sign in sign up
<<
Home , Wood drying

Dynamic Control of Moisture During Hot Pressing of Composites

Cheng Piao, Todd R Shupe, and Chung I? Hse

Abgtrsct Hotpmssingfssnhportantsfepinthe~ I. invedgatetheMCrn~entthughdynamic ture of wood composites. In the conventional pns- CQntrnl of moisture in hot pnssing, ing system, hot press output often acts as acanstrslint 2. dammine the thennal conductivity, perambiz- to~pmdudian.Sevleredryingofthefurnish ity, speci6c heat of themajor wood compos- bg*,gwicles* flab,or fibers) required 'by this pm- ites (medium density fibedmd [MDn part- ~SU~~~theman~costicleboad, and oriented ~~[OSBI), aadg.leates~iFpo~~~ortsd~~rgani~3. deVi?lop~~~f9reachpressing armpoatnds, which are a main source of air pollution stagedthe~tidpressisgprocess, bthe finest products in- Wood composites madeinthem~tional~hawhfghinteKnal 4. devebpmathematicalm&forthevacud ~andlowdimensionalstabitityd~.The perforated platen pnsbg process to define dedqmmtdapressingtechndogytoimpmthe tempature, mmand vapor pns-= p- ~is~tedbythenelea~e~fvaparpressure htsin a hot pmdng cycle, ad during hotpmssing. A perfmated caul system may en- 5. ~thevaalum~anthephysicald lage the area for the moisture to escape fkn the ine&nicalperfonnanaofthesedcom- baardrnatduringpressing,andand,makepossiMe pasites. high moiaum content (MC)~aessing.% objectives Apleliminarysbrdgon~plateas~ oftbis!3tudyareto: that the rate of core tempera- changes lwneased with inueasing flake MC and number of platen holes. Modulus ofrup~~e(MOR) was highest when the MC saapc: was at 8 percent and decmsdthereafter as MC in- ~~,Sdtodd~Natlaat~, ~statcItsfverJfty~~Rdulps,LA creased forail three amounts of platen holes. Internal HmK bond~~withtheincl.leaseofplaten PaJt~ResearchScienrid,USDAFonraSaulce,Poaprt per at MC Produds--w hde number unit area all Jd except the 2 ffa: pementMCWand~wfththein~of ~wood~UglAFanartSaoioe.SaabanRb fiake MC levels. Sbis study showed that the @rated seuch Station, Phdk,LA Ihiepapa(No. 06~)isplbllehsd~thtrgpm#ldtbeDi- caulhadthefuactioaofreleasingmoistureandvapor rectorof the lmManaAgiculW lkpe&m@statiosL presure during hot p~pgging. (Kamh and Casey 1988a, 1988b). The fow edges and axmm am the ody amas hmwhich the mobtun? ~ard~J9ringBot~can escape during hot pressing. Thus, mobtm, tem- In the man- of dry-process wood ampas- ~,aadvapor~~t31tsmformed ites. min-coated wad furnishes me consolidated fram face to core and from con to edges (Maloney into panels between two heated platens. The moishu'e 1977, Stickler 1959, !hcbdand 1962). The mat edges inamatisvagorizedd~inthe~cal are dcompared to the total surfaoe area of the and horizontal directions from' the mat rndaces. bOgTd,~,onlyasmallamarntof11~)~es- Since mat edges are the only areas exposed to the out- capes thmugh the edges and oomm; most moistwe side enviroment, vapor presswe is formed in the is released thxwgh the psnel Eaces upon opening of mat and is responsible for problems such as low care the hot press (Maloney 1977). Since excessive mob densityandintemaZbond(T]B)strength,blawsandde- tun may cause blows or delamination, the MC in a lamination, d resin wash-in and wash-out effects. conventional process is rigodycontrolled in a low To avoid these pblems, woad furnishes are dried to ranseh&dtyiqs. a low moisture content (MC). The drying process also MC and its distribution are important hctom gav- dtsin volatile c#.ganic compound (VOC) emis- ethe pmperties of wood composites. Heat and sions. The low MC nquhmcnt of the conventional moisture soften the wood furnish and facilitate pressing system lowers the productim of the furnish plasticization so the wood can be bent without break- dryer as well as the entire production line. The rigid ing. It is reported that the &mum condition for requbnent on mat MC makes the safety window of wood deformatkin is 15 tr, 20 percent MC at 100% the conventional process very nanowand substantial (Spelter 1999). The MC of a oonventional mat is nor- ]assesawincurred Thesolutiontothis problem re mally below this range and high swebgwould be ex- quires the use of furnish MC that 19 higher than what pected after the panels retake moistrim Rice (1960) iscurrmtlyusedinc~mm~plamsandanunderc found both bendhg