Scholarlink

Objective <p>This study aims to propose a high-dimensional coupled system based on the FitzHugh–Nagumo (FHN) neuron model to explore the stochastic resonance (SR) phenomenon driven by both external periodic excitation and random noise, and apply it to bearing fault diagnosis to improve the ability of fault signal detection.</p> Methods <p>Firstly, the equivalent potential function of the single FHN neuron model is derived to determine the range of bistable characteristics of neuron discharge dynamics. Then, the stationary probability density (SPD), transition rate, and output signal-to-noise ratio (SNR) of the single FHN neuron model are derived using the two-state theory. Next, the analysis is extended to a three-dimensional coupled FHN neuron model, and the influence of parameters on the output SNR is studied through numerical simulation. Finally, a high-dimensional bidirectional coupled FHN neuron method is proposed, and the adaptive genetic algorithm (AGA) is used to determine the optimal output of the system.</p> Results <p>The coupled FHN neuron model shows a significant SR effect under the combined influence of noise and coupling; bidirectional coupling achieves the highest output mean signal-to-noise ratio (MSNR) in the FHN system, and there always exists an optimal dimension <i>n</i> that maximizes the MSNR of the output signal; this method outperforms the single FHN system in detecting weak fault signals, and coupling at the end yields higher SNR compared to coupling at the center.</p> Conclusion <p>This research significantly enhances the ability of fault signal detection, provides a more robust method compared to the traditional single neuron system, and improves the reliability and accuracy of fault diagnosis in various 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

A Novel High-Dimensional Coupled FHN Neuron Stochastic Resonance Model and its Performance in Faults Recognition

  • 期刊论文

  • 发表日期：

    2025年1月3日

查看全文
  • Lifang He,
  • Xiaoxiao Huang,
  • Jiachen Hou

摘要

Objective

This study aims to propose a high-dimensional coupled system based on the FitzHugh–Nagumo (FHN) neuron model to explore the stochastic resonance (SR) phenomenon driven by both external periodic excitation and random noise, and apply it to bearing fault diagnosis to improve the ability of fault signal detection.

Methods

Firstly, the equivalent potential function of the single FHN neuron model is derived to determine the range of bistable characteristics of neuron discharge dynamics. Then, the stationary probability density (SPD), transition rate, and output signal-to-noise ratio (SNR) of the single FHN neuron model are derived using the two-state theory. Next, the analysis is extended to a three-dimensional coupled FHN neuron model, and the influence of parameters on the output SNR is studied through numerical simulation. Finally, a high-dimensional bidirectional coupled FHN neuron method is proposed, and the adaptive genetic algorithm (AGA) is used to determine the optimal output of the system.

Results

The coupled FHN neuron model shows a significant SR effect under the combined influence of noise and coupling; bidirectional coupling achieves the highest output mean signal-to-noise ratio (MSNR) in the FHN system, and there always exists an optimal dimension n that maximizes the MSNR of the output signal; this method outperforms the single FHN system in detecting weak fault signals, and coupling at the end yields higher SNR compared to coupling at the center.

Conclusion

This research significantly enhances the ability of fault signal detection, provides a more robust method compared to the traditional single neuron system, and improves the reliability and accuracy of fault diagnosis in various applications.