Scholarlink

<p>Complex power quality disturbances (PQDs) pose significant challenges to power systems, causing equipment damage, power loss, and disruptions in industrial processes. Accurate detection of PQDs is crucial for effective mitigation and prevention strategies. Existing methods for PQD detection often suffer from limitations such as low accuracy, limited class coverage, and high computational complexity. To address these limitations, we propose a novel algorithm for detecting and classifying PQDs. A total of 37 PQD patterns are synthesized based on mathematical equations according to the IEEE 1159–2019 Standard. These patterns include single PQDs, double PQDs, and triple PQDs. The modified short-time Fourier transform (MSTFT) is enhanced based on window selection, optimal window size, and reduced boundary effects. Feature selection capabilities are improved by reducing distance, normalizing distances, and using early stopping criteria with the improved single-objective grasshopper optimization algorithm (ISO-GOA). Adaptive sample weighting boosted random forest (ASW-BRF) is employed, iterating through the boosting process and training a new decision tree at each step to improve classification performance. Additionally, PQDs are split using tenfold validation, and hyperparameter tuning is performed using grid search to achieve the highest possible performance. The results demonstrate that the proposed method effectively detects PQD patterns. Only five features were selected and used to achieve high classification performance at an accuracy of 99.84%. The low number of features obtained and the high accuracy of the results demonstrate the suitability of the proposed method for real-world applications.</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

Power quality disturbance detection using improved grasshopper optimization and adaptive boosted random forest

  • 期刊论文

  • 发表日期：

    2025年7月12日

查看全文
  • Supakan Janthong,
  • Pornchai Phukpattaranont

摘要

Complex power quality disturbances (PQDs) pose significant challenges to power systems, causing equipment damage, power loss, and disruptions in industrial processes. Accurate detection of PQDs is crucial for effective mitigation and prevention strategies. Existing methods for PQD detection often suffer from limitations such as low accuracy, limited class coverage, and high computational complexity. To address these limitations, we propose a novel algorithm for detecting and classifying PQDs. A total of 37 PQD patterns are synthesized based on mathematical equations according to the IEEE 1159–2019 Standard. These patterns include single PQDs, double PQDs, and triple PQDs. The modified short-time Fourier transform (MSTFT) is enhanced based on window selection, optimal window size, and reduced boundary effects. Feature selection capabilities are improved by reducing distance, normalizing distances, and using early stopping criteria with the improved single-objective grasshopper optimization algorithm (ISO-GOA). Adaptive sample weighting boosted random forest (ASW-BRF) is employed, iterating through the boosting process and training a new decision tree at each step to improve classification performance. Additionally, PQDs are split using tenfold validation, and hyperparameter tuning is performed using grid search to achieve the highest possible performance. The results demonstrate that the proposed method effectively detects PQD patterns. Only five features were selected and used to achieve high classification performance at an accuracy of 99.84%. The low number of features obtained and the high accuracy of the results demonstrate the suitability of the proposed method for real-world applications.