Scholarlink

<p>Optimal placement and sizing of distributed energy resources (DERs) in radial distribution networks are vital for improving efficiency, reliability, and economic viability. However, conventional optimization techniques often face slow convergence, limited adaptability, and poor uncertainty handling. To overcome these challenges, this study introduces a hybrid framework that integrates a modified cheetah optimizer (MCO) with a regression-based machine learning model. The MCO, inspired by the hunting dynamics of cheetahs, adaptively balances exploration and exploitation, while the ML model predicts voltage profiles to guide the search process and enhance convergence speed. The framework jointly optimizes power loss, voltage deviation, and voltage stability, and incorporates a region-specific economic model that includes capital costs, tariffs, and grid penalties. Renewable energy variability is modelled using NASA-SSE solar and wind data via Weibull distributions. A multi-run strategy with 30 independent executions ensures robustness, while ANOVA analysis validates the statistical superiority of the method over existing algorithms. Tested on the IEEE 33-bus system, the approach achieves a 94.2% reduction in power losses, improves voltage stability index to 0.951, and delivers annual cost savings exceeding $77,000. Convergence curves, boxplots, and sensitivity analyses confirm the framework’s reliability, scalability, and practical value for DER planning under uncertainty.</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

Modified cheetah optimizer for optimal techno-economic placement of distributed energy source planning

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2025年10月21日

查看全文
  • Sangeeta Debbarman,
  • Kumari Namrata,
  • Akshit Samadhiya,
  • Ahmad Taher Azar,
  • Saim Ahmed,
  • Ahmed Redha Mahlous,
  • Walid El-Shafai

摘要

Optimal placement and sizing of distributed energy resources (DERs) in radial distribution networks are vital for improving efficiency, reliability, and economic viability. However, conventional optimization techniques often face slow convergence, limited adaptability, and poor uncertainty handling. To overcome these challenges, this study introduces a hybrid framework that integrates a modified cheetah optimizer (MCO) with a regression-based machine learning model. The MCO, inspired by the hunting dynamics of cheetahs, adaptively balances exploration and exploitation, while the ML model predicts voltage profiles to guide the search process and enhance convergence speed. The framework jointly optimizes power loss, voltage deviation, and voltage stability, and incorporates a region-specific economic model that includes capital costs, tariffs, and grid penalties. Renewable energy variability is modelled using NASA-SSE solar and wind data via Weibull distributions. A multi-run strategy with 30 independent executions ensures robustness, while ANOVA analysis validates the statistical superiority of the method over existing algorithms. Tested on the IEEE 33-bus system, the approach achieves a 94.2% reduction in power losses, improves voltage stability index to 0.951, and delivers annual cost savings exceeding $77,000. Convergence curves, boxplots, and sensitivity analyses confirm the framework’s reliability, scalability, and practical value for DER planning under uncertainty.