Developing NARX Neural Networks for Accurate Water Level Forecasting

• Main Authors: Basri H., Razak M.A., Sidek L.M.
• Other Authors: 57065823300
• Format: Book chapter
• Published: Springer Science and Business Media Deutschland GmbH 2024
• Subjects: Dam safety; Hydropower; NARX; Temengor reservoir; Water level

A reliable model for predicting fluctuations in water levels in the reservoir is essential for effective planning to manage the potential risks of flooding. A nonlinear autoregressive network with exogenous inputs (NARX) model is proposed to predict the water level of the Temengor Reservoir, Perak in Malaysia. The hyper-parameters of the proposed model have been optimized to enhance the accuracy of the proposed model while the Levenberg-Marquardt method was used to train the model. The NARX algorithm is capable of accurately predicting water levels with a high degree of accuracy. The use of the such technique for water level monitoring can be beneficial in the design of mitigation strategies for future flooding events, as it provides a critical parameter for gauging the potential severity of a flooding event. By understanding the changes in water level, emergency management teams can better prepare for and respond to floods, helping to minimize the damage and destruction they can cause. � The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2023.