The Feasibility Study Of Damage Detection On Concrete Dam Based On Vibration

In China,thousands of dams have been constructed in past years. Damage may occur in the period of service of each concrete dam. Dam will become dangerous even time bomb and endanger properties and people if these damages can't be found timely and develop arbitrarily. Therefore,damage analysis,safety assessment to damaged dam become one of important subjects to study for field of hydro-structure.Crack is one of the most common in variety of damages. Generally,crack appears in the local of dam early and then will make dam failure if there is no effective means and remedies to prevent it ceaseless spread. It is very significative to detect the crack as soon as possible. Displacement field,temperature field,stress field of dam will change only in very little area of it when someplace of dam appears damage. So it is difficult to detect damage in the early stage if no measurement point in the damage area by means of displacement measurement,temperature measurement or stress measurement. But the method based on vibration may overcome the difficulties. We notice that changes of structural parameter due to damage must change dynamic characteristic of structure. It may be possible to detect and identify structural damage by detecting the change of dynamic characteristic and detect structural damage on the whole even damage in area beyond of reach.Xianghongdian concrete gravity-arch dam was analyzed using numerical simulation in this paper. In different place of dam crack with different stripe and various depth was simulated and corresponding vibration frequency and shape were given. The frequency of damaged dam polluted by noise was also discussed and the results were presented using statistical neural network to identify. The effect of light change of water lever to dynamic characteristic was also considered. The study results prove that it is possible to detect damage of concrete dam using vibration test. The more deep the damage extent is,the more the frequency reduction is. Damage can be identified using statistical neural network when noise lever almost equals to percent of frequency reduction of dam due to damage. Of course,the method is limited by environmental factor and precision of instrument when used in practice.