نوع مقاله : مقاله علمی پژوهشی
نویسنده
کارشناس ارشد جغرافیا و برنامه ریزی شهری دانشگاه شهید چمران اهواز، اهواز، ایران
کلیدواژهها
موضوعات
عنوان مقاله English
نویسنده English
Abstract
Introduction and statement of the problem: The use of neural intelligence in predicting the variables of water resources, including underground water, is widely increasing. Purpose: This research through artificial intelligence and FNN-LM model pursues several goals, which include determining the parameters affecting the fluctuations of the underground water level in the Khuzestan plain, as well as investigating the spatial and temporal effects of the water level parameters through 10-year time data and Then, the modeling of groundwater level fluctuations in selected piezometers in the plain is studied. Method: The use of artificial intelligence and the FNN-LM model method, and at the end, by changing the percentage of the last month of the input data in the model, hypothetical conditions were created and according to the obtained neural network models, the fluctuations of the underground water level were predicted. In this hypothetical situation, it was discussed. Findings: The effect of the discharge parameter from the wells is far more than the effect of the rainfall parameter, so that the prediction of drought and drought conditions is only due to the change of the rainfall. Conclusion: By using the created neural network models for each observation well and using the most accepted method of geostatistical models, an appropriate spatial and temporal prediction of the groundwater level was made. The best modeling of water level fluctuations with the FNN-LM model was achieved by choosing appropriate parameters and with the most acceptable time delay.
Keywords: artificial intelligence, underground water level, FNN-LM model, Khuzestan plain.
کلیدواژهها English