properties and dhensional sta- standing of the mechanism of heat and mass transfer bility pqmtiesare impmved by inaedng the mat during~pressing. MC from 9 to 15 percent. Moistum may speed the heat The moisture in a mat of a typical dty-process tmnsfer from face to core (Kamke and Casey 1988b, wood~t~~prlnsarilycolnesfrom~Soemin 1974, SMckler 1959). Drying wood materi$ls sources, i.e., the furnish, the resin applied to the fur- is the dominate somof VOC emissions frwr the nish, and the mndensatfon reacton after the rcesin wood products indw (Englund and Nussbaum ams.h a typical proce%s, laat MC 2000). Englund and Nussbaum (2000) showed that fiDmthefudshmaybebetweffl2~tto6pcr- emissions of tapenes and VOC per wendry weight cent based an mdtyfurnish weight. For stmcmd rapidly inuwem when the MC is reduced to pels, wakr-bsed phenol-fddchyde (PF) resin below 12 percwit (Granstrom 2003). Also* high mat is widely used. The solid content of PF resin is about MC may reduce closing time (Maloney 1977), board 45 to SO percent; that is, if 1 percent mare of solid moistmy adsorption and thickness swelling (Maku msin is added to a mat, more than 1 pacent of water and Hamada 1955), and internal strain (Kelly 1977, is also added to the mat. The typid resin conteat of a Sucbs?and1962). MC addyinfluences other pm- conventional oriented strandboard (OSm manufao- ptdes. Excessive moisture may cause adhesive cur- twingprooess b 3 to 6 percent. That means that about ing pmhlems, longer pxxssing time, low care density, 3 to 6 parent of water is added to the mat along with and thus low indbond (IB) strength (Johnson the resin addition. For the third so-, a small and Karnke 1994, KeUy 1977, Malmey 1977). High amount ofextra moisture is added to the board mat MC k#ds to high gas press= in the mat (Kamke and by combjning two resin molecules and splitting off Casey 1988b). If moisture Is not deased sufficiently one watermolde. The total MC of a mat is from 7 to during pFwsiag, blows or delamination can occur: 12 pement, depending on the particular process. Thae are techniques to reduce or ehnimte the 3n the pmces of consolidation, moisture in a mat negdw eflects of MC and take advantage of its posi- fkstEhangestovaporraderheatand~md tive effeds. Several endeavors have been 3mplemm- migrates to the cooler anas of the mat am,edges, tedtoachi~vethisgdOnetechniqueistodistribute and comers along with volatile dvesof the higherrnoistutreinthefacethminthecaretop wood furnish and nonreactive adhesive mmwnents duce the steam-shock effect (Sosnh 1974, Strider 1959). Steam-shock is an advmce h both pressing and a flowing fluid when they are at Wrat tan- theory and technology. Strldih (1959) showed that -. Thermal radiation occurs as a surEace the heat transfer mte increases with increasing faoe emit* energy in the form of dectmmagnetic furnish MC, but modulus of ruptun! (MOR), modulus waves. Allof thesa transfermodes are present during of elasticity (MOE), and IB 1tnmgt.h decrease when hot pnssing of wood composites. the corn MC is at 9 mt. Numerous studies have been conducted on the de- Steam-injeaion pressing was designed to reduce velopment of mathematical models of heat and mass pnssine time d frnprwe dimensional gtabiliity of transfer in hot pressing. Bolton and Humphrey wood composites. It hcludes open and cldsys- (1988) conductad a review af the literatun? on heat tems. Both systems use perforated platens to heat the and momtransfer. MacLeau (1942) and Car- mat thmugh hot stepun that is ejected fmm the holes. ruthers (1959) measured the tempemwe changes in Intheclosedsystem, thematissealedandconsali- during hot pmsing. Kull calculated heat dated under elevated temperature and pnsme (Gei- derin Elakeboard and bdt a onedimembnal mer and Kwon 1999, Geirner et al. 1998, Shen 1973). model (1988) to predict the tempemtum changes in Attheendofsteaming,pressureis~leassdandthe~ the mat. Since heat and moisture transfer in winpos- mainhg steam is exhadto the atmosphere. In the ites during hot pmssing cxxws in three dimensions, open system, mat edges arc exposed to the envir9n- ollbdimensional models can dy@ct &mois- ment, as in the conventional pressing pmss. Steam is ture and temperature changes in the vedcal