The Stochastic Finite Element Method And The Structure Reliability Analysis In The Dam

In this thesis, the available methods for estimation of the safety of a concrete dam are first reviewed. Then a new method which combines stochastic finite elements with structural reliability analysis is presented for analysis of reliability of concrete dams. And a systematic study for stochastic finite element methods and structural reliability analysis methods are also performed.The research work in this thesis can be summarized as follow:1. A spectral approach based on Karhunen-Loeve expansion is used for discretization of the random field. It has been proved to have advantages of rapid convergence and high accuracy.2. Comparisons are made for the advantages and disadvantages among the reliability method, perturbation method, spectral approach and Monte Carlo Simulation(MCS). A improved Neumann expansion method is proposed and the corresponding equation is derived. The method is then used in nonlinear analysis. A better convergence and good accuracy has been reached by using some special techniques.3. The maximum entropy method is presented to calculate the reliability of concrete dams. It is shown that this method is more effective than the other methods, such as first order secondary moment and MCS. In this method, the results do not rely on the tail of probability distribution function of the random variables. Thus the method is especially suitable to the cases that the distribution functions are not exactly known.4. The crack band theory is used to derive the constitutive relationship of concrete after cracking. And the limit state functions for tension and compression strength and stability are constructed and used in analysis of the reliability of concrete dams. A true example of concrete dam is studied by using the above methods.5. A computer program(named SSDAP) is coded in FORTRAN. It is compiled by using Lahey FORTRAN system. The code obtained is more efficiency than that got by other compiler. The program can deal with static(linear or nonlinear), dynamic as well as stochastic problems.