The College of Engineering at Kirkuk University discussed a master's thesis entitled:
(Prediction of Seepage in Earth Dams Using Artificial Intelligence, Case Study: Duhok and Darbandikhan Dams). The thesis, presented by student Derin Raouf Saber, aimed to develop models for predicting water seepage in earth dams using artificial intelligence techniques. These techniques included artificial neural networks with backpropagation and scaled conjugate algorithms, long-term and short-term memory neural networks (LSTM), adaptive fuzzy neural inference systems (ANFIS) with hybrid and backpropagation training algorithms, and support vector regression (SVR).
The message concluded that the artificial neural networks performed exceptionally well using the Scaled Conjugate algorithm, as the value of (R²) reached (1) in most piezometers, while the value of (MAE) recorded an amount of (0.005) in piezometer (P1), thus surpassing the rest of the models in terms of matching accuracy. The backpropagation algorithm achieved near-perfect performance, and the (LSTM) models showed high stability and efficiency in leakage prediction.
