Ship bearing piezoelectric energy recovery device and control method

CN122371730APending Publication Date: 2026-07-10WUHAN UNIV OF SCI & TECH
View PDF 0 Cites -1 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
WUHAN UNIV OF SCI & TECH
Filing Date
2026-03-31
Publication Date
2026-07-10

AI Technical Summary

Technical Problem

Existing technologies have limitations in adaptability for energy harvesting from self-excited vibrations of ship bearing friction. In particular, energy output is unstable in confined spaces and complex marine environments. Furthermore, deep learning models lack diagnostic accuracy in small sample scenarios and cannot be adapted to passive monitoring scenarios.

Method used

A ship bearing piezoelectric energy recovery device based on convolutional residual neural network is adopted. By combining Gram angle difference field conversion, multi-scale denoising module, attention module and feature enhancement module, GAM-DRCN model is constructed to realize real-time monitoring and energy harvesting of bearing vibration signal.

Benefits of technology

The diagnostic accuracy is improved to 99.2% in noisy, small-sample scenarios, the model parameters are reduced by 25%, it is adapted to passive monitoring scenarios, improves fault classification accuracy and energy harvesting efficiency, and is suitable for passive monitoring and energy recovery of ship bearings.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN122371730A_ABST
    Figure CN122371730A_ABST
Patent Text Reader

Abstract

This invention discloses a piezoelectric energy recovery device and control method for ship bearings. A piezoelectric cantilever beam vibration energy harvesting device (6) is fixed on the lower surface of the outer ring of the bearing (4). A GAM-DRCN monitoring system (9) incorporates a GAM-DRCN model, which includes a sequentially connected Gram angle difference field conversion layer, an MSS multi-scale denoising module, a CBAM attention module, and an FEM feature enhancement module. The MSS multi-scale denoising module adopts a Res2NeXt structure combined with SE channel attention and a soft threshold denoising mechanism. The system is configured to receive vibration signals collected by sensors and output the voltage prediction results of the piezoelectric device. This invention can adapt to passive monitoring scenarios in actual ship navigation, as well as real-world scenarios where vibration signals are affected by multiple factors such as marine environmental noise and equipment coupling interference during actual ship navigation. It is suitable for rapid processing and monitoring of noisy signals in scenarios with low computational load and small sample size.
Need to check novelty before this filing date? Find Prior Art