Artificial intelligence detects cutting tool failures in real time

A new artificial intelligence system which detects problems with cutting tools on manufacturing lines with 100 per cent accuracy could prevent catastrophic failures – and even one day monitor machines’ day to day wear and tear.

The technology was created by researchers at Nottingham Trent University, led by Prof Amin Al-Habaibeh, professor of intelligent engineering systems.

Primary applications include use within the milling and drilling processes employed in car engine manufacture, the aerospace industry and metal cutting.

The system, developed by PhD researcher Milad Elgargni under the supervision of Prof Al-Habaibeh and Professor Ahmad Lotfi, of the university’s school of science and technology, uses a combination of infrared cameras and artificial neural networks to consistently detect when cutting tools are broken or missing.

The technology does not require any contact with the manufacturing machinery, and could provide live feedback via computer to alert operators in order to help prevent catastrophic tool damage.

It can detect problems in real time, which is difficult to achieve by common methods, while the artificial intelligence system can learn, making it possible to monitor various cutting tools and increasing the system’s flexibility for users.

‘Existing monitoring systems perhaps only look at the power of the spindle to see if a problem is occurring,’ said Professor Amin Al-Habaibeh. ‘However, there may be a collision if the tool is broken. Our system uses infrared to make sure the tool is working in the right place and under the right conditions at all times. We are also shortly to publish research about the use of the technology to monitor gradual wear.’

As the technology is based on using a simple infrared camera, it should be easy for manufacturers to put it in place without disrupting existing machinery.