Advanced intelligent cutting tool wear monitoring system based on thermography
Advanced intelligent cutting tool wear monitoring system based on thermography
Project duration: 1 October 2023 - 30 September 2025
Project leader: Dr. Nika Brili
(https://cris.cobiss.net/ecris/si/sl/researcher/41552)
Project type: postdoctoral basic research project
Funding level: 1,00 FTE per year, category B
The control of cutting tool wear in turning contributes to improving product quality, optimising tool costs and reducing unwanted events (production stoppage, workpiece ejection, etc.). It is important both for the quality of the final product and for the optimisation of processes. The problem is particularly pronounced in small batch and individual production, where the decision on the suitability of the cutting tool is left to the machine operator, who makes decisions based on experience and personal judgement (informal knowledge). The main idea of the research project is the automated detection of cutting tool wear and the decision to change the cutting tool will be made, independently of the operator's knowledge and experience.
The main objective of the research is: to develop a reliable tool condition monitoring (TCM) system based on the thermographic image of the cutting tool using deep learning methods.
The project will produce a comprehensive database of thermographic images that will be available to other scientists and will make a major contribution to the development of the science of tool condition monitoring using artificial intelligence.
As it is difficult to provide a sufficiently large database for turning with worn tools, the Generative Adversarial Network (GAN) method will be used to produce additional images based on the captured images. This will ensure a balanced database for all categories.
Project phases:
- DP1 - PREEXPERIMENTAL PHASE (D1.1 Detailed experimental design)
- DP2 - EXPERIMENT (D2.1 Image Database, D2.2 Extended Image Database, D2.3 Big Data)
- DP3 - GAN (D3.1 GAN-based extended database)
- DP4 - CLASSIFICATION (D4.1 Learned model for classification, D4.2 Deep neural network)
The research project is funded by the Slovenian Agency for Scientific Research and Innovation