Hybrid grade scanning system for green mill application
MILLS THE world over are enjoying brisk markets, and scrambling to stay ahead of the curve with improved processes. Pukepine Sawmills is one of those operations that is proceeding with vigorous plans to update its technology and improve its throughput and value recovery. With several projects under way, it selected USNR to install a unique green grading line that is beating targets for return on investment.
Pukepine’s management didn't mind being the first to try a new concept. They understood that it can sometimes take time to achieve full results from a new design, but believed the benefits would prove its worth. They have a strong history with USNR and were confident they would receive good value for their investment. With several months run time under their belt, the results are exceeding their expectations.
Pukepine Sawmills, located at Te Puke, New Zealand, is owned and operated by the Tanner family; Ken Tanner is CEO, joined by his sons Jeff (general manager) and Paul (engineering and maintenance). The operation comprises green and dry mills, lumber kilns, a wood treating plant, a secondary processing plant that produces finger-joint and moulding materials, and a paint plant. Pukepine employs 120 on a single shift.
The sawmill produces 50% structural timber products for the New Zealand and Australian markets, and 50% high quality appearance-grade products for New Zealand, Australia, North America and all of Europe. The mill’s capacity is 100,000m3 annually, up from 65,000m3 with recent improvements.
David Sandford is a consultant who works with the Tanners, assisting and advising on day-today operations as well as management of capital investment projects. As such, he worked closely with Jeff and Paul Tanner to manage this project. He said, “We selected USNR for this project based on a good working relationship in the past with previous projects, and we were keen to continue building that relationship.
“We hoped to have a oneyear payback on the project based on conversion and grade recovery. We haven’t gone a full year yet, but we are on course to exceed those expectations.”
Partnering with USNR
The sawmill’s headrig carriage utilizes USNR’s MillExpert optimization. The Newnes-McGehee (now USNR) gang and edger were previously relocated from a North American mill, and employ the Newnes Sawmill Suite (NSS) software platform. Trimming comprised a manual grading, valley trimmer line that was very limited on capacity, with room for only one grader on the line. Constrained by both trim line layout and technical capability, David said, “Our goal was to increase capacity and improve grade recovery.”
The project starts with a lugged transfer (Pukepine supplied) feeding the precision belted scanner transfer. This trimmer optimization system is comprised of a new scan frame fitted with 22 BioLuma™ 2900LVG+ sensors.
Next in the line, a new Multi-Track Fence utilizes 13 paddles to position lumber on the positioning transfer for trimming. From here, the equipment was supplied by Pukepine. Positioned lumber is fed to the multi-saw trimmer, and select trimmed boards are rerouted at the smart tipple gate.
There are several aspects of this automated grading line that are unique to Pukepine. Firstly, this configuration is atypical from the standard green grading systems in that the sensors are mounted off-axis to provide 4-sided scanning with only 2 rows of sensors (1 each, top and bottom). This configuration is typical for the Transverse High Grader (THG) system used for automated grading in dry, planed applications. Most green grading BioVision lines require only top and bottom faces to be scanned, where the sensors are mounted above and below in a true differential scanning arrangement. With the types of defects encountered in Radiata pine, and the grades that Pukepine’s products specified, the 4-sided scanning arrangement was required.
Radiata pine has some properties that are not common in North American species, such as needle fleck (bird’s eye), that required additional optimization coding to detect and correctly classify these characteristics for the products Pukepine is producing.
For this application, Pukepine selected BioLuma 2900LVG+ sensors that collect and integrate high resolution laser profile data (0.3”/8mm x 2500 Hz), high density color vision data (0.01”/0.25mm), as well as GrainMap™ data (0.3”/8mm x 500 Hz) from all 4 surfaces of the board. GrainMap grain angle measurement provides more comprehensive defect analysis, and allows the optimizer to detect and classify defects that do not offer significant visual contrast. This system provides 3D modeling with far more detail than any other scanning system on the market today.
Another aspect of this green grading system that is unique is the utilization of an end grain scanner that is commonly used with the THG for dry mill grade scanning. A BioLuma 2900V vision sensor is mounted in the lumber line end of the scan frame where its field of view is horizontal, and it scans the ends of the boards as they pass by. The end grain scanner provides valuable data by locating the board relative to the pith in its “parental” log. The optimizer uses this information about pith location in determining knot projection and connectivity within the piece.