direc- injected~separatelyheatedphplatensandthematis tion. The onedimensional model cannot accurately consolidated by the heat fmm the pktens and the predict moisaire and tempera- changes in a mat. steam (Ceimer and Kwon 1999, Hse et al. 1991, John- Bowem (1970) evaluated the convection effects, and sonand Kamke 1994). Composite panels pdwith Cefahrt (1977) and Kayibn and Johnson (1983) in- the steam-injection method have high dimensional cluded convection effects in their ontxhemional stab- This was attributed to the combination of models. Weplasticization, Egrh flow, and chemical modifi- Humphrey and Rolton (1989) studied the heat and cation (Geimer and Kwon 1999). However, high MC mass Sansfer in pressing of partidebad. A ma& in a steaminjected mat dtsin poor mechanical matical model was developed using a finite ~~ pmpdesofthemat. Thisalsoistheresultintheam method to predict temperature, MC,and vapor pres- ventional pressing process when the mat has high a sure changes in the middle of the p-ticlebcmrd mat MC (Palardy et al. 1989). during pressing. Both conduction and mmeaion The solution to thepzoblem of high mohtumpness- were included in the madeling and chmlar baards ing is dependent on the release of excessive moistme wemz used to simulate dmedhensional hot-press- inthe~eofthehot~process.Inatypical ing ~tiozls.The initial temp3rature in- ~cycle,thesteamgressurp!ishi@estinthe predidedbythemoddf3tdwiththeeperimental middle ofa mat and diminishes toward hedges, resultr. Thed~trendof~11por~pre- whemthesteamisre1eased.ItiS~tedtbatone dictedby the model agrees with the memud+, arb& ofthetwomatfaasmay beusedasavenues b&thepdktedvaporgressurevaluesexdtheex- forvaprtoescapefromthematd hotpmsbg peiimemtal ones often by a factor of thee.. The differ- wenlarge the mktwxscaping m.7 e way to im- ence was partially attributed to the long flow paths ~thisfstohavehdesintheplatemsathatthe and the hoeofpress dosun in the model. The mqmatedsteammaybe~kasedthmughtheholes mocklfaiiedto~ctthesecand~intemper- &o the atmwphem while the board is being hot a~andwash~donlybrtheinitialandsub -. quent vamollstages. Kayihan and Johnson (1983) developed a one- Mdalbgonhertradsslws~ dimensional theoretical model fbr heat and moisture Heat transfer in a medium or between media is tmhrwhen pressing particleboard. The model was due to tempera- differences. The are thrree based on three simultaneous systems of equations de- modes af heat tram&, i.e., conduction, convection, rivedby Luikov (27). Both conduction and mvection and diation. If heat ttanski- occurs across a ma were included in the model. The model divided a half dium, then the heat is tmmferred by conduction. of the mat thickness into a numk of consecutive Convection heat trader occurs between a surface corn-tJ, each of which was divided into two phases according to the solid to void do:void space pared with the measured tempera-. The dlfFennce aradwodtsub&nce.Tofnchrdethebulkflowofva- ~thetwotempem~w~edasaniadicator porintothernade1,aItlirdngaw;lsusedasamea- oftlaLe~onand~~on,whicS,~ere suretosimulatedebulkmavlementhmhigh-u, usedto pdid t&e-pl~file. Thepdc?dden- low-pmme regions. The lateral removal of vapor s~ty~tsagreedwellwithwhatwasmeasuredex- was tnated as %&age" in the mdThe model can ceptwhe!nthetracLedTg~bdawthe~tem- beusedtopredidboundmb~,~~~mois-pemue (Wolcott et al. 1990). ?'he aumacy of the me, vapor pmsme, density, mpor mafsture, and to- techniques is dependent on the &dent measur~ tal pressw pnrfiles at diikmt mat depths. The tem- matof the difference between the two temperatures. peratwedropginte~~ll~oftenrpem~pmfifffintheFor numerical simulation, emexist to assume a expdments with fibe&od are consistent with the condnuumandlocal~~c~um,but concept of "lealdnf fntrudud into the modeel. errors can be minimized by meshing the simulated wl- Ikd et al. (2001a, 2001b) developed a finitcelanent ume dciently small. The huge praoessing capacity modelta~cttbe~~tun?,gasgressure,of modern c~mputemcan reduce the errom to sd6- MC, and VDPs of wood and straw composites. Some den@ small. The techniques of tracking Tgcombined heat conduction and convection jmmmters =ob- with numerical methods for MC simulation would be tained from apmhents and ovfPers