The rationale for the Charles Ingram mill to turn over the grading process to an automated grading machine was to eliminate the speed and grading limitations attributed to human grading. But an additional requirement critical to success was the expectation that no matter what method was chosen, Ingram’s mill would be capable of generating more value-added products. Stated in a different way, Ingram wanted to be able to end up with a system that would allow them to make frequent changes at the planer mill so they could quickly respond to sales opportunities that would generate more profit.
Human graders are amazing in that they can make complex decisions very quickly and accurately. Through repetitive experience, a person’s brain can be trained to build up efficient neural pathways that allow them to process what they see, hear, and feel because they develop shortcuts. However, performance takes a hit when a human is suddenly asked to make dramatic changes to incorporate a different set of rules. For example, the results of human graders are not so good right after a planer mill changes from a Dimension to an Appearance grade, or if the supervisor asks graders to increase the number of 8 foot #1 boards to fill an order, but not to sacrifice too many long boards to get them. “Oh, and make sure you don’t make too many cut-in-twos in a row because our production volume will drop too much.”
Charles Ingram produces about 100MMBF annually at its single site facility in Effingham, South Carolina. The bulk of its products include Dimension grades of Southern Yellow Pine as well as multiple appearance grades, with nominal sizes ranging from 1x4s and 1x6s, 5/4” radius-edged decking, 2x4s to 2x10s, 4x4s and 4x6s, plus 6x6 Timbers. Its planer line required three or four people to keep up with the production volume because graders needed at least two to three seconds on any one board to maintain a quality output.
There was an opportunity to serve a higher-value market, but that meant meeting more exacting customer expectations. So not only were the graders being asked to make more important decisions in less time, but the frequent changes to satisfy special orders required graders to quickly “reprogram” themselves.
The mill realized that it was using scanners and computers at almost every step of the breakdown process in the sawmill to get the best value out of the fiber, but depended on real-time human decisions for determination of the value for each board once the lumber reached the planer mill. That meant that even though they had some of the latest technologies at work during the green sawing process, any additional gains they generated were still restricted by the inefficiencies in the planer mill. It was apparent that automating the grading process was the solution.
The mill put together a team representing management and quality control, and set off to select a scanning system supplier. Their intention was not to pick the fastest or simplest scanner. Instead, they wanted to focus on accuracy. They believed that if the accuracy was there, the obvious result would be increased value from their fiber. They stressed that in choosing a scanning system they would place a lot of emphasis is on how the system makes money and that they would rely heavily on input from their quality control people for input on grading results, flexibility for future changes, and ease-of-use.
Charles Ingram’s own lumber was graded by hand and then sent to both lineal and transverse scanner suppliers for evaluation. Several days were spent visiting each supplier’s factory or running the lumber through a production scanner in a mill. This was a challenging undertaking, as each supplier’s situation was different. For example, one supplier fed boards through their scanner by hand, while another had a large facility with material handling equipment to automate the test process.
Although Ingram was very interested in calculating a return on investment so they could justify the initial expense of a scanner project, the primary emphasis for the evaluation was to develop a good understanding of what could be seen by a scanner. The expectation was that if it could identify enough of the fiber correctly to take over the grading process, it would make them money. Therefore, instead of shipping a batch of mill-run lumber and comparing grades between man and machine, approximately 200 boards were carefully selected to represent fairly common but difficult characteristics that were typical of a production run. Included in this lumber were some “easier” high-grade boards, too. Worm-eaten pitch and blonde knots were present on the more difficult boards as the occurrence of these was expected to increase in the future as more plantation fiber would be the norm. In case accuracy might vary over a range of sizes, narrow and wide Dimension samples were included and a 5/4 Decking sample represented appearance products.
Each board was given a unique number and its characteristics recorded, noting any differences between defect identification and sizing between man and machine. For example, various knots, shake, and wane areas were measured with calipers and steel tape with discrepancies noted. Although not the goal for this testing, the scanner’s optimized solution for each board was helpful for identifying critical areas to investigate. Answering the question, “Why didn’t this board make a #1”, was a typical step in flagging potential problem areas since the answer might lead to a knot displacement or light skip defect that might have been measured or identified differently between the scanner and the Ingram team.
After doing the rounds the first time, Ingram’s purchase decision was delayed a year because they decided that scanners were close but not quite good enough for them, and the depressed market increased the perceived risk. This delay forced a reconsideration of the testing process – with the result being a significant change that Ingram felt could not have been foreseen without first going through the initial effort. And the key factors driving the adjustment were “weathered” wood and training.
A production scanning system located after a planer expects to see lumber that is bright, clean, and at production size. The fiber analysis depends on some consistency, or it cannot identify and locate fiber characteristics accurately. By the time the lumber had been shipped multiple times and exposed to various handling actions, it was dirty, greasy, damaged, and had shrunk so badly that each successive supplier had an increasingly more difficult time adjusting their system to ignore weathering that would not be present in a real application.