Because this is a green line, a cleanup saw is positioned ahead of the scanner to present a freshly sawn, square end of the board for end grain scanning.
This green grading system is based on USNR’s THG. The hardware and scanning technologies required are the same, whereas the primary difference is in the software that is configured to detect and classify characteristics is green vs. dry, and in un-planed vs. planed wood.
When grading green vs. dry wood, some defects change color once they are dried, and this can positively or negatively impact the capability of the grading system to accurately detect and classify those defects. As well, planed lumber allows more subtle use of the system’s GrainMap feature, as the surface is smoother and more uniform. Planing also creates a cleaner surface free from most grease and machining marks, making it easier for the system’s range of technologies to detect minute variations in the wood.
Facing these challenges means making adjustments in the code and training the system to accurately recognize those variations, then correctly classify them within the parameters of specified products. Training the system equates to teaching it to classify each characteristic based on accumulated data that has been fed to the optimizer through scanning a large quantity of boards. Training sets are created for each defect type and for each species that is to be run. The classifier retains this database of characteristics, and compares the scan data to each of these as it develops a solution.
For some characteristics such as stain, the system is tuned to identify areas of interest that are then passed to the classifier to compare the data against its database and correctly classify that particular feature.
Another challenge that is overcome with experience is the natural variations that occur with different species of wood. While Radiata Pine is similar in its properties to Southern Yellow Pine, there tends to be less contrast between knots and the sorrounding wood fibre in Radiate Pine. This demands the utmost in capability from the grading system. It requires minute GrainMap measurements, ultra-high resolution color imaging, and highly dense end grain data combined through DataFusion® to compare and verify all the data inputs before a solution is generated.
The mill’s upgraded its existing NSS transverse cant optimization system to version 7.2, and the operating system to Windows 7. This addressed any obsolescence concerns from the previous version which dated back over 15 years. The new system also offers much more computing power. David commented, “This version is a lot more robust; we’re very happy with it.”
Regarding the new Multi-Track Fence, David related that the mill wanted to be prepared for future increases in speed and throughput. “The Multi-Track Fence is probably overkill for our nees, but we were thinking ahead. We're very satisfied
with the fence’s operation.”
Better than expected
With the new grading line operating for several months, David says, “It is exceeding our expectations.” He didn’t want to disclose any specifics other than to say the projects it will take Pukepine about a year to determine exactly how far the system has exceeded their goals.
He went on to relate that, over and above the increase in value recovery the system is providing, it now sets the stage for further capital improvements at the site. “This project has been based on recovery and grade improvement. And now it also allows us to increase production considerably. We’re in the middle of adding more kiln and boiler capacity, and that allows the sawmill to produce more.”
David was pleased with the way the project progressed overall. He said that in spite of the short timeline, USNR met the timeline and the project was installed during the mill’s 2013 Christmas break. “The grading system commissioned up very well over about 5 months with this being the first 4-sided.
green scanning system; the technicians did a good job!”
David gives much credit to Stephen Wardrop, USNR’s technician who started up and commissioned the BioVision system, and to William Cleland, Pukepine’s optimization technician who works on all the mill’s optimization systems. He said that William traveled to USNR’s Salmon Arm, BC facility for a week’s training on the system prior to installation, then worked closely with Stephen during the start-up and commissioning phases. Other USNR personnel who worked closely with Pukepine’s team were Nadim Karmali, project manager, Steve Woods and Natalie Peles, software designers, and Shayne Cameron, service technician.
He also gave credit to Pukepine’s in-house electricians, maintenance team, engineers and construction crew who handled the disassembly of the old line and installation of the new equipment. And he related that the trim line's configuration and layout had to be reworked significantly to accommodate the new processes. He also gave credit to Tui Technology that handled the controls for the project.
With four new kilns and boiler upgrades currently in the works, the Pukepine team is looking ahead with anticipation of continued strong markets. With the investments they’ve already put into place Pukepine is ready to take on fresh challenges.