Study On The Mathematic Model And Artificial Neural Networks Model Of High Arch Dam Atomization By Jumping Jet

In recent year, the construction of high arch dam is accelerated. Since operation of the Ertan project by the end of last century, a number of high arch dam projects including Xiaowan, Goupitan, Xiluodu, Wudongde, etc. are carried out one after another. The impinging of two jets from surface and mid-level outlets used for energy dissipation brings serious flood discharging atomization in these projects. This paper uses the mathematical model to predict the range and rainfall intensity of atomization for high arch dams, and furthermore, two Artificial Neural Networks models of atomization by jumping jet are established. The followings are studied in this paper:(1) The orbit is evaluated according to a three-dimensional numerical model of the aerated water jet. Based on the study of spray characteristic of the jet when it plunges into the tailwater, a three-dimensional stochastic model including gravity, air resistance, air buoyancy and wind conditions is constructed to evaluate the range and the intensity of the rainfall in the gorges, which can be computed using Monte-Carlo analysis and Runge-Kutta method.(2) In the computation for the range of spray and intensity of rainfall, considering nappe-wind generated by flood discharge, the results of comparing the computational solutions with the prototype data observed from Ertan indicate strong affection of nappe-wind and a good adaptation for nappe-wind.(3) Using the mathematic model, the feedback and verifying analysis of atomization for Ertan hydropower station are carried out, and the results fit the prototype data well.(4) Using the mathematic model, the prediction of atomization for Wudongde and Goupitan hydropower station is carried out, and advice for mitigating the damages of atomization is proposed.(5) The comparison between the data of Wudongde hydropower station from the mathematic model and the collected data from prototype observations or model experiments shows that the computational solutions are correct and rational.(6) On the basis of the analysis on the factors effecting atomization, the Back-Propagation Neural Networks model of storm identification for the impinging energy dissipation is established and the storm ranges for Wudongde and Goupitan Hydropower Station are identified by the Back-Propagation Neural Networks Model of storm identification.(7) The Back-Propagation Neural Networks model of rainfall intensity distribution...