Author(s): M. R. Kavianpour; M. R. Najafi; Z. Kavianpour
Linked Author(s):
Keywords: Genetic Algorithm; Subtractive Clustering; ANFIS; Bottom Outlet; Aeration
Abstract: In recent years, the construction of high dams with high velocity flows and discharges over their spillways or/and within their bottom outlet conduits has had a rapid progress. With increasing the flow velocity over the surface of these structures, the potential for cavitation damage is also increased. Amongst various methods, aeration has been found as a cheap and easy way to reduce the cavitation tendency. Based on various studies, it has also been found that flow aeration eliminate cavitation risk downstream of gates in outlet tunnels. So far, many works have been done and various relationships have been introduced to predict the quantity of entrained air downstream of outlet gates, but still there are uncertainties in using those expressions for various conditions. On the other way, in recent years the applications of artificial intelligence, such as neural networks, fuzzy logic, and genetic algorithm have attracted the attention of many investigators. They are known as powerful tools to solve engineering problems with uncertainties. Therefore, in this study, a model based on adaptive network based fuzzy inference system has been used to estimate the quantity of air demand downstream of gates in outlet conduits. Fuzzy subtractive-clustering is used to determine the principle structure of the model and, to train it ANFIS is used based on experimental information from various physical model studies. The measurements were made at Water Research Institute of Iran on scaled physical models (1: 12-1: 20) of bottom outlet conduits of Jareh, Alborz, Dasht-e-Abbas, Jegin, Karkheh, Kosar, Taham and Seymareh dams, which are under construction in Iran. The model shows a very good prediction of the quantity of required air, compared to the experimental information from the physical models.
Year: 2007