Research On Monitoring Model Of Seepage Flow Around Hydraulic Structure

With the development of water resources exploitation and utilization, the problems involved in hydraulic structure safety become increasingly remarkable. It is important to predict the monitoring data of hydraulic structure precisely for supervising and decision-making of hydraulic structure operation.Aiming at seepage flow monitoring problems related to safety hydraulic structure closely, some theories and methods are presented for seepage flow monitoring model based on the mathematical and mechanical knowledge, such as probability statistics, methods of optimization, wavelet network, and so on, together with prototype observation date from Tianhuangping pumped-storage hydroelectric plant, the ordinary statistical model, the optimum statistical model on lagging effect and the wavelet network model are established respectively, we can draw the following conclusions from obtaining solution by the way of the numerical analysis method.1. After analyzing factors of seepage flow quantity, the ordinary statistical model is established by selecting factors reasonably. The example shows that ordinary statistical model is effective in analyzing monitoring data of linear seepage flow, however, when it is less effective in analyzing monitoring data of non-linear seepage flow.2.Based on analysis of lagging effect of reservoir level and rainfall on the seepage flow, the seepage flow monitoring models of the dam are researched, and the seepage flow optimum statistical model on lagging effect is established. The applied result shows that optimum statistical model is more precise than the ordinary statistical model in analyzing homology monitoring data, but the predicting of model is not very accurate in analyzing monitoring data of non-linear seepage flow.3.Based on the research of the essential characteristic of the wavelet and the wavelet transform, the wavelet network is introduced, the wavelet network model and its structure and learning rules are particularly researched, taking the advantage of wavelet network on the problem of approximation of function, seepage flow waveletnetwork model is established. The result shows that, comparing with ordinary statistical model and optimum statistical model, it take better forecast precision in the forecast of linear and non-linear seepage flow of the dam.