Vai al contenuto

thermography

Technologies and techniques for detecting product defects through thermographic analysis, real-time measurements, data analysis, computer vision, artificial intelligence and machine learning

Subjects involved

Dumarey Group (manufacturing company), AITEM S.r.l. (solution provider). The use case was promoted by the CIM 4.0 Competence Center.

Benefits

Improved product quality (identification of over 40% of defective parts), cost reduction (potential annual savings of €250,000), maintenance optimisation (higher productivity and profitability thanks to a proactive and predictive approach to maintenance)

The Challenge

The Dumarey Group develops and supplies engines and transmissions worldwide, energy storage systems, control systems, sub-assemblies and components, with proven integration capabilities for small, medium and large-scale applications. Thanks to its flexible, partnership-oriented approach and its established experience as an OEM, it ensures the reliability of its products both in production and in use. Dumarey is committed to promoting sustainable mobility for all, developing hydrogen and electric technologies and supporting the circular economy through reuse, remanufacturing and recycling.

AITEM was founded in 2020 as a start-up, more precisely a spin-off of the Dumarey Group, and develops artificial intelligence-based software using cutting-edge, innovative technologies. What makes it unique is the combination of technological expertise, market diversification and a strategic business approach, with the speed of a start-up and the experience of a multinational.

The problem highlighted by Dumarey lay in its high-pressure die-casting process, used in the production of gearboxes in aluminium, which could generate defects that often could not be detected until a later machining stage, usually outsourced to third parties.
Even though parts were scrapped after machining because of these defects, the work carried out still had to be paid for.
This failure to identify defects in advance, together with the customer’s/supplier’s freedom to define what counted as a defect, led to economic losses and a reduction in production capacity.

The Solution

The system developed by AITEM and adopted by Dumarey can analyse data from:

  • the use of a thermal camera
  • specific inline measurements
  • the collection and analysis of production process data,

and by applying:

  • computer vision techniques
  • machine learning,

it can accurately predict the presence of defects in components before they are sent to third parties. Rejecting defective parts avoids the need for external machining and makes it possible to put them back into the production cycle.

The results and benefits achieved are particularly evident in the following areas:

  • Improved product quality (identification of over 40% of defective parts),
  • Cost reduction (potential annual savings of €250,000),
  • Maintenance optimisation (higher productivity and profitability thanks to a proactive and predictive approach to maintenance,

Are you interested in adopting this methodology?