混凝土重力坝动力系统的可靠性方法

为满足巨大的能源需求,一批混凝土重力坝项目将要在我国多震的西南地区建设开工。这些工程库容达数十亿到数百亿立方米,并且大部分处于强烈地震活动区,若发生破坏将带来严重的后果。强震对混凝土重力坝的设计提出了更高的要求。因此,有必要对混凝土重力坝的动力可靠性进行系统性的分析。本文以此为目的,从合成人工地震波等荷载和设置人工边界条件入手,结合数值计算,对混凝土重力坝的弹塑性损伤过程进行了动力分析,探讨了混凝土重力坝的破坏模式及其可靠度的计算方法,并深入的研究了混凝土重力坝溃坝的风险评价方法。为工程设计人员有针对性地采取工程措施进行重力坝抗震设计提供参考。主要研究内容如下：(1)基于遗传算法和小波理论分别改进了单点人工地震波和空间相关人工地震波的合成方法。采用小波函数优化了初始的人工地震波,并采用遗传算法分别对单点人工地震波和空间相关人工地震波合成的迭代过程进行了优化,使优化后生成的人工地震波的计算反应谱与目标反应谱的误差更小(2)结合透射边界和黏弹性人工边界理论提出了一种改进的黏弹边界,从而考虑了结构—地基交界面侧向位移对于结构的影响；基于阻尼溶剂抽取法理论又提出了另一种改进的人工边界,使阻尼溶剂抽取法易于运用在有限元数值计算中。(3)采用无限单元模拟了无限地基对混凝土重力坝的影响,在此基础上采用弹塑性损伤模型模拟了混凝土重力坝的损伤破坏过程。并对混凝土重力坝的损伤分布和动力响应进行了分析比较。利用能够反映混凝土拉压方向不同强度特性的Hsieh-Ting-Chen四参数屈服函数,基于热力学耗散原理提出了一种混凝土损伤本构模型,其拉、压损伤演化定律采用相同的计算公式,能够较好的反映混凝土在复杂应力状态下的拉压不同的损伤特性及应力应变本构关系。(4)基于摄动理论和虚拟激励法建立了对混凝土重力坝损伤分布的概率分析模型。通过计算得到了混凝土重力坝损伤因子期望值和方差的分布云图,并得到了其损伤概率的分布云图,同时检验本模型的收敛性。(5)基于虚拟激励法,计算了失效路径上的单元的应力的概率特征。运用虚拟变量法,计算了单元的条件失效概率。引入了马尔科夫过程的相关理论计算了失效路径的失效概率。采用串联模型可靠度计算方法对坝体的体系可靠度进行了计算。(6)基于奇异值分解及加权回归方法对响应面法进行了改进,并建立了新的模型用来求解验算点。通过以上方法使其抽样点的个数与迭代的次数都明显降低,从而,将其可以应用在坝体等大型结构的可靠度的分析中。(7)针对岩基深层滑动面的随机性和模糊性的特点,运用概率推理理论对岩基深层抗滑稳定的模糊可靠度进行了分析。(8)通过坝头,上游折坡面,建基面三处的抗滑失效概率推导出了溃口的宽度,在此基础上计算出溃口洪水流量随时间的变化规律,从而计算得出了溃坝后的生命以及财产损失。并基于系统动力学,建立了灰色系统动力学评价方法用于评估溃坝后的损失。又利用矩阵特征值分解理论和模糊理论建立了对溃坝后损失的评价方法。并对毕节地区运用本模型进行了数值模拟,检验了本模型的准确性。
工业羊毛毡
To meet the enormous energy demands, several concrete gravity dams are being constructed in earthquake regions of Southwestern China. The reservoir capacities of these projects reach several billion cubic meters, and most of these projects build in strong earthquake regions. Once the dams are failure, it will bring very serious results. A strong earthquake is therefore a great challenge to dynamic reliability design for concrete gravity dams. Therefore it is necessary to systematic analyze dynamic reliability for concrete gravity dams. For this purpose, by means of synthesizing artificial seismic wave and other loads and setting for artificial boundary condition, finite element method(FEM) based numerical calculations were adopted to make dynamic analysis for the elastic-plastic damage process, failure mode and dynamic reliability of conerete gravity dam. Moreover, it made deeply study on evaluation methods for dam-break risk. The results offer important reference to engineering designers and.anti-seismic design. The major contributions are summarized as follow:(1) Single-point and spatial correlation artificial seismic wave synthesis methods were improved based on genetic algorithm and wavelet theory respectively. Initial single-point artificial seismic wave synthesis was optimized by wavelet theory. The iteration processes were optimized by genetic algorithm for single-point and spatial correlation artificial seismic wave synthesis. The result showed that the deviation between calculational and objective response spectrum was smaller by this paper’s method than conventional method.(2)An improved viscous-spring artificial boundary condition was proposed based on conventional transmitting and viscous-spring artificial boundary condition in order to eliminate influence of the lateral displacement on structure-foundation interface. Another artificial boundary condition was proposed based on damping-solvent extraction method in order to make damping-solvent extraction method easily used in finite element method(FEM) based numerical calculations.(3) Infinite element method (IEM) was adopted to simulate the influence of infinite foundation to concrete gravity dam, and then elasto-plastic damage constitutive relations model was adopted to simulate the damage process of dam. Moreover, damage distribution and dynamic response of concrete gravity dam were analyzed. By using Hsieh-Ting-Chen four-parameter yield function that could reflect the different strength characteristics of concrete in tension and compression, a damage constitutive model for concrete was proposed based on the energy dissipation principle. It had the same damage evolution law under tension and compression load and it was good for determining stress-strain constitutive and damage characteristic in complex stress state. (4) A new method was proposed based on perturbation theory and pseudo excitation method in order to make the probability analysis for the damage distribution of concrete gravity dam. The expected value, variance and probability of damage of concrete gravity dam were obtained by numerical calculation, and then astringency of this modal was examined.(5)Based on pseudo-excitation method, the probability characteristics of element stress in failure path were calculated. The conditional failure probability of elements in failure path was obtained by method of dummy variable. Through leading into markoff process theory, failure probability of failure path was obtained. And then calculation method for reliability of series model was adopted to calculate systematic reliability of concrete gravity dam.(6) Singular value decomposition(SVD) and weighted regression methods were adopted to improve response surface method. And a new modal was established to calculate checking point. The number of sample points and iteration times was decreased significantly by using the method above, so this method could be applied to analyze the reliability of large structures such as concrete gravity dam and so on.(7) Aiming at random and fuzzy characteristic of deep sliding surface, probability reasoning theory was adopted to analyze the fuzzy reliability for deep sliding stability of dam foundation.(8) The width of breach was deduced through the anti-slide failure probability of dam head, upstream broken-line sloping surface and foundation surface, and then relationship between volume of dam-break floods and time was calculated. Moreover, the losses of lives and property were obtained. A risk evaluation model was established based on system dynamics and grey theory to evaluate the dam-break losses. By decomposing the related indexes and confirming membership degree with the optimized method, another risk evaluation model was proposed based on fuzzy theory. And then this modal was applied to bijie region to to examine the correctness.