Application of Gene Expression Programming Approach to Estimate the Aeration Coefficient of Bottom Outlet Gates of Dams

Abstract
Background and objectives: The use of storage dams plays a key role in the development of industry, agriculture and employment communities Bottom outlet tunnels are one of the most significant components of the reservoir dams which are used in flood evacuation and control. They consist of inlet duct, main conveyance tunnel and flow regulator structures including gates and valves. A major problem with bottom outlet gate of dams is cavitation which happens in the high flow discharge. This phenomenon would destroy the surface of structure. It has been demonstrated that flow aeration is an effective way to reduce the cavitation damages. In this regard, the flow aeration rate is an important discussion that must be noted. Since, in this paper aeration coefficient evaluation is assessed.
Materials and methods: This study, is to estimate the aeration coefficient of bottom outlet gate of four dams (Alborz, Zhaveh, Gotvand Olia, Jareh) using Gene Expression Programming (GEP) approach. To achieve this aim, experimental data were used collecting from hydraulic structures laboratory of Tehran Water Research Institute to train and test the model. The aeration coefficient was influenced by compressed Froude number (Frc) and aerator area to gate area ratio (Aa/Ag). 30 chromosomes and 3 genes were chosen to GEP performance. The model ability was assessed by two statistical parameters of correlation coefficient (R2) and root of mean square error (RMSE).
Results: The results show that GEP predicted the aeration coefficient of bottom outlet gates of dams with R2 of 0.803 and 0.639 and RMSE of 0.096 and 0.125 for training and testing stages, respectively. This model gave better results compared by regression equation with R2 of 0.718 and 0.402 and RMSE of 0.114 and 0.171 for training and testing parts, respectively. In the other words, the error of aeration coefficient prediction was decreased about 28% using GEP approach.
Conclusion: The results show that GEP intelligence approach is an adequate model to predict aeration coefficient of bottom outlet gates of dams. Also, the results of traditional regression equations were improved using this method. In the other words, these results indicated that GEP is reliable to evaluate the aeration coefficient of bottom outlet gates of dams by more accurate estimation to prevent cavitation phenomenon. So, use of this way is suggested in future studies related to this topic.


Conclusion: The results show that GEP intelligence approach is an adequate model to predict aeration coefficient of bottom outlet gates of dams. Also, the results of traditional regression equations were improved using this method. In the other words, these results indicated that GEP is reliable to evaluate the aeration coefficient of bottom outlet gates of dams by more accurate estimation to prevent cavitation phenomenon. So, use of this way is suggested in future studies related to this topic.