| Autor: |
Tobias Gaul, Frank Duckhorn, Mareike Stephan, Stefan Bonerz, Lars Schubert |
| Jazyk: |
German<br />English<br />Spanish; Castilian<br />French<br />Italian |
| Rok vydání: |
2024 |
| Předmět: |
|
| Zdroj: |
e-Journal of Nondestructive Testing, Vol 29, Iss 10 (2024) |
| Druh dokumentu: |
article |
| ISSN: |
1435-4934 |
| DOI: |
10.58286/30485 |
| Popis: |
Numerous sensors are integrated into modern machine tools to detect and control the forces acting on the tool or workpiece, or to monitor mechanical components such as ball bearings. In addition to the machine itself, the tool is also a sensitive component. In extreme cases, spontaneous failure of the tool can lead to a collision with the clamped workpiece and cause irreparable damage. This results in subsequent costs due to material loss, repair, or production downtime, which can be considerable high. It is therefore important to detect possible wear or failure of tools during the production process. Tool monitoring is particularly useful for flexible production systems which, due to their dynamic application situation, do not allow an estimation of tool life and therefore no fixed tool maintenance intervals. Consequently, the integration of sensors into the machine tool is intended to detect wear at an early stage and minimize tool failure. The integration of a sensor system was carried out in collaboration with Ott Jacob Spanntechnik during a joint research project. The purely mechanical component of the rotary union of a motor spindle was expanded to include an integrated acoustic emission sensor. The focus was on the development of inexpensive sensors for integration into the machine and an evaluation procedure for detecting tool breakage. This paper presents the evaluation principles used and examines their applicability in a manufacturing process. At first, acoustic emission signals from tool breakages were generated in laboratory on a simplified cooling channel setup. The measurement data were used to detect breakage events using different acoustic emission parameters and machine learning methods. It was shown that both methods can identify drill break in the signal, regardless of the diameter. The measurements were then repeated on a machine tool under realistic operating conditions and extended to include wear detection. |
| Databáze: |
Directory of Open Access Journals |
| Externí odkaz: |
|
