, which is due to the effect of stiffness. It is also observed that the tested damping is not well accord to the damping predicted from analysis when amplitude is small, and the observed effect of small friction force of the damper is confirmed to be the reason. The effect of this small friction force is also observed during tests of other kinds of dampers. It was found that due to the supporter stiffness, the tested average damping of oil damper was the smallest compare to the other three kinds of dampers. It is concluded that the importance of damper supporter should be emphasized. It is found that the damping of cable with attached voltage strengthened MR damper could be divided into two parts, which is confirmed to be due to the binlinear mechanical behaviour of MR damper. The damping of cable-MR damper system will rely on strengthening voltage, mode number and amplitude.The maximum damping predicted from analysis is observed in the test of friction damper. The damping vs. amplitude curve of the theoretical is similar to the test. However, the starting amplitude of the test results is smaller than the analysis; the reason is thoutht to be due to the non rational assumption.The free vibration of a taut cable with linear viscous damper and spring in parallel is investigated. An approximate analytical solution is derived for the optimum damping constant. It is found that the stiffness will greatly reduce the maximum attainable damping of the cable with attached damper near to anchorage. It is also found that there are also universal design curve for viscous damper with stiffness. The influence of fractional damper with stiffness is investigated and verified by full scale experiment results of nonlinear viscous damper.The free vibration of a taut cable with attached bilinear damper is investigated. It is found that the damping vs. amplitude curve of cable could be divided into two parts. The part one is at larger amplitude, the damper works in bilinear states. The part two is at smaller amplitude, the damper works in linear state. It is found that the critical changing point of damping will depend on initial state of cable vibration. The theoretical study is verified by full scale experiment results of voltage strengthened MR damper. It is found that the changing trend of damping is well predicted.Semi-active control method of cable vibration is proposed by using Shape Memory Alloy (SMA) as actuator. The SMA actuator is in series with cable near cable anchorage. The control rule is Kobori’s Switch/off control in separated time. A semi-active control experiment of model cable with SMA actuator is conducted. It is found that time delay of SMA actuator is only about 0.3s. This time is shorter than period of many stay-cable. The experiment found that the effect of control is not good for the free vibration of model cable vibration. The numerical simulation results also verified the experiment results. However, the numerical simulations find that the control effect will be much better when the control method is change SMA actuator according to phase of cable vibration, but this control method will need a much faster SMA actuator.更多还原