基于混沌理论的高速水流和流激振动特性研究

发布时间：2018-05-17 10:56

  本文选题：高速水流 + 流激振动　； 参考：《武汉大学》2014年博士论文

【摘要】：高速水流和流激振动现象是水利工程中十分常见并且非常重要的问题,在建设高水头、大流量水利工程时,如果在设计、施工、运行管理等环节上稍有不慎,就有可能因这两个不利现象而威胁到整个枢纽的安全,因此对高速水流和流激振动现象进行深入研究就显得十分必要。国内外关于这两方面的研究也已经积累了丰厚的成果,但是绝大部分都是在工程意义上的,这意味着无论在试验数据分析、还是数值模拟方面,基本都是以雷诺统计平均思想为理论基础的,这种处理方法虽然在一定程度上能够满足工程上的需求,但它忽视了高速水流或者流激振动现象中所蕴含的某种复杂规律性。因此将混沌理论引入到高速水流和流激振动的研究中来,是一种在工程意义下,将与湍流和振动有关的研究回归到湍流以及流固耦合现象本源性质的全新思想,并能为工程设计人员更好地理解高速水流和流激振动现象中的复杂性提供一定的理论基础,具有普遍的意义。论文首先研究了边界条件对闸门振动特性的影响规律,随后基于混沌理论,以模型试验中所测得的数据为研究背景,采用混沌初步识别方法、相空间重构理论、混沌特征量的对比分析等方法,初步研究了窄缝式消能工、阶梯式消能工、消力池底流消能以及平板闸门流激振动四种特定情况下高速水流脉动压力或振动加速度响应中所蕴含的复杂性规律,并揭示了高速水流和流激振动中存在的混沌特性。以龙开口水电站深孔平板闸门为对象,通过Block Lanczos法进行模态分析,采用约束刚度连续变化及附加质量法研究了边界支承条件和流固耦合两方面因素对其自振特性的影响,发现顺流向、侧向及竖直向约束刚度的变化在某一范围内对闸门自振特性的影响非常显著,其规律与矩形薄板横向振动的规律一致；自振频率与其振型振动方向所对应的约束条件紧密相关,与其他方向的约束条件基本无关,通过分析闸门自振频率随约束变化的规律,可以推求其相应振型的振动方向；在考虑流固耦合的情况下,门前水体通过附加质量进行模拟,结果表明水位在一定范围内的变化对闸门自振频率的影响可以忽略,且约束条件的影响相对较小。随后,在对四种不同模型试验中测量得到的试验数据进行混沌特性分析时,普遍得到了以下三条结论：(1)整体来看,对实测数据时间序列进行混沌特征的初始判别时,采用主分量分析法和0-1测试法是完全可行的,而且概念简单、可操作性强、判别方法直观有效。但这两种方法只能定性判别时间序列是否具备混沌特性,而不能定量表示和区分混沌程度的强弱差异等。(2)相空间重构的嵌入维数采用平均伪最近邻域法(Cao方法)进行计算,研究发现实测数据中包含一定水平的噪声,可能对吸引子的重现和Lyapunov指数的计算造成影响,不过文献[147]指出对于高维的时间演化过程,Cao方法对噪声具有更好的鲁棒性,该法在实测数据混沌分析的应用中是完全可行的,后来在关联维数的计算中也证明了这一点。(3)噪声对Kolmogorov熵的计算有较明显的影响,使得K只能对实测时间序列进行定性的混沌特性识别,而不能反映混沌程度的大小；而最大Lyapunov指数在脉压序列中也没有很明显的分布规律,但在加速度信号的分析中则存在明显规律。在相空间重构及混沌特征量的分布规律两个方面,不同模型试验存在一定的差异,主要结论如下：(1)窄缝消能工选取一级收缩方案(FC)和二级收缩方案(SC)两种体型进行对比分析,得到相空间重构嵌入参数τ在6-13之间,m在11-16之间,在FC方案中,τ口m随流向都存在一定的规律性。饱和关联维数D2更能反映出一定的规律性,表明在相同体型条件下,上游水位越高,流量越大,相应的流动结构越复杂,紊动随机性更高；而在SC体型的第二个收缩段底板附近,水流紊动程度并不取决于水位、流量等初始条件,而取决于二级收缩段边壁的突然转折；窄缝消能工中边墙的突然转折使得底板附近水流结构的复杂性及紊动程度比边墙附近水流更加显著,但收缩段并没有从本质上改变水流的动力结构,而只是微小的扰动。(2)阶梯式消能工选取两组阶梯组合,并通过单宽流量和弗劳德数来控制来流条件,研究表明：相空间重构的最佳嵌入参数范围为τ-=7-18,m=8-17,结合关联维数随来流条件的变化规律,发现阶梯式消能工上水流内部结构的演化过程并不直接取决于来流条件或者阶梯体型,而是取决于阶梯上的水流流态,流态不同,水流状态空间的演化规律就不同。竖直面凸角处的脉压序列存在一定水平的噪声或其附近流场的复杂度较高,可能由于跌落水流在竖直面凸角处存在较大空腔,空腔的脉动比水流脉动更复杂。混沌特征量方面,λ1的分布规律表明竖直面凸角测点的脉压序列在跌落水流时比滑行水流更显出随机性,而水平面测点的脉压序列在滑行水流时比跌落水流的混沌程度更大。D2的分布规律表明在滑行水流时,竖直面和水平面凸角处的测点均有沿程下降的趋势。(3)消力池中选取一种新型消能结构为试验方案,拟定三组试验工况。结果表明,消力池底板测点的植整体上略大于消力中墩测点,而m值则相反。D2总体分布在5.238-8.854之间,消力池底板测点数据的D2较消力中墩要小；在大流量工况(2%)时,消力池底板测点D2值有随流向增大的趋势；小流量工况(50%)时,D2值呈现出随流向减小的趋势；中墩测点的D2值大小与来流条件无关；随流量的减小,中墩前底板测点的D2值有先减后增的趋势,中墩后测点则是递减趋势,这与该优化方案的消能机理有关。在对λ1分布规律的分析中也间接表明了该优化方案在较小流量时的消能效果更好。(4)研究了水弹性模型试验中所测量的闸门加速度响应数据,拟定了1/8～7/8共7个开度条件,上游水位控制在设计水位。除了1/8开度外,侧向振动的m值比其他方向更大,而1/8和7/8开度在各振动方向时的m值比其他中间开度都要小。在该闸门的竖向振动与顺流向振动中呈现出了较低维(3.342～5.130)的混沌吸引子,表明对闸门流激振动系统进行建模,只需要更少的独立控制变量,就可以基本描述闸门在振动过程中所呈现出来的复杂性和非线性规律。不同测点的λ1随闸门开度的变化规律呈现“两边小中间大”的趋势,表明除了1/8和7/8开度,在其他局部开启条件下,闸门在水流激振影响下的振动复杂性更大,这就体现在实际工程中,闸门在2/8开度到6/8开度之间的振动情况存在更多不确定性,包括强烈振动的情况。
[Abstract]:High speed water flow and flow excited vibration are very common and very important problems in water conservancy projects. In the construction of high water head and large flow water conservancy projects, if they are inadvertent in the design, construction, operation management and other links, it is possible to threaten the safety of the whole hub because of these two unfavorable phenomena. Therefore, the high speed flow and flow exciting vibration can be caused. It is very necessary to study the dynamic phenomena in depth. The research on these two aspects has also accumulated rich achievements, but most of them are in the engineering significance. This means that both the analysis of experimental data and the numerical simulation are basically based on the theoretical basis of the Reynolds statistical mean thought. Although the method can meet the needs of engineering to some extent, it ignores some kind of complex regularity contained in the phenomenon of high speed flow or flow induced vibration. Therefore, the introduction of chaos theory into the study of high speed water flow and flow excited vibration is a kind of research in engineering, which will return to the study of turbulence and vibration. The new idea of turbulence and the intrinsic properties of fluid solid coupling can provide a certain theoretical basis for engineering designers to better understand the complexity of high speed flow and flow excited vibration. The paper first studies the influence of boundary conditions on the vibration specificity of the gate, and then based on the chaos theory, the model is based on the theory of chaos. The data measured in the type test is the research background, the initial chaotic identification