Skip to main content

Working smarter through automation: Wood and fiber quality research team develops new way to analyze wood

The smartest woodworking tool at the University of Georgia lives in a brick building at the edge of Whitehall Forest.

Here, amid machines that test wood for its stiffness and strength, sits a router that automatically determines the size and depth of a wood disk, and then adjust its cuts to match the disk dimensions. Place a wood disk on its deck and the machine knows what to do, as it hovers, calculates, whirrs and cuts sleek, smooth lines across the wood’s surface.

When it’s done, you’re left with a disk of wood that’s machined smooth, making the tree rings easy to discern. This detail is key, because for as smart as this computer numerical control router is, it has no idea how important those rings are in its larger purpose: To help determine a quicker method of measuring the yearly growth rate and the volume of wood in a southern pine tree.

“Now we have wood disks that are clean and can be imaged,” says Joe Dahlen, associate professor of wood quality and forest products at the University of Georgia Warnell School of Forestry and Natural Resources. “We’ve sampled 400 trees, and we’re looking at determining the ring-width information from the disks. Given the clean disks, let’s see how much more information we can get out of them.”

The router’s smooth cuts were necessary because when trees are harvested and cut the traditional way, with a chain saw, it leaves a rough edge that obscures the tree’s rings. Because of the green wood’s sap, sanding it just becomes a sticky mess. And when the wood disks are dried, they crack, which ruins the sample.

So, Dahlen and his team took a different approach. They started with a basic computer numerical control router, then built and programmed the router to automatically determine the shape for each piece of wood laid upon its deck. The program adjusts the movements for the router itself, using a bit to cut across the wood in wide, even swaths.

Once the clean cuts are made, the wood disks are taken to a specialized photo studio of sorts—also built by Dahlen and his research assistants—where they take images of each machined piece using a variety of white and blue lights to illuminate unique features of the wood.

Developing the router to quickly machine green disks was a breakthrough for the team, which was struggling with a way to reliably capture images of the samples they collect for their research. Dahlen is optimistic that the imaging techniques they’re developing will replace the current way to measure wood volume using water displacement— where wood and bark pieces sit in a tub until they’re saturated, then they’re gently inserted into a small water pool to measure their displacement.

While the process of cutting and imaging the disks may take about the same amount of time as the traditional water displacement method to determine volume, Dahlen says they’ll end up with far more information using the new method. The team is now working to develop a computer model that can be used to determine a variety of factors.

“The first is actual wood volume and bark volume,” says Sameen Raut, a master’s student who has been instrumental in developing the process and computer model. “The traditional way is time-consuming and from images we’ll have equally accurate information.”

But not all trees produce straight and stiff lumber; if a tree grows unevenly, the tree produces “compression wood,” which warps during drying and has low stiffness.

“Being able to quantify how much compression wood is important because we haven’t been able to quantify that before,” adds Raut, “And while a tree may seem straight, we are still detecting a decent amount of compression wood within these straight, defect-free trees.”

Raut says he also wants to quantify the rings within the disks in order to measure growth rate over time.

All of these calculations tie into a larger project that Dahlen’s team is tackling, which is looking at the performance of longleaf pine trees grown in plantations. But once the algorithms are tested and refined, they will have a model that can then be developed for any kind of tree, says Dahlen.

“For a long time, I wanted to capture images of disks—we didn’t have a visual record of the disks we were working with,” he says. “It took a lot of design, engineering and programming to come up with this. We’ll be able to use the photos from the disks to gauge how well our wood quality predictive models are working, and we can use the information to better predict product quality.”

 

Associated Personnel:
Article Type:
Video: