Raqamli Transformatsiya va Sun’iy Intellekt
OPTIMIZATION OF HYDRAULIC PROCESSES IN OPEN CHANNELS BASED ON DEEP LEARNING MODELS
1 University of World Economy and Diplomacy
2 Research Institute of Irrigation and Water Problems
3 National University of Uzbekistan named after Mirzo Ulugbek
Zenodo DOI: 10.5281/zenodo.21912269
Keywords
Abstract
Under conditions of water resource scarcity and climate change, the issue of efficient water management in irrigation systems is becoming increasingly important. Accurate prediction of hydraulic processes in open channels is complex due to the nonlinear behavior of these systems and hydraulic inertia. In this study, a data-driven approach based on a Long Short-Term Memory (LSTM) neural network is proposed for predicting water discharge in open channels. The Mirishkor irrigation canal in the Republic of Uzbekistan was selected as the research object. The model was trained using time-series data of water level, gate opening degree, and inflow discharge. A dataset reflecting the channel dynamics was generated by considering time delay and inertia. According to the experimental results, the linear regression model failed to represent the system dynamics (R² = 0.0039), whereas the LSTM model demonstrated high accuracy (R² = 0.7963). The RMSE and MAE indicators decreased by 54% and 55%, respectively. The results show that sequence-based neural networks are effective in modeling open channel systems. The obtained findings confirm that deep learning models are an effective tool for developing optimal control and decision-making systems in open channels.
References
- Mamataliev A. B., “Optimization of water use in the Mirishkor canal system using digital technologies.” Irrigation and Melioration, 4(12), 15–21. (2022).
- Murtazaev A., Rakhimov S. “Automation of water metering in main canals: Problems and solutions.” Central Asian Journal of Water Resources, 9(1) (2023)., 45-58.
- Zhang J., “Water level prediction based on LSTM and attention mechanism: A case study of the Yangtze River.” IEEE Access, 8, 2020, 839–848b.
- Hochreiter S., & Schmidhuber J. “Long Short-Term Memory.” Neural Computation. 1997.
- Seytov A., Аbdurаxmonov O., Choriyorov N., Maxkamov А. Development of аlgorithms for plаnning аnd mаnаgement of wаter resources in Irrigаtion systems in the conditions of climаte chаnging, SCIENTIFIC BULLETIN Physical and Mathematical Research 2024/№1(6).
- Wang Y., “Daily runoff forecasting using a deep learning model based on LSTM networks.” Journal of Hydrology, 594, (2021).
- LeCun Y., Bengio Y., Hinton, G. “Deep learning.” Nature, 521. (2015), 436–444. (2015). (Chuqur o‘rganish (Deep Learning) metodologiyasi bo‘yicha fundamental qo‘llanma).
- Kamyab H., “The latest innovative avenues for the utilization of AI and big data analytics in water resource management.” Results in Engineering 2023.
- Turaev R., Seytov A., Haydarova R., Abduraxmonov O., Choriyorov N. “Optimal water management in the channels of machine water raise systems”, O‘zMU xabarlari 21.1, 2024.