Another realization was that the same log sent through two different mills does produce a different looking board – and there are regional differences between where a tree is grown that show up after planing. For example, consider the sawmill process – there are always differences in how a mill effects the conversion from log to board. Just a few of the differences that show up on a planed board originate from the debarker leaving bark and related gouges, sawing patterns and log sizes changing the kinds of knots, drying damage, handling differences causing marks, direct-fired kilns leaving black areas on unsurfaced sections, and planer knife patterns shaping the surface. These differences between mills require that a scanner be customized for each site. Therefore, the lumber Ingram was packing around was going to be seen differently at each scanner supplier, depending on how similar their experiences matched the Ingram boards. That didn’t level the playing field for the suppliers and the results were even more skewed by how much effort a supplier could/would put into customized training for the Ingram wood.
So the big shift Ingram made was to leave its own lumber home and visit mills to see a potential supplier’s scanner in production, analyzing how well it did on lumber that it was already trained for. Assuming they could be trained for Ingram’s situation, potential scanners were evaluated in the mills, with the team spending sometimes several hours looking closely at the solutions to learn the strengths and weaknesses.
In the end, they chose Lucidyne’s GradeScan system. According to the selection team, there were two primary reasons driving their decision, with the first being the stated goal of accuracy. The second was the Lucidyne staff. The selection team’s overriding belief was, “Despite the higher cost of the Lucidyne scanner, its accuracy was excellent and its people were phenomenal. The Lucidyne engineers were enthusiastic about doing the very best they could at grading each board.”
The largest part of the customizing necessary to prepare the GradeScan® system for Ingram’s specific production requirements and fiber required analyzing Ingram’s wood.
Ingram sent a couple-hundred boards to Lucidyne’s home in Corvallis, Oregon, for training. The lumber was scanned through Lucidyne’s in-house GradeScan and the results reviewed by both companies over the period of a few days. Lucidyne’s staff made adjustments to its scanner image processing program to suit Ingram’s needs. This was also the ideal time to identify any optimization-related modifications necessary for customization. Lucidyne worked with Ingram to develop sets of grade rules that would satisfy the mill’s requirements for standard industry grades as well as proprietary “house” grades.
Once Ingram’s new GradeScan was built, a second batch of lumber was sent through the scanner as part of a training exercise for Ingram personnel, and as an acceptance test. It was important to both companies that once the scanner was installed that it work immediately. Ingram sent a team that included technical and management personnel and all participated in learning how to do everything from calibrating the scanner, making grade rule changes, operating the system on a day-to-day basis, and basic troubleshooting. The Ingram team learned that although the system was simple to operate, there was a lot of flexibility and power accessible just underneath the surface that they could tap for fine tuning and future opportunities. Management’s representation at the training wasn’t merely ceremonial; it was understood that the scanner would be making decisions that directly affected Ingram’s bottom line. So the more their top managers knew about its capabilities and what it could do on a day-to-day basis would be critical to their directing of the entire Charles Ingram operation.
Lucidyne’s lineal GradeScan system was installed in line with Ingram’s planer in March of 2012. The sorter control system was installed two weeks before by Lucidyne and utilized Rockwell’s ControlLogix. Then the mill shut down on Friday morning for the mechanical changeover. The scanner mechanical and planer outfeed conveyor changes were started on Saturday, electrical on Sunday, and testing began Monday morning. The scanner was running its first production run Monday evening. They had an ALS inspection three weeks after start-up and the inspector said that it started up better than any he had seen. He also said that it looked like the mill already had it dialed in. But from the mill’s view, they were just getting started because there still were lots of ways to tune the system to do even better.
The scanner itself is located about 16 feet from the planer and fed by a rollcase, which places it outside the planer room and eliminates a lot of debris. Most of the mechanical changes to downstream equipment were done during normal downtime and tested with the trimmer/sorter system startup to work out any kinks, especially the new trimmer infinite fence.
Lucidyne’s new Warp Tunnel was installed after the scanner on the planer outfeed belt to measure warp when boards are relaxed and not being confined by rolls or fencing. The mill staff said that they have been surprised a few times when a board or Timber looked like a top grade but was downgraded for warp. In fact, it wasn’t obvious by just looking at them until they laid them on the floor and actually ran a string to measure the bow, crook, and twist. One comment was, “The warp measurement is more accurate than a human, in fact it is too accurate. It can measure warp that a person can’t see.”
Ingram had Lucidyne upgrade its Grade Mark Reader (GMR) and uses it on reentry lumber since that material does not pass through the scanner. The GMR is used far less than before, in GradeScan system has some optional features that make it possible to reduce the number of these incomplete packages so they don’t end up displacing valuable production time later.
A “Check Grader” monitors material flow between the planer and the sorter to watch for problems with lumber handling. They also will manually grade lumber that is reentered so they must be qualified to grade. A large color monitor mounted near the grading area shows the status and performance of the system. The Check Grader’s role is important due to the nature of managing wood flow and they can be a “first responder” if any problems to arise, although their actual grading skill is rarely utilized.
Lumber is tracked from the scanner to the trimmer lug chain using a patented invention Lucidyne calls the “True-Q™”. True-Q tracks lumber using the “fingerprint” of the board, with no ink or paint marks. It looks like a GMR and has a simple camera system in it to take a detailed snapshot of a small section of the board once it gets to the lug chain. This picture is compared to what the GradeScan saw; the scanner looked at all four faces of the board so it doesn’t matter if the board flipped over (or rolled in the case of a 4x4 or 6x6) before it got to the True-Q camera. The end result is a very low reject rate compared to other methods, with essentially zero maintenance.
