Hy­brid grade scan­ning sys­tem for green mill ap­pli­ca­tion

Australasian Timber - - SCANNING -

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.

Uniquely Pukepine

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.

Grad­ing chal­lenges

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.

Ad­di­tional projects

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!”

Giv­ing credit

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.

Look­ing ahead

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.

Newspapers in English

Newspapers from Australia

© PressReader. All rights reserved.