Modernizing Flood Control with Intelligent Dam Gate Prediction – A Survey

Generally, dams are built across the rivers to stop the river flow into the seas and oceans. In this way, dams will store water in it. This stored water will be utilized for many purposes, like electricity generation, cropping, etc. Apart from this, one of the major reasons for building the dam is to store water during heavy rain. Nowadays, this rainfall is unpredictable because of the large amount of uncertain rainfall. Dams are getting damaged, and it is costing the lives of the people who are living in lowland areas. This work explores the machine learning model, which can predict the dam gate opening and closing and act as an alarm for people living in lowland areas. In this work, classification models like Naïve Bayes, K means, KNN, SVM, DT Classifier, RF Classifier, GB Classifier, ANN, LSTM, and regression models like KNN Regressor, SVR, DT Regressor, RF Regressor, and GB Regressor were tested and trained. The model was tested and trained with the daily precipitation and dam gate data. The evaluation metrics for classifier models are F1 score, accuracy, and precision. MSE, MAE, RMSE, and R-squared are the evaluation metrics used for regressor models. Finally, after training the different classifier and regression models, it was concluded that Naïve Bayes, K-means, Random Forest classifier, and ANN classifier are the best-suited models for this work. The main objective of this model is to predict the dam gate position using precipitation data and previous dam operation data.