Intelligent calibration method for multi-stage gearbox dynamics model

By constructing a dynamic model of a multi-stage gearbox that considers the flexibility of the drive shaft and gearbox housing, and combining a surrogate model and optimization algorithm, the problem of large deviation between simulation and actual results in the existing technology is solved, and efficient and accurate prediction of the dynamic response of multi-stage gearboxes is achieved, supporting engineering applications.

CN122133274APending Publication Date: 2026-06-02ZHEJIANG TONGLI HEAVY GEAR
View PDF 0 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
ZHEJIANG TONGLI HEAVY GEAR
Filing Date
2026-02-10
Publication Date
2026-06-02

AI Technical Summary

Technical Problem

Existing multi-stage gearbox dynamic models neglect the flexibility of the transmission shaft and gearbox structure when considering the gear system, resulting in large deviations between simulation and actual results. Furthermore, traditional calibration methods are time-consuming to calculate and are prone to getting stuck in local optima or failing to converge, making it difficult to achieve high-precision predictions.

Method used

An intelligent calibration method combining surrogate models, parameter sensitivity analysis, and optimization algorithms is adopted to construct a multi-stage gearbox dynamic model that considers the flexibility of the drive shaft and gearbox structure. Key parameters are screened through Pearson correlation analysis, and iterative intelligent calibration is performed using a Kriging surrogate model to reduce computational costs and time.

Benefits of technology

It improves the prediction accuracy of the dynamic model of multi-stage gearboxes, shortens the single simulation time to the second level, reduces the computational cost, realizes accurate dynamic response prediction of multi-stage gearboxes under real working conditions, and supports vibration performance evaluation and fatigue life prediction.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN122133274A_ABST
    Figure CN122133274A_ABST
Patent Text Reader

Abstract

This invention relates to an intelligent calibration method for a multi-stage gearbox dynamic model. First, a multi-stage gearbox dynamic model considering the flexibility of the drive shaft and gearbox structure is constructed. The spatial vibration response at five measuring points, collected from vibration characteristic tests of the multi-stage gearbox, serves as the benchmark for intelligent calibration. Then, Pearson correlation analysis is used to screen correction parameters for the multi-stage gearbox dynamic model. Optimal Latin hypercube sampling and solving the multi-stage gearbox dynamic model yield the data required to construct a surrogate model. A Kriging model is then used to construct the surrogate model for the multi-stage gearbox dynamic model. Finally, minimizing the error between the simulation and experimental vibration results at each measuring point is the objective of intelligent calibration. The surrogate model and intelligent optimization algorithm are combined to complete the intelligent calibration of the multi-stage gearbox dynamic model. This invention significantly improves the prediction accuracy and calibration efficiency of the multi-stage gearbox dynamic model while reducing computational costs.
Need to check novelty before this filing date? Find Prior Art