Document Type : Research/Original/Regular Article
Authors
1
Ph.D. candidate, Department of Hydraulic and Water Resources Engineering, AWTi, Arba Minch University, Arba Minch, Ethiopia
2
Associate Professor, Department of Hydraulic and Water Resources Engineering, AWTi, Arba Minch University, Arba Minch, Ethiopia
Abstract
The Gibe Cascaded reservoir system in Ethiopia is vital for national energy security, but complex hydrology and water scarcity create significant operational challenges. This study addresses these gaps by developing an integrated optimization-extraction framework, using successive linear programming (SLP) to manage the non-linear relationship between reservoir head and power generation. Ethiopian Electric Power (2004–2022) provided reservoir operations, whereas the Ministry of Water and Energy (1995–2024) provided hydrology, 12.5 m DEM, land use/cover and soil data. While dam features and demand data derived from Webuild Group design documents, meteorological data sourced from Ethiopian Meteorological Institute (1990–2024). The SLP model was evaluated against HEC-ResSim and historical data using reliability, resilience, and vulnerability (RRV) analysis. The SLP model achieves superior efficiency, outperforming Baseline (23,760 MWh) and Observed (18,610 MWh) scenarios with daily mean energy gains of 26,180 MWh and high reliability scores of 0.91 and 0.94 for Gibe-I and Gibe-II reservoirs. The SLP model increases annual energy by 10% (883,000 MWh) and reduces vulnerability (0.23 vs. 0.44). This suggests optimization ensures infrequent, shallow failures, while traditional rule-based management prioritizes more rapid recovery. The SLP optimization successfully buffered short-term deficits in single dry years via strategic storage drawdowns, sustaining high reliability (≥ 82%), elevated resilience (≥ 59%), and minimal vulnerability (≤ 7%). The SLP approach exhibits high resilience to short-term hydrological stress, validating its effectiveness for operational decision-making and long-term planning in drought-susceptible basins. To apply the optimized results in practice, three rule-curve extraction methods were tested: Symbolic Regression (STORAGE_SR), the integration of K-Means clustering and Fourier series, and Frequency Interval Analysis. The STORAGE_SR proved the most reliable, with R² and NSE above 0.80 for both Gibe-I and Gibe-III. By balancing hydropower production and release requirements, this study provides a practical strategy to mitigate energy shortages and enhance the resilience of the flow uncertainty. Ultimately, these advancements directly support Sustainable Development Goal (SDG) 7 (Affordable and Clean Energy) by boosting renewable output, and SDG 13 (Climate Action) by strengthening infrastructure resilience against climate-driven hydrological variability.
Keywords
Main Subjects
)