Self-standing wireless sensor node and machine condition classification
作者
Takaharu Yamada,Haruhiko ASANUMA
出处
期刊:Dynamics & Design Conference [The Japan Society of Mechanical Engineers] 日期:2023-01-01卷期号:2023: 452-452
标识
DOI:10.1299/jsmedmc.2023.452
摘要
We investigated a method for collecting vibration acceleration waveforms using wireless sensor node (WSN) powered by piezoelectric vibration energy harvester (PVEH) and classifying bearing fault states using 1D-CNN. First, the power output characteristics of the PVEHs were measured by charging the aluminum electrolytic capacitors, and it was found that the series-type PVEH had larger power output in this study. Next, we proved that the PVEH can power the WSN by shaking it with the sum of the sine wave at the resonant frequency of the PVEH and the vibration of the bearing. Finally, we input the vibration waveforms measured by the WSN into the 1D-CNN to learn the fault state classification and validate the classification accuracy. As a result, a classification accuracy of 95.64% was achieved in the cross-validation method. For future research, we will conduct experiments to verify whether the classification can be performed for a larger number of bearing failure states.