method, the phase space reconstruction theory and the chaotic characteristic comparison analysis are used to study the pulsating pressure or vibration of the high speed water flow under the four specific conditions of narrow slit energy dissipator, staircase type energy dissipator, stilling pool bottom current dissipation and flat gate flow induced vibration. The complex law contained in the acceleration response is presented, and the chaotic characteristics of the high speed water flow and the flow excited vibration are revealed. The modal analysis is carried out by the Block Lanczos method, taking the deep hole flat gate of the long shedding hydropower station as the object, and the two aspects of the boundary support conditions and the fluid solid coupling are studied by the continuous change of the constrained stiffness and the additional mass method. The influence of the factors on the natural vibration characteristics shows that the change of the lateral and vertical restraint stiffness has a very significant influence on the self vibration characteristics of the gate in a certain range. The law is consistent with the law of the transverse vibration of the rectangular plate, and the frequency of the vibration is closely related to the constraint conditions corresponding to the vibration direction of the vibration mode, and it is about the other direction. The beam condition is basically irrelevant. By analyzing the law of the vibration frequency of the gate, the direction of its corresponding vibration can be calculated. In the case of fluid solid coupling, the water body in front of the gate is simulated by the additional mass. The result shows that the influence of the change of the water level in a certain range on the vibration frequency of the gate can be ignored, and the restraint bar can be ignored. The influence of the parts is relatively small. Then, in the analysis of the chaotic characteristics of the experimental data obtained from four different model tests, the following three conclusions are generally obtained: (1) in the whole, the principal component analysis and the 0-1 test method are completely feasible for the initial discrimination of the chaotic characteristics of the measured data time series. The two methods can only determine whether the time series has chaotic characteristics, but can not quantify and distinguish the difference of the intensity of chaos. (2) the embedding dimension of the phase space reconstruction is calculated by the average pseudo nearest neighborhood method (Cao method), and the actual number is found. The noise in a certain level may affect the recurrence of the attractor and the calculation of the Lyapunov exponent. However, literature [147] points out that the Cao method has better robustness to the noise in the time evolution process of the high dimension. The method is completely feasible in the application of the measured data chaos analysis, and later in the calculation of the correlation dimension. It is also proved that (3) noise has a significant influence on the calculation of Kolmogorov entropy, so that K can only identify the qualitative chaotic characteristics of the measured time series, but can not reflect the size of chaos, but the maximum Lyapunov exponent has no obvious distribution in the pulse pressure sequence, but it is stored in the acceleration signal analysis. There are certain differences in the two aspects of the phase space reconstruction and the distribution law of the chaotic characteristic quantity. The main conclusions are as follows: (1) the narrow slit energy dissipator selects the first order contraction scheme (FC) and the two stage contraction scheme (SC) for the analysis of the ratio, obtains the phase space reconstruction embedding parameter tau between 6-13, and m in 11- 16, in the FC scheme, the m of the tau port has a certain regularity with the flow direction. The saturation correlation dimension D2 can reflect a certain regularity. The higher the water level, the greater the flow, the more complex the flow structure and the higher turbulence, and the turbulence in the second contraction segments of the body shape. The degree does not depend on the initial conditions such as water level, flow and other initial conditions, but it depends on the abrupt turning of the side wall of the two stage contraction section. The sudden turning of the side