Mohammad Taghi Aalami

Reservoir dams play a pivotal role in water resource management. Accurate prediction of inflow to reservoirs significantly enhances operational performance. While standalone artificial intelligence methods have recently been frequently used to predict inflow, hybrid models have shown quite more satisfactory success. In this study, various deep learning models, including MLP, GRU, LSTM, CNN, CNN-MLP, CNN GRU, CNN-LSTM, CNN-GRU-MLP, and CNN-LSTM-MLP, were utilized to predict the monthly inflow to the Aras reservoir in the Azerbaijan-Iran region. The results were compared with the Neural Basis Expansion Analysis for Time Series Forecasting (N-BEATS) model for univariate forecasting and the NBEATSx model for multivariate forecasting using a monthly inflow time series dataset. To enhance prediction accuracy, the hyperparameters of the models were optimized. Additionally, to evaluate the impact of feature selection on model performance, five different scenarios were developed as combinations of input variables for forecasting one future time step. The evaluation metrics revealed that among the scenarios, Scenario 5 (comprising lagged inflows at months 1, 11, and 12; lagged average monthly precipitation in the upstream basin at months 1 and 12; the solar month counter; and a three-month moving average of monthly inflow) yielded the best results. Among the models, the hybrid CNN-LSTM-MLP demonstrated the highest prediction accuracy. Specifically, the performance metrics for this model and the best scenario included MAE, RMSE, PBIAS, R², KGE, and NSE, which were 8.78 m³/s, 12.95 m³/s, 1.5%, 0.89, 0.91, and 0.89, respectively. Conversely, the NBEATSx model exhibited suboptimal performance, with reduced accuracy as the number of input features increased, although the N-BEATS model performed well in univariate forecasting. This study highlights the high potential of hybrid deep learning models in accurately forecasting reservoir inflows and underscores their utility in enhancing water resource and reservoir operation management.

As one of the largest super-saline lakes in the world, Lake Urmia in northwestern Iran has been facing severe drying in recent years. Drought and rapid expansion of agricultural activities are considered to be the main driving factors in the shrinking of the lake. To address this problem, an analysis of the spatiotemporal dynamics of land use/land cover (LULC) is important. This research implemented a multi-source satellite image analysis through support vector machine (SVM) for mapping LULC distributions for the years 2000, 2010, and 2020. Cellular automata (CA)–Markov was prepared for modeling the future landscape changes for 2030 and 2040. In the last step, the water requirement of agriculture in the catchment area of the Urmia Lake was simulated through the NETWAT model. Through the employed future LULC modeling, it was found that the areas covered by irrigated agriculture and gardens will grow from 1,450 and 395 km2 to 3,600 and 1,650 km2 in 2040, respectively, as deduced from the changes that occurred from 2000 to 2020. This will increase the water requirement of agriculture from 1.5 billion cubic metres in 2000 to more than 4.1 billion cubic metres in 2040.

In this paper, the failure costs of an arch dam’s optimum shape were investigated, taking into account the interaction between dam, reservoir, foundation, and hydrodynamic effects resulting from seismic loading. Construction, lifecycle, and failure costs were considered to address the risk-based arch dam optimization problem. Both the flexibility and damping effects of the foundation rock were considered. The optimization was carried out using a hybrid algorithm on the basis of the charged system search and big bang-big crunch methods. A real-world arch dam was investigated as a numerical example to demonstrate the efficiency of the proposed method. The numerical results revealed the computational advantages of the new method and the new defined risk optimization problem for optimal design of arch dams.

This paper focuses on the operational modal analysis of the Karun IV dam, the highest concrete arch dam in Iran, based on the responses obtained from ambient vibration tests, which are carried out on the dam during 19 March 2014, and 3 March 2015 earthquake with an epicenter about 133 km northeast of the dam. The non-parametric FDD-Wavelet method is used for identification and extraction of modal parameters including natural frequencies, mode shapes and damping ratios. Nine and four natural frequencies are identified from ambient and seismic responses, respectively. Stabilization diagram and averaged coherence spectrum (ACS) are used to detect and eliminate spurious form true modes. Comparison of the results indicates that identification is performed properly and reliably.

This paper describes the potential use of artificial intelligence (AI)-based techniques to predict monthly groundwater level (GWL) and filling missing GWL data. The adaptive neuro-fuzzy inference system (ANFIS) model conjugated to data pre-processing methods was used for GWL modeling of a Tampa Bay plain at three distinct scenarios. In the proposed methodology, the self-organizing map (SOM) clustering method, as spatial preprocessor, and the discrete wavelet transform (DWT), as temporal data preprocessor, were linked to the ANFIS framework. The original time series of precipitation, runoff, and GWL parameters were decomposed into multifrequency subseries by DWT, and then the subseries were imposed as input data into an ANFIS model for each cluster identified by the SOM, for prediction of the GWL value one and multitime step ahead and to fill missing GWL data. The wavelet transform coherence (WTC) technique was also used for selecting dominant input variables of the ANFIS model. The performance of ...