Application of Teaching-Learning based optimization algorithms in the operation of Eleviyan reservoir considering environmental demand

Abstract
Background and Objectives
Generally, Iran has an arid and semi-arid climate with a high population. In recent years, the shortages of surface water and groundwater have become a main national challenge in Iran. The optimum operation of reservoirs is one of the most critical challenges in water resources management. In a situation where the drying crisis of the Urmia Lake is serious, comprehensive management of water resources in this basin, allocation of environmental water rights, and the optimal operation of dam reservoirs are the most principled methods of allocation to combat the drying crisis of this lake. The choice of management policies and optimization method depends on the system specification, availability of data, type of objective function, constraints, and variables. The Eleviyan dam is located on the Sufi-Chay River and is one of the main existing dams in the Urmia Lake basin. Therefore, in this research, the performance of the Teaching Learning Based Optimization algorithm (TLBO) is compared with Improved Harmony Search (IHS) and Particle Swarm Optimization (PSO) algorithms in order to optimize the Eleviyan reservoir operation, located across the Sufi-Chay River.
Methodology
Considering the vital importance of environmental flow to prevent the death of Urmia Lake, the minimum environmental flow of the Sufi-Chay River was estimated by two hydrological methods: Tennant (first scenario) and Flow Duration Curve analysis (second scenario). The Teaching Learning Based Optimization algorithm, Improved Harmony Search, and Particle Swarm Optimization algorithms were used for optimization. The optimization models were written in the form of two scenarios, taking into account the complete provision of the minimum environmental flow, the municipal demands, and industry, and minimizing the severity of agricultural shortages. In order to check the performance of the studied algorithms in the optimal exploitation of the reservoir, the performance indicators of the reservoir including reliability, vulnerability, and stability index of the reservoir were used.
Findings
The results in this study showed better performance of the TLBO algorithm method compared with both scenarios. In TLBO algorithm, the objective function values for both scenarios are calculated, respectively 2.43 and 7.54. While with Improved Harmony Search algorithm objective function values are calculated respectively 2.81 and 8, and with the Particle Swarm Optimization algorithm objective function values are calculated 3.34 and 8.45 with both Tenant and Flow Duration Curve (FDC) methods. Based on the obtained results, the TLBO algorithm has been able to supply the downstream demand of the dam. So, for the first scenario, it will provide 81.40% and for the second scenario, 62.55% of the demands. The release of observation has been able to provide about 70% of the total demand and according to the situation of Urmia Lake, the current policy of the dam is not suitable and should be revised. Taking place according to the second scenario (flow duration curve method), in case of reducing about 8% of agricultural expenses and allocating it to the environmental flow, the problem of water rights of Urmia Lake from the Sufi-Chay River will also be solved to some extent.
Conclusion
One of the main causes of drying of the Urmia Lake is the reduction of inflow to the Lake due to the development of agricultural lands, allocation of irregular water resources to the agricultural sector regardless of the environmental right of water, and irregular yields of surface and groundwater resources. Regarding the scenarios, it seems that if the downstream agricultural demand is reduced by about 8% in methods such as lessened irrigation of gardens, changes to crop cultivation, and revising the curve of dams around the Lake, it would aid the recovery of Urmia Lake and play an important role in adequately supplying the downstream needs. The results show that managing the Eleviyan dam is required to reduce the agricultural demand through the methods such as reducing the crop area and changing the cropping patterns such as low consumption plant cultivation or changing the type of irrigation method in order to optimize the operation of water resources in the basin. Otherwise, water stress will increase in the basin. The results in this study showed a higher performance of the Teaching algorithm compared with the other methods employed in the optimum release from the Eleviyan reservoir with considering environmental flow.