University research aimed at improving sawmill instrumentation

ORONO — Sawmill instrumentation isn’t usually considered high tech; after all, what does it take to cut lumber? Spin the saws, grade the lumber, sell it to a customer.

Not so in today’s competitive market for sawn lumber.

Bruce Segee, an assistant professor of electrical and computer engineering at the University of Maine, has researched sawmill instrumentation. One project he completed four years remains in operation, and he and other UM researchers are developing high-tech instruments that will be easy for regular maintenance people to troubleshoot.

Counting lumber with retired computers and an encoder

International Paper owned a sawmill in Madison, N.H., a mill described by Segee as “fairly low tech. They really had no idea how much lumber they were producing.”

Mill employees tracked daily production by counting the bundles of lumber, not the individual boards within the bundles. Employees annually selected some bundles “as representative of what they usually produced” and measured every board in those bundles, Segee said. This information was used to calculate how much lumber the mill was probably producing.

IP decided to install a counting system to measure individual boards during production, as well as the board dimensions available in each bundle. Segee was working at the University of New Hampshire when IP asked him to help develop a new counting system.

“The ultimate goal was to create an autonomous system and to run it in a nasty environment,” Segee said. “A sawmill is a dusty, noisy place, hard on delicate sensors and computers. We had to figure out a way to count the lumber without having the (sensory) system breaking down all the time.”

Segee was still researching the system when he came to UM in summer 1992. The system became operational in November 1992, and “it’s been running ever since,” he said.

“Cost was a factor, so the project was designed to be simple,” according to Segee. He and his graduate students created a system that uses “retired” IP office computers and eight optic sensors.

The computers were equipped with an insulated power source and were placed around the mill. The sites are a programmer’s nightmare, Segee hinted, because the “computers aren’t where it’s air-conditioned, and it’s not exactly heated, either.” He explained the computers were heated with 100-watt light bulbs in the winter and cooled by electric fans in the summer.

The sensors “look down on the moving boards and pick up a reflection when a board is going by,” Segee said. “There is no reflection when a board isn’t there.”

A sensor is located every 2 feet. Each time a sensor detects a reflection, it “trips.” The system determines a board’s length by the number of sensors that trip consecutively.

A rotary encoder fastened to a chain axle on the counting table calculates the width of each board. The spinning axle moves the chains that move lumber on the table. With each axle revolution, the encoder sends 500 electric pulses to the computer, which determines how far the chain has moved.

The mill was cutting lumber to a set thickness, according to Segee. “If I know how far the chains moved while the sensors were tripped, I can calculate how wide the board is,” he said.

Power-line instrumentation

That research showed Segee “what sawmill operators are looking for in such a system, one they wouldn’t have to hire a computer programmer or engineer to run.”

Hiring a technical troubleshooter to maintain high-tech equipment is a big expense for sawmill owners, Segee stated. What if alternative instrumentation, one easier for electricians and millwrights to fix, could be developed?

He and Steve Shaler, an assistant professor of wood sciences and technology at UM, obtained a USDA grant to research power-line instrumentation. Power-line refers to the 110-volt house wiring, not the incoming electrical supply.

“Our project’s purpose was to address a real need in the less technology-driven industries, where there is a need for, as well as a fear of, technology. The need could be as simple as counting the number of boards, but sawmills do not always have a computer tech on staff,” Segee said. “Why not come up with the electronics and instrumentation a mechanic or electrician could use?”

The project’s goal is to develop instrumentation that can be plugged into the operational process anywhere in a sawmill. An instrument could be a rotary encoder, a thermometer, a photosensor, an actuator, a humidity sensor: “Instrumentation is the easy part,” Segee said. “It’s limited only by your imagination.”

The instruments would be able to “talk” to each other or to a computer, he explained, and would not require special connectors or coaxial cable. The instruments would use a small voltage (less than 10 volts) as a “frequency” by which to communicate.

An electric screen would prevent other electrical frequences from erasing the instrument-collected data, Segee said. The instruments would work on a 60-hertz electric current, using a higher-frequency signal from that current for communications and a lower-frequency signal for power transmission.

A computer and its associated software are needed “to use the data. Once you get the instrument in place, it can generate all the data it wants, but you need to know what to do with it,” Segee said.

He believes this system would require less maintenance than a complicated sensor system employing computerized sensors. “If you have power, and the instrument isn’t working, you need to plug in another instrument,” Segee said.

Under the one-year grant, researchers have already developed a system in use at Barrows Hall. The instruments are talking to each other and computers, Segee reported, but the system has not yet been baptized in a sawmill.

“We don’t know how it will work in an industrial environment. That’s what we need to learn next,” Segee said.