نوع مقاله : مقاله پژوهشی
نویسندگان
1 کارشناس ارشد شرکت آب و فاضلاب مشهد
2 دانشیار گروه مهندسی آب، دانشگاه علوم کشاورزی و منابع طبیعی گرگان
3 دانشیار گروه مهندسی عمران، دانشکده مهندسی، دانشگاه بیرجند، بیرجند، ایران
4 دانشیارگروه مهندسی آب، دانشگاه علوم کشاورزی و منابع طبیعی گرگان
چکیده
کلیدواژهها
عنوان مقاله [English]
نویسندگان [English]
Groundwater has been discussed As one of the most important sources of drinking and agriculture water supply in arid and semi-arid especially. Birjand plain by being in the arid region and the use of ground water is as the only source of fresh water ahead. Predict the groundwater level fluctuations can be help in planning and future decision making, to provide long-term drinking water, agriculture and industry. The aim of this study was to compare two methods of MODFLOW numerical modeling and artificial neural network in forecasting of the Birjand aquifer groundwater table. ANN has been one of the smart ways, with the use of intrinsic, non-linear relationships between the thought and generalize the results to other modes. Numerical model receives all information about aquifer GIS layers, which is able to predict the groundwater level in the future. GMS software uses for the numerical solution of motion equation and of the finite element and finite-difference methods. In this study, the finite difference method was used. Numerical models run for steady state, unsteady state and for three scenarios wet, normal and dry were compared. The neural network model inputs were taken from the extraction wells, the amount of input water to each polygon in terms of cubic meters (caused by area rainfall) and the water level in the piezometers, in the step before, and the model, the water level is at the current time step. The results show that using artificial neural network can be predicted with reasonable accuracy the level of underground water for up to 12 months later. Moreover, in a number of piezometers predict the level of groundwater is of sufficient accuracy to 18 months. MODFLOW numerical model predicts with more accuracy than the artificial neural network changes the water level within 24 months.
کلیدواژهها [English]