Sediment load forecasting of Gobindsagar reservoir using machine learning techniques

With ever advancing computer technology in machine learning, sediment load prediction inside the reservoirs has been computed using various artificially intelligent techniques. The catchment region of Gobindsagar reservoir India is forecasted this study utilizing data collected for years 1971–2003 several models algorithms. Firstly, multi-layered perceptron artificial neural network (MLP-ANN), basic recurrent (RNN), and other RNN based including long-short term memory (LSTM), gated unit (GRU) are implemented to validate predict reservoir. proposed learning validated three influencing factors on yearly basis [rainfall (R a ), water inflow (I w storage capacity (C r )]. results demonstrate that suggested MLP-ANN, RNN, LSTM, GRU produce better with maximum errors reduced from 24.6% 8.05%, 7.52%, 1.77%, 0.05% respectively. For future next 22 years, were first predicted ETS forecasting model help 33 years. Additionally, it was noted each prediction’s error lower than reference model. Furthermore, concluded predicts its alternatives. Secondly, by comparing precision all established study, can be evidently shown LSTM superior MLP-ANN models. It also observed among all, took best due highest R 0.9654 VAF 91.7689%, lowest MAE 0.7777, RMSE 1.1522 MAPE 0.3786%. superiority ensured Taylor’s diagram. Lastly, Garson’s algorithm Olden’s as well perturbation method models, used test sensitivity analysis factor forecasting. discovered most sensitive annual rainfall.