Abstract:

Besides VietNam, there are some countries such as Laos, Cambodia…that care of developing the hydropower works. However, forecasting about the displacement and deformation of the hydropower dam has not focused yet, therefore warning of the dam’s status has not updated frequently so it is easy to cause catastrophe in case of the broken-down dam. Now deformation and displacement monitoring of hydropower works is divided in two types: surface monitoring and dam core monitoring, surface monitoring is mainly observed by geodetic methods and periodical measurement because it is expensive and complex to estalish a continuous monitoring system. Dam displacement forecasting on the basis of the cyclic data is extremely important, it help warning the risks of dam and reducing damages on material and human. The article researched the application of artificial neural network on subsidence modeling of the dam surface monitoring points on the base of data that was observed in previous periods, then predict the subsidence of the next ones. Results of the experiment on building subsidence model of Yaly hydropower dam demonstrated that the artificial neural network model has good accuracy, forecasting values are the same as measures and then they proved the applicability of this model in predicting the subsidence of the hydropower dam.