EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Study on fault detection of generator slot wedge tightness based on improved KNN

Longjiang Dou, Wenjie Li, Bingnan Wang, Tengbo Zhang, Qian Cao

Machine Learning: Engineering · 2025

DOI: 10.1088/3049-4761/ae07e2 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Machine Learning: Engineering, 2025. Verfasst von Longjiang Dou und 4 weiteren Autor:innen. Die Arbeit nennt 20 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: National Natural Science Foundation of China, 考虑实际风速时空分布的风力发电机组故障特征分析与机电联合故障识别研究 保定市科技计划基础研究专项基金, supported by the Fundamental Research Funds for the Central Universities.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

This paper proposes a fault detection method for detecting generator slot wedge looseness, utilizing an improved K-Nearest Neighbors (KNN) approach. The primary challenge addressed is the over-reliance on manual experience and the low accuracy in detecting slot wedge faults in generators. Initially, three distinct states of slot wedges—severe looseness, slight looseness, and normal state—are simulated on a test bench, with corresponding sound signals collected. Subsequently, these sound signals undergo Fourier transformation, and the looseness fault feature parameters are extracted. This paper identifies decay time as an effective parameter for diagnosing slot wedge looseness through experimental analysis, marking its first use in such a context. Alongside maximum peak frequency and bandwidth area, these final feature parameters are employed for detecting slot wedge looseness faults. Experimental results demonstrate that the accuracy of the results obtained using the selected fault feature parameters for slot wedge fault detection is higher with the improved KNN, compared to traditional KNN, extreme learning machine, support vector machine, and Backpropagation neural network. The improved KNN algorithm, in particular, shows higher fault diagnosis accuracy, reaching up to 99.6%. This level of precision suggests that the method can be fully applied to the fault detection of generator slot wedge looseness.

Förderung

  • National Natural Science Foundation of China
  • 考虑实际风速时空分布的风力发电机组故障特征分析与机电联合故障识别研究 保定市科技计划基础研究专项基金
  • supported by the Fundamental Research Funds for the Central Universities
  • 中央高校基本科研业务费

Zitieren

APA

Dou, L., Li, W., Wang, B., Zhang, T., & Cao, Q. (2025). Study on fault detection of generator slot wedge tightness based on improved KNN. Machine Learning: Engineering. https://doi.org/10.1088/3049-4761/ae07e2

Vancouver

Dou L, Li W, Wang B, Zhang T, Cao Q. Study on fault detection of generator slot wedge tightness based on improved KNN. Machine Learning: Engineering. 2025. doi:10.1088/3049-4761/ae07e2

BibTeX
@article{Dou2025,
  author = {Dou, Longjiang and Li, Wenjie and Wang, Bingnan and Zhang, Tengbo and Cao, Qian},
  title = {Study on fault detection of generator slot wedge tightness based on improved KNN},
  journal = {Machine Learning: Engineering},
  year = {2025},
  doi = {10.1088/3049-4761/ae07e2}
}
Kostenlos starten →