wall of the narrow gap energy dissipator makes the complexity and turbulence degree of the flow structure near the floor more significant than the flow near the side wall, but the contraction section does not have the dynamic structure that essentially changes the flow of the flow, but only the dynamic structure of the flow is not essentially changed. It is a small disturbance. (2) the ladder type energy dissipator selects two sets of ladder combinations and controls the flow condition through the single wide flow and Froude number. The study shows that the optimum embedding parameter range of phase space reconstruction is tau -=7-18, m=8-17, combining the relation dimension with the changing law of the flow condition, it is found that the internal structure of the water flow in the staircase energy dissipator is shown. The process does not depend directly on the flow condition or the staircase shape, but depends on the flow pattern on the ladder, the flow state is different, the evolution law of the flow state space is different. The pulse pressure sequence at the convex angle of the vertical face has a certain level of noise or the complexity of the flow field near the vertical plane, which may be due to the falling flow at the convex angle of the vertical face. There is a larger cavity, the pulsation of the cavity is more complex than the flow pulsation. In the aspect of chaotic characteristic, the distribution of lambda 1 shows that the pulse pressure sequence of the vertical plane convex angle measuring point is more random than the gliding flow in the fall flow, and the distribution law of the pulse pressure sequence of the horizontal point measuring point is larger than the falling water flow in the gliding flow. It is shown that in the sliding flow, the measuring points at the vertical and horizontal angles all have a downward trend along the distance. (3) a new energy dissipation structure is selected as the test scheme in the stilling pool, and the three sets of test conditions are drawn up. The results show that the planting of the bottom plate of the floor of the stilling pool is slightly larger than the stilling piers, while the m value is generally distributed in the 5.238-8.85 4, the D2 of the bottom plate of the stilling pool is small, and the D2 value of the bottom plate of the stilling pool has a tendency to increase with the flow direction in the large flow condition (2%); when the small flow condition (50%), the value of the D2 value decreases with the flow direction; the size of the piers measuring point is independent of the flow condition; as the flow rate decreases, the front bottom plate measurement of the middle pier is measured. The D2 value of the point has a tendency to decrease and increase first, and the post test point of the middle pier is the decreasing trend, which is related to the energy dissipation mechanism of the optimization scheme. In the analysis of the distribution law of the lambda 1, the energy dissipation effect of the optimized scheme is better. (4) the acceleration response data measured in the hydroelastic model test are studied. A total of 7 opening conditions of 1/8 to 7/8 are determined. The upstream water level is controlled at the design water level. In addition to the 1/8 opening, the m value of the lateral vibration is larger than that in the other directions, while the m value of 1/8 and 7/8 opening is smaller than the other intermediate opening degrees in each direction of vibration. The lower dimension (3.342 to 5.130) of the chaos is presented in the vertical and downstream vibration of the gate. The attractor indicates that only less independent control variables are needed to model the gate flow excited vibration system. The complexity and the nonlinear law of the gate in the vibration process can be described basically. The trend of the change of the gate opening of the different measuring points with the opening of the gate is "small in the middle of the two sides", which shows the 1/8 and 7/8 opening. Under the conditions of other local opening, the vibration of the gate is more complex under the influence of the flow excitation, which is reflected in the actual engineering. There are more uncertainties in the vibration of the gate between the 2/8 opening and the 6/8 opening, including the strong vibration.
【学位授予单位】：武汉大学
【学位级别】：博士
【学位授予年份】：2014
【分类号】：TV131.3
，

本文编号：1901095