Author Topic: Using Near-Infrared Detection to Determine Moisture Content  (Read 3926 times)

Offline StavrosA

  • Forum Expert
  • Newbie
  • *****
  • Posts: 5
    • View Profile
Using Near-Infrared Detection to Determine Moisture Content
« on: September 16, 2014, 08:54:42 AM »
Near-infrared detection holds great promise for determining moisture content, species, and density of wood in mill operations. Not only that, it can be done fairly quickly, speeding production on the line.

Hi, I'm Stavros Avramidis, a Professor at the University of British Columbia, Canada's, Department of Wood Science, and a Kilndrying.org expert, and I've found that when it comes to such factors as Moisture Content (MC), species type, and wood density in mill operations, high variation 'in' leads to high variation 'out'.

So what exactly am I talking about? Without presorting prior to kiln drying, you introduce a high variation of species, density or MC, leading to a high variation in product output. This is really an issue throughout the whole industry.

As you probably already know, not all species are created equal. Different species have different drying characteristics. If you don't know the species difference, then the sort can be based on either density or MC.

But telling the difference when pre-sorting in the forest can be very expensive. If you are in the mill and can't tell the species difference, then you can use MC instead.

Tight export standards can pose a problem, however. Typically, exports can have a variance of +1.5%, so if you input wood with a high variance in MC, you will get wood out with an equally high variance of MC and the whole shipment will be returned because it doesn't meet the export quality standards.

What is the solution to this thorny problem? Near-infrared spectroscopy. After extensive study on this topic, we've found that near-infrared spectroscopy in the 1300-2100 nm range of the spectrum coupled with multivariate statistical models provides a way to rapidly and accurately predict the surface moisture content of wood at line speeds of from 100 mm/s up to 1,000 mm/s, and of determining entire surface moisture distribution rapidly, as well as detecting wet-pockets in lamina for industry applications. The speed of the process also avoids problems caused by debris flying around inside the mill. Near-infrared technology allows a mill to determine MC, density, and species immediately in the mill, a great time-saver that can translate into increased productivity as well as increased profits.

What are your thoughts on this technique? Let us know at Kilndrying.org.



References

Watanabe, Mansfield, Avramidis (2012) Detection of wet-pockets on the surface of Tsuga heterophylla (Raf.) Sarg. by near infrared (NIR) spectroscopy

Watanabe, Mansfield, Avramidis (2011) Application of NIR spectroscopy for moisture-based sorting of green hem-fir timber

Watanabe et al (2010) Wet-pocket classification in Abies lasiocarpa using spectroscopy
in the visible and near infrared range

Offline MichaelM

  • Forum Expert
  • Kiln Drying Enthusiest
  • *****
  • Posts: 94
    • View Profile
Re: Using Near-Infrared Detection to Determine Moisture Content
« Reply #1 on: November 03, 2015, 12:37:17 PM »
We did this about 15 years ago and it does work great as a moisture meter.  It is used this way in particleboard mills and on chip conveyors where a surface MC is ok.  I think it is even used for veneer sorting.  Atnikov, in the Western Dry Kiln Association Proceedings had a paper on using IR (not NIR) for a moisture meter.

Our objective in the past work was to use NIR on the wood prior to drying to predict the pieces that would be wet after drying.  This would allow one to sort not based on green MC or green density, but and estimate of the final MC.  The hypothesis was that there was a chemical difference (due to extractives, or other compounds) that caused pieces to dry differently.  We were trying to detect the chemical signature with the NIR, not really the water signature.

Well, obviously we were not successful and I don't think we published the results.  One reason was that the water signature was huge and tended to hide the other chemical information.  There's a part of me that still believes this technique should work.  However, I agree with Stavros, it would make a great meter for surface MC. 

 


Search


* Forum Experts

MichaelM MichaelM
Forum Expert
TimothyD TimothyD
Forum Expert
PhilM PhilM
Forum Expert
StavrosA StavrosA
Forum Expert

* Subscribe to Newsletter

\
SimplePortal 2.3.6 © 2008-2014, SimplePortal