Another benefit is GradeScan’s ability to have a solution ready before a board gets to the lug loader – so the Trimmer/Sorter Control system can create empty lugs when needed for cut-in-two solutions. This eliminates the need to split the lug chain or install a recirculation system to get the necessary empty lug.
The most important benefit to Ingram is how GradeScan® benefits their business. Going in, it was clear that the scanner had to be simple enough that Ingram’s current staff could learn to run the scanner. Ingram has found that indeed, one person can easily manage the system on a day-to-day basis. They sent three people to be trained so they would have some depth during vacations and shift operations. The three represent quality control and management, but curiously no technically-educated personnel were included in the system training. Mill management stated, “You need a techno-country boy to run this machine. One person can easily do it, but they need to have some grading experience and not be afraid of computers.” When asked if they need an engineer, one response was, “Hell no, you don’t even need an electrician for calibration or replacing a camera. But you do need someone who has common sense.”
The system has been in operation for less than a year but the Ingram team has already made the following bottom-line observations:
Planer setup is critical. It used to be that with manual graders the second shift production results were behind the first shift because the more experienced employees worked days. But when the scanner took over, the second shift volume of their highest value product would occasionally be double that of the day shift. It quickly became obvious that now success can hinge on one person getting the planer dialed in correctly.
When their machine-graded lumber hit the market, the mill received no complaints, even though their above-grade values dropped significantly. There was one new customer that they didn’t approach before they had a scanner because it would have been very tough to satisfy his custom grading requirements. However, the mill recently approached him and was able to set up some rules for his product so the first pass got them pretty close to what he wanted – at 4% below grade because of a pitted knot defect. Apparently this was a lot better than what he was getting from another mill that had a different brand of grading scanner because he said he would buy all Ingram could make. After some more some additional adjustments, the next pass got the below grade down to 1%. Needless to say Ingram’s customer is pretty happy.
Other ways the mill benefits from GradeScan are through trimming boards with wet pockets and having the option to utilize the scanner’s ability to strength-grade lumber for MSR and/or MEL. Since GradeScan is certified for stand-alone MSR/MEL grading, the mill expects to take advantage of this feature if they are required to or can benefit from adding strength-graded products to their offering in the near future. Continuous data from a Wagner in-line moisture meter is currently processed by GradeScan with the scanning solution to identify areas that exceed predetermined moisture limits. Instead of automatically downgrading the whole board, these pockets are treated like other defects with specific rules to limit their size and location – many times the best valued solution is a trim instead of simply downgrading the whole board.
Ingram does not need to have any assistance when they adjust parameters or even design a custom grade for a new product. Whenever they do need help, Lucidyne has watched remotely via the Internet to answer questions and step in if needed. The mill had one part failure right after startup when they lost a geometric camera – and replaced it themselves. Other than daily cleaning procedures, the scanner only gets calibrated once every six months; usually when Lucidyne is here for regularly scheduled maintenance service. A telling comment from the staff was that the toughest part was learning the new terminology for scanning. They also said that, “We probably haven’t used half of the capabilities that the machine has, so we haven’t run into any limitations.”
Ingram expects to see continuous improvements with grade results as they learn more about operating their system. They also have already seen value from new software for lumber tracking that was just installed by Lucidyne six months after their start-up. This improved algorithm was part of the continuous upgrade philosophy of Lucidyne’s maintenance program, and all installed (and new) GradeScan owners will benefit as they too, get their new software. Lucidyne’s perspective is that mills will profit more from incremental upgrades since they will always have the “latest and greatest” in their machine rather than just let the system lag behind for ten years and then install a major retrofit. Lucidyne’s owner and President, George Carman commented, “The technology and software techniques used in an automated grading system are as advanced as any other industry, and as such are constantly changing. Unlike a straight-forward geometric scanner for an edger or trimmer, complete fiber analysis requires far more information from sensors to truly understand the make-up of a board. It’s a shame for a mill to spend millions of dollars for the best there is and a few months later fall behind the next new scanner. And if our customers are successful and stay competitive, it is good for Lucidyne too. We have been in business for over 25 years and are growing conservatively as we take on more scanning opportunities.”
The Charles Ingram mill claims that production has increased with automated scanning. They did not see a need to speed up their process before automating their grading process so did not prepare for higher rates – but now they plan to install a new planer and increase their speed. Today they believe that they made a good decision with GradeScan. An important measure is that they have had an increase in their high grades with no increase in lower grades. And at the same time, this has been a tough year – they have been in the process of changing over kilns over to continuous, so have had a lot of wood sitting in the yard too long getting moldy. Once the kiln conversion is done, they expect to see even better results and claim that they continue to see improvement in their bottom line the more they learn how to use their GradeScan’s capabilities. The Ingram team firmly believes that if a mill has a good grading system it offers them an important edge in a crummy market when survival is the goal. An improving lumber market leverages that advantage even more.