Scholarlink

Purpose <p>Gear stiffness degradation is a critical factor influencing the vibration response and operational stability of helical gear drives (HGD). It can significantly affect the remaining useful life of the gearbox, potentially leading to failure. Accurate prediction of dynamic responses and stiffness degradation is essential for early detection of faults.</p> Methods <p>A novel cross-timescale prediction (CTSP) method for dynamic response and stiffness degradation prediction of HGD is proposed in this study. A duffing oscillator is introduced to generate a nonlinear prediction model of HGD, in which the deterministic dynamic mesh force and the random one is evaluated on two timescales, transient one and operational one, so that the correlation between the gear stiffness degradation and time can be quantified. The state parameters, including displacement, velocity and stiffness responses, on the transient timescale are evaluated using the CTSP method, and the stiffness degradation process on the operational time scale is further predicted based on the former estimated data points.</p> Results <p>The results show that the dynamic mesh force and the motion in the line of action (LOA) of HGD is periodic. The predicted results of displacement and velocity agree well with their ground truth, and the stiffness degradation prediction results above 98% confidence interval. The accuracy of the CTSP method in dynamic response and stiffness degradation prediction are slightly affected by the input torque and rotate speed, which verifies the excellent performance and applicability of the CTSP method. Moreover, the minimum MAPE of CTSP method, AR model and ARIMA model is 0.29%, 1.25% and 0.82%, and the minimum RMSE is 1.78×106 N/m, 9.43×106 N/m and 4.25×106 N/m, respectively.</p> Conclusion <p>The proposed CTSP method effectively predicts both the dynamic response and stiffness degradation of HGD, showing high accuracy and minimal error fluctuation compared to other models. The proposed method offers significant potential for the early detection of gear faults and performance degradation, contributing to improved reliability and extended lifespan of HGD.</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 Cross-Timescale Prediction Method for Vibration and Stiffness Degradation of Helical Gear Drive

  • 期刊论文

  • 发表日期：

    2025年1月17日

查看全文
  • Shuiguang Tong,
  • Xianmiao Yang,
  • Zheming Tong,
  • Yiwei Hu,
  • Huabin Li

摘要

Purpose

Gear stiffness degradation is a critical factor influencing the vibration response and operational stability of helical gear drives (HGD). It can significantly affect the remaining useful life of the gearbox, potentially leading to failure. Accurate prediction of dynamic responses and stiffness degradation is essential for early detection of faults.

Methods

A novel cross-timescale prediction (CTSP) method for dynamic response and stiffness degradation prediction of HGD is proposed in this study. A duffing oscillator is introduced to generate a nonlinear prediction model of HGD, in which the deterministic dynamic mesh force and the random one is evaluated on two timescales, transient one and operational one, so that the correlation between the gear stiffness degradation and time can be quantified. The state parameters, including displacement, velocity and stiffness responses, on the transient timescale are evaluated using the CTSP method, and the stiffness degradation process on the operational time scale is further predicted based on the former estimated data points.

Results

The results show that the dynamic mesh force and the motion in the line of action (LOA) of HGD is periodic. The predicted results of displacement and velocity agree well with their ground truth, and the stiffness degradation prediction results above 98% confidence interval. The accuracy of the CTSP method in dynamic response and stiffness degradation prediction are slightly affected by the input torque and rotate speed, which verifies the excellent performance and applicability of the CTSP method. Moreover, the minimum MAPE of CTSP method, AR model and ARIMA model is 0.29%, 1.25% and 0.82%, and the minimum RMSE is 1.78×106 N/m, 9.43×106 N/m and 4.25×106 N/m, respectively.

Conclusion

The proposed CTSP method effectively predicts both the dynamic response and stiffness degradation of HGD, showing high accuracy and minimal error fluctuation compared to other models. The proposed method offers significant potential for the early detection of gear faults and performance degradation, contributing to improved reliability and extended lifespan of HGD.