| Title: | Smart AI-based system for turning tool condition monitoring |
|---|
| Authors: | ID Brili, Nika (Author) |
|---|
| Files: |  URL - Source URL, visit https://itis.fis.unm.si/wp-content/uploads/2026/01/ITIS-2025-Proceedings_FINAL.pdf
 PDF - Presentation file, download (12,76 MB)
MD5: EC7343CA10134BAE709B019667700F4A
|
|---|
| Language: | English |
|---|
| Typology: | 1.12 - Published Scientific Conference Contribution Abstract |
|---|
| Organization: |  RUDOLFOVO - Rudolfovo - Science and Technology Centre Novo Mesto
|
|---|
| Abstract: | The turning process is a widely used cutting operation in industry. Any optimization of this process can significantly improve product quality, streamline costs, or reduce unwanted events. With automatic monitoring of turning tools, we can reduce costs, increase efficiency, and decrease the number of undesirable events that occur during machining (scrap, tool breakage, etc.). In single-piece or small-batch production, tool wear is monitored by the machine operator; however, such wear assessment is left to subjective judgment and requires intervention in the process. The presented solution eliminates this problem with automated monitoring of the cutting tool’s condition. An IR camera was used for process monitoring, which also captures the thermographic state. The camera was properly protected and mounted right next to the turning tool, enabling close-up observation of the machining. During the experiment, constant cutting parameters were set for turning the workpiece (low-alloy steel designated 1.7225, i.e.,42CrMo4) without the use of coolant. Using turning inserts with varying levels of wear, a database of more than 6,000 images was created during the turning process. With a convolutional neural network (CNN), a model was developed to predict wear and damage to the cutting tool. Based on the captured thermographic image during turning, the model automatically determines the cutting tool’s condition (no wear, minor wear, severe wear).The achieved classification accuracy was 99.55%, confirming the suitability of the proposed method. Such a system enables immediate action in the event of tool wear or breakage, regardless of the operator’s knowledge and training. |
|---|
| Keywords: | deep learning, tool condition monitoring, turning, tool wear |
|---|
| Publication status: | Published |
|---|
| Publication version: | Version of Record |
|---|
| Year of publishing: | 2025 |
|---|
| Number of pages: | Str. [214] |
|---|
| PID: | 20.500.12556/DiRROS-27216  |
|---|
| UDC: | 004.8:004.92:621.941 |
|---|
| COBISS.SI-ID: | 265308163 |
|---|
| Note: | Nasl. z nasl. zaslona;
Opis vira z dne 20. 1. 2026;
|
|---|
| Publication date in DiRROS: | 03.02.2026 |
|---|
| Views: | 129 |
|---|
| Downloads: | 97 |
|---|
| Metadata: |    |
|---|
|
:
|
Copy citation |
|---|
| | | | Share: |  |
|---|
Hover the mouse pointer over a document title to show the abstract or click
on the title to get all document metadata. |
