A new white paper from DBI demonstrates how infrared machine vision and […]
A new white paper from DBI demonstrates how infrared machine vision and data processing uncover hidden variations in insulation materials during fire-exposure testing – and points to a method that could transform testing of other construction products.
When insulation manufacturers fail a full-scale fire test, they often lack a clear understanding of why. A single thermocouple reading on the unexposed side may cause a failure, while revealing little about what happened inside the material. According to Martin Sturdy, Research Consultant at DBI’s Advanced Fire Testing division, this lack of insight is one of the challenges Infrared Machine Vision (IMV) is designed to address.
– Manufacturers come to us because they need to prepare for a fire test or have trouble passing one, and the process helps them learn more about their material – not just whether it passed or failed. Thermocouples provide valuable readings, but only locally. They don’t always reflect the full behaviour of the material, says Martin Sturdy.
DBI’s new white paper demonstrates how IMV, combined with advanced data processing, can reveal differences in product homogeneousness that traditional measurements may overlook.
Understanding homogeneousness through variance
Thermocouples remain essential in both small- and full-scale fire testing, but they capture temperature at only a few fixed points. This means they cannot detect variations that occur outside those locations. A thermocouple might sit on a compressed or wrinkled region that heats up faster than the surrounding material, while in another test it might land on an undisturbed area. Both readings comply with the standard, yet they tell different stories.
IMV supplements these point measurements with full-surface infrared imaging. During testing, the camera captures an image every five seconds, generating thousands of infrared images. Each image is stored as a data matrix containing pixel-based temperature values. DBI’s processing pipeline converts these matrices into temperature maps and histograms, ultimately producing a numerical variance value that reflects how uniformly the material behaves.
A low variance indicates a more homogeneous product, while a high variance highlights irregularities that affect fire performance.
– Anyone can point an infrared camera at a test and see a hotspot. The real value comes afterwards, when we provide objective measurement and process the data. That’s where we can give manufacturers a number that reflects how uniformly their material behaves – something objective they can use to compare samples, says Martin Sturdy.
What the white paper shows
The study evaluates ceramic wool, mineral wool slabs and mineral wool rolls under identical heating conditions. Although intended for similar applications, the materials behaved differently. In this test series, the ceramic wool sample – which had a lower nominal thickness than the two mineral wool samples – nonetheless displayed a highly uniform temperature distribution with low variance. Mineral wool slabs also performed uniformly, reflecting the stability of their production method.
Mineral wool rolls, by contrast, showed significantly higher variance caused by wrinkles and compression introduced during rolling and packaging. These irregularities created localised hotspots that would not necessarily be captured by thermocouples alone.
– Once we added IMV to our small-scale tests, it became clear that the material wasn’t always as homogeneous as expected. The rolls in particular had areas that consistently heated up faster, and that explained why thermocouple readings differed so much from test to test, says Martin Sturdy.
Preparing for full-scale tests
While some producers approach DBI after a failed test, others now use IMV earlier in their development process. Screening materials before booking a full-scale test gives them a better understanding of their product’s behaviour and reduces the risk of unexpected outcomes.
– Producers can come to us before the full-scale test. If the variance is high, they know the material is not ready for testing. If it’s low, they can go in with confidence. It’s not just about explaining a failure – it’s about preventing one, says Martin Sturdy.
He highlights a recent example: A manufacturer who had failed several full-scale tests finally passed after IMV revealed inhomogeneous regions that had previously gone unnoticed.
Instead of spending test after test and wondering why something went wrong, they can do the homework first and come better prepared for the big test, says Martin Sturdy.
A tool with wider applications
Although the current study focuses on insulation products, IMV is not limited to mineral wool. Other construction materials – such as gypsum boards or bio-based materials like hempcrete – could also benefit from IMV if manufacturers want deeper insight into how their material behaves during fire exposure. According to Sturdy, the same IMV-based testing approach can be applied to these materials as well.
– If a producer wants to measure something else than variance, we can also do that. The technology is flexible because it captures millions of data points. What we do with those data depends on what they need to learn about their material, says Martin Sturdy.
This flexibility positions IMV as a promising tool for future fire-testing challenges across a much broader range of materials and applications.
Read about small-scale tests, and download the white paper ‘Revealing the hidden fire performance of insulation’
What IMV and data processing can show
Infrared Machine Vision (IMV)
DBI’s data processing
Key benefits of IMV screening
Menu
Members
Knowledge