Artificial intelligence is transforming the way Domtar’s sawmills operate, augmenting human expertise and completing routine tasks through machine learning. With AI modeling, we’re finding new ways to improve quality and efficiency.
The wood manufacturing industry is highly efficient. Every part of every tree we harvest can be made into useful products. Whatever can’t be used as lumber is chipped to become pulp and paper, and other residues, like bark, shavings and sawdust, are used to generate energy as biomass.
Domtar has identified ways to leverage AI modeling and machine learning to significantly improve product quality over the past two years at our mills in Glenwood, Arkansas; Cross City, Florida; Thunder Bay and Atikokan in Ontario; and Normandin, La Doré and Senneterre in Quebec. We can also use AI modeling to assess wood for compliance with customer specifications.
“AI allows us to be more precise and consistent,” says Carl Lévêque, quality superintendent for all planer mills in the company’s Wood Products business unit. “There’s a lot less waste.”
AI Enhances the Sawmilling Process
Soon after arriving at the planer mill, boards are scanned by sensors and photographed from multiple angles. The scanner analyzes the wood’s composition, grading it visually and identifying inconsistencies that are undetectable to the human eye, such as separations, discoloration or holes created by insects after forest fires. This was once a labor-intensive and potentially hazardous task that was prone to human error. Now, it takes seconds and is highly accurate.
Programmers create AI models to assess wood for each individual specification at a level of detail that would be impossible for people to attain.
“The Normandin mill makes around 15 unique lumber products, and each one needs to meet customer specifications,” explains Carl. “Wood that will be used to build homes is subject to strict building code standards, and customers may have additional specifications.”
Using AI modeling ensures the same high standards of quality are applied consistently to every product that leaves the mill.
AI Modeling Complements Human Ingenuity
With human guidance, the machines are trained over time to become more precise. This ensures every board is sorted in a way that optimizes its value. But that doesn’t mean the process is fully automated. Effective decision making still requires human expertise to support AI. Process engineers and technicians train the AI by tagging pixels of the defect that we want to be able to identify.
Our quality and optimization supervisors perform regular field tests on each AI model to ensure the machine’s assessments are aligned with their specifications. They constantly evaluate and analyze data to keep the technology up to date. Sometimes, an AI model identifies an irregularity that is not actually a defect, erroneously flagging a board as unfit for a particular purpose. When this happens, mill employees work with the machine supplier to adjust the model, reducing waste.
Wood is a plentiful, renewable natural resource. By combining the use of AI modeling and other emerging technologies with human ingenuity and expertise, we can extend the value of each tree, and the value and consistency of our products, more than ever before.
