Scholarlink

<p>This study optimizes reservoir operation for water quality enhancement by determining optimal withdrawal amounts. A meta-model-based optimization simulation approach is employed to improve outflow quality, addressing downstream water demands. Hydrodynamic and water quality simulations for the Ekbatan Dam were performed using the CE-QUAL-W2 (Comprehensive Water Quality - Water Quality Model 2) model. To mitigate computational costs associated with multiple CE-QUAL-W2 calls, a Supervised Learning (SL) surrogate model was developed and integrated with the Fruit-fly Optimization Algorithm (FOA), forming FOA-SL, for estimating Total Dissolved Solids (TDS) and minimizing outflow TDS concentration. The CE-QUAL-W2 model was calibrated and validated using 2019–2020 data, achieving acceptable performance metrics (<i>NSE</i>, <i>MAE</i>, <i>RMSE</i>). FOA-SL demonstrated rapid convergence, reaching solutions in just 300 iterations compared to 1,000,000 iterations for the standalone FOA. A comparative analysis with the Genetic Algorithm (GA) for operational optimization revealed that FOA-SL achieved superior objective function values (lower TDS). While GA offered faster individual run times (approximately 38&#xa0;min), FOA-SL achieved greater accuracy. The optimized operation led to an approximate 0.3% reduction in outflow TDS on March 18, 2020, with a final Relative Bias (RB) of 1.7&#xa0;mg/l. This strategy involves increased withdrawal rates, particularly during high TDS periods, to reduce pollutant accumulation and improve overall water quality. This research offers an effective approach for reservoir water quality management.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Simulation-Optimization Approach Based on the Meta-Model in the Optimal Operation of a Single Reservoir Using the Development of the FOA-SL Surrogate Model

  • 期刊论文

  • 发表日期：

    2026年5月18日

查看全文
  • Seyedeh Zahra Hassani,
  • Parisa-Sadat Ashofteh

摘要

This study optimizes reservoir operation for water quality enhancement by determining optimal withdrawal amounts. A meta-model-based optimization simulation approach is employed to improve outflow quality, addressing downstream water demands. Hydrodynamic and water quality simulations for the Ekbatan Dam were performed using the CE-QUAL-W2 (Comprehensive Water Quality - Water Quality Model 2) model. To mitigate computational costs associated with multiple CE-QUAL-W2 calls, a Supervised Learning (SL) surrogate model was developed and integrated with the Fruit-fly Optimization Algorithm (FOA), forming FOA-SL, for estimating Total Dissolved Solids (TDS) and minimizing outflow TDS concentration. The CE-QUAL-W2 model was calibrated and validated using 2019–2020 data, achieving acceptable performance metrics (NSE, MAE, RMSE). FOA-SL demonstrated rapid convergence, reaching solutions in just 300 iterations compared to 1,000,000 iterations for the standalone FOA. A comparative analysis with the Genetic Algorithm (GA) for operational optimization revealed that FOA-SL achieved superior objective function values (lower TDS). While GA offered faster individual run times (approximately 38 min), FOA-SL achieved greater accuracy. The optimized operation led to an approximate 0.3% reduction in outflow TDS on March 18, 2020, with a final Relative Bias (RB) of 1.7 mg/l. This strategy involves increased withdrawal rates, particularly during high TDS periods, to reduce pollutant accumulation and improve overall water quality. This research offers an effective approach for reservoir water quality management.