| Project/Area Number |
18K03907
|
|---|
| Research Category |
Grant-in-Aid for Scientific Research (C)
|
| Allocation Type | Multi-year Fund |
|---|
| Section | 一般 |
|---|
| Review Section |
Basic Section 18040:Machine elements and tribology-related
|
|---|
| Research Institution | Kyoto Institute of Technology |
|---|
Principal Investigator |
Iba Daisuke 京都工芸繊維大学, 機械工学系, 教授 (10402984)
|
|---|
| Project Period (FY) |
2018-04-01 – 2021-03-31
|
| Project Status |
Completed (Fiscal Year 2020)
|
| Budget Amount *help |
¥4,290,000 (Direct Cost: ¥3,300,000、Indirect Cost: ¥990,000)
Fiscal Year 2020: ¥390,000 (Direct Cost: ¥300,000、Indirect Cost: ¥90,000)
Fiscal Year 2019: ¥1,170,000 (Direct Cost: ¥900,000、Indirect Cost: ¥270,000)
Fiscal Year 2018: ¥2,730,000 (Direct Cost: ¥2,100,000、Indirect Cost: ¥630,000)
|
|---|
| Keywords | 歯車 / かみ合い振動 / 人工知能 / 損傷検知 / 樹脂歯車 / 故障診断 / データセット創生 |
|---|
| Outline of Final Research Achievements |
In this study, I have developed a system that automatically generates a large amount of teacher data by collecting various data (engagement vibration, tooth crack images, etc.) that represent the condition of plastic gears. It has become possible to detect the occurrence of tooth cracks by imaging the meshing vibration and learning and analyzing it with artificial intelligence. I also worked on the detection of signs that occur immediately before the occurrence of a tooth crack. As a result of decomposing the measured meshing vibration for each frequency band and investigating the progress of the operation test of the resin gear and the change in the frequency response for each band, changes occurred in some frequency bands before the occurrence of tooth cracks. I confirmed the phenomenon. I proposed a method of using this feature as a change point and using it as a label for learning data for artificial intelligence, and detected signs of tooth cracks.
|
| Academic Significance and Societal Importance of the Research Achievements |
天然の資源に変わり，世界の産業や経済を牽引する広義な意味での新しい資源は「情報」であるとされる．機械要素である歯車においても，「情報資源」を掘り当てて回収・加工し，それを利用する新しい方法の創造が求められている．そこで本研究では樹脂歯車を対象に，資源として樹脂歯車の状態を表す諸量（かみ合い振動や歯元のき裂画像）を自動回収するシステムを開発し，回収した情報から創成した学習用データセットを深層構造を持つ人工知能に学習させることによって，歯車の振動情報のみから損傷検知及び損傷モードの分類を行うシステムを開発した．
|