wem taken fbm mom~andviableforthepn?dictionof?empkr- the litemhue. The pmdicted dtsof temperature, va- abre,vaporpssure,andmoisntreMaswellas pary,ad MC agreed well with the -t. density gdients. Harfess et al. (1987) developed a one-dimeasional Of cwrse,the main material in wood composites nmmical model to predict the density profile of is wood. Thus, the model used to predict heat and particleboard. The model that Included heat and moisture of wood can be adopted and further devel- mass transfer simulated the physical and mechanical oped to model the heat and moisturetmwfer ofcom- pnx?esses that occur in the press and mat system. The posites. Kamke and Wolmtt (1991) sixmdated the dosing pmams was modeled as a desof closing variation of temperature, vaporpressure, and MC in stages, wM& are dMnguished by input values for a pressing mat based on a mathematid model in ram mypower supply, and rate of closure. A mat wood drying. The model first defined the environ- was described as a series of levels, each of which oon- mental conditions surmunding an individual flake sisted of two or more mat lam.Mat layers within a omthe entire pressing cycle byestbating the equi- level were assumed and identical with llbrium moisture content (EMC)using measured lo- respect to NC, bulk density, and specific gmvity. At cal temperature and vapor pressure. Then a one- the end of any simulated pmss cycle, the predicted dimensional model was developed to desuibe the bitypfUempmse!nts layers at difForent tempera- heat and mass transfer inside of the individual wood tumsandfll~~values.~caldaCawasusedto flalw based on the enviromental condition and the calculate the strain development as a of tern- wooddrying model. Modeling implicitly assumed pa*ture and stress without onsibtionof MC. that the pdcted domain was infinitely large and The&iss~te?nperaftnr:ofapolymer,TT,,is edge boundaries were neglected. Tempemture and thetemperatmethatcot.respondetoa-in* EMC of a flake aad average flake MC were estixnated when a specific volume is measured again& tempera- by the ~IEodel.The predicted tempemture variation ture (1983). The amorphous dalsin wood, hemi- agreed with the measurement. and lignh, am glassy m@mes when the Zomborl (2001) dewloped a two-dimensional tern- is below Tg and rubbery when the tem- model using a finite diffemnce method to @ct the paatuxeisaboveTgForhydmscopicplymermatai- variation of tanpemum, vapor pmsme, and MC of als such as wood, T,decmses with an in- of MC single and thme-layer OSB. Madeling was Jso devel- The variation of TgEor hemioellhand with oped on the basis of a model on wood drying.As in MC was desuibed by the Kwei Equation (Kelly d al. most other models, a subcstantial dkrepancy existed 1987). Wokott et al. (1990) used a one-dimensfoaal between the predicted and expedmental values. The heat and mass transfer model in conjunction with model qualitatively agrees with the experhentat measured temperature and gas pmssure to predict lo- data. This was attributed to the materids properties caked MC. The TB of haniceMoses and lignin in that wme obtained from the libmtme and not the wdodwastrac%edthnwglhthe~cycleandarm- ones on which the modd was built. Howevez; two ad- vances were made in this modeling#that is, the Mu- isafunaionofparticlegeometty.Thesteamdissidisaipa- shof dosing time and heat generated from the exo- tionofflakeboardisslowerandtbepr~sstimesare thennfc reaction of resinpolymerizatio~~A sqamted longer than those of particleboard and MDF. Kamke prommodel was also developed based on hke and Casey (1 988b) hmd that a more permeable mat gemn&y and density and used to simulate the stme the presswe buildup and thus lead to a 1- huaI changes of the mat during compression. ~pera~plateauinthecore.Thedifferew#infur- Most of the above models began the simulation nishshapes also affects the voids in the boards (Lenth hthe point a.pm c1om-e. The heat trans- and Kamke 1996). Quantitative pdctiom depend fend by radiation and time difference of the upper on the accunwy of the ma& pametem on which and lower mat surfaces in aolltadq with the platens the model is built. The way to gemdze the model to were neglected. Closing time is typically mom than 10 beviableforothermaterialsistoindudethe~~ percent of the total pmsing time. From the moment characteristcs in the model. when the mat is put in the press to the moment when the mat is pressed to target thickness, the suditce lay- Current Approach ersafthenmtmwarmedandtempemtumandmois- MOaimg-rrrlm tun~entsareformed.Enwswould&6mm the selection of boundary conditions in a model with- When hot pressing wood composites, the ktto mtincl~thepresscl~TheddcPeloped moistwe vapo&ns water to vapor, but vaporization is by Zombori (2001) sixdated the enthe pressing pfo- limitedbythevapar~inthemat.Afterphtens oess including closing time, but the predicted temper- contadthemat, heatis~rnainlybyconduc- atnmandtatalpressurewerepoaiy~with tim (Bawen 1970,Kamke and Casey 1988b) and mois- theexperhmtalmeasurementinthepmsdosing [email protected] period.Amther

when:

, HL I of kg-I), VtheE htheat vaporization (J H, = (J E~ = energythat00wsintothewde, adsorptkw heat b-9, 4, ~~~density(kgm-~),and E~ = eneqy~tedhthe~and - Vv = vaporvolume (m3 ). . E~ =energythatismd Accdhg to Dam$s and Fiis laws (Siau 1995), the In~onalspace~eeehnodehessix~vaporvohrme V, in Equation [S]cm be calculated by bc#ingnotes(farrarre~inFig.2).~ex- *tion 163 d.langeisinfl-dbydaand-onb tweenm, n, k,anditssixadjoiinodes,aswellasbg geneation. "ha Eqyation [Z] may be expresd as: * K = specific permeability of wood composites (m2)* whem A = cmss sectidarea of flow pth (d)? gtiwmrrfi - conduction Tate ad, t = duration of flow perlod (sec.), L * ~0- length (m), tiom. Since the txad-er of heat and moistme is 9 = upstream pm=u= CN m-2h coupled togethw, the moisture and vapor PI = downmeam pressure (N m-2), gtadientswiltbeobtainedsimul~in~cal~ n = vapor dynamic viscosity &g m-'~~), lation of temperature gradients. Dm = vapor diffusivity (kg masm1), pw = normal density of water (kg m-3), M-=-F-= G = spec& gravity of wood (I ,000 kg me3). ~ounQustandheatandmass~indepth, and spe&Uydesiipedomtedplatenswi.lIbeusedto aAa = MC difbeme (%). dease the vapor pmssme drning pmsshg. The hod- zontal cham& connected to the perhxations urSI1 be cdto a vacuum. Fipm?la shows the mois- twe mavement in a mat pressed wi?hsuch a platen wittmut a vacuum. To &ted heat Wadk the vacuum wiU not be applied before the aretempera- tun 80" to 10006. When moishm is vapor- where: ized, -a part of the vapor produced may descend in the mat, as in the canwntional process. The mn&der finds its way &mu& the holes to the end-nt, as shown in Flgun lb. Thus. "U"shapes of vapor tracb wuuk?be apeaid -out the mat thickness. Va- porprssure in regionsnearholes &be lowerthan a =percentageofarndensador~d that in the conventid pmcess. However, perha- (96). dons haw little effect on heat transfer due! to rela- The heat generated- by the exohrmk reaction of tivelysmallareasofthehales.Afterthemo~in gbcanbeexpressedas(Zam~2001): thematco~reaches~to100"C.mostwoodfur- nish hm&out the path of moistwe is plasticized. Keepingm~inthematanyfintherwouldbe d-' rtal to the board Tfre saturated moisture would the further irzrrease of temperatme in them(KamkeandCasey 1988b)andwashoutthe resin on furnish swfaces. The rnohtum should be rp movdlasquidyas~blempxztctical.~is & = ~t~of8hte(itgm-~l.d ~avaclsumstarts.Ifavacuumisapplidtothe AF = reactionindex-. holesatpointB inFigutw Icandd, vapor will be The esletgy stored in the node m, n, k may be ex- cmauted and mpor pmsure will dmp, as indicated pre!Bed as: bythedpbdlineshowninFigm Id. Also,anega- tire~wilSallowpossible=mainingfrec:water to be quickly vapized and evacuated out of the mat. Partsofthemsin thatwas "wasbdinawhenthefur- w- nish was heated &mu@ condensation of water woddLikdybeuurasheddtothe furnish sufaceas p =l-massaty&theM the mohhm quicldy evaporated. The detrimental ef- mat, fects of "wash-aut"can also be &tmed or avoided. c = spec& heat Q kg-'IC1), and Thiswouldbeachidby~the~time dt = timeinaease(s8(:*). to start the vacuum. The corn tempe!mture wauld in- After the calculation of the comecdon and conduo crease again with the quick remod of moisture and tiantransfw18tesftrrmthesixncighborlngnodes, pmistm. The vaporization period, B to C in Figure Equation 131 can be used to calculate the temperature Ic, will be neduced to B to C due to the evacuaton of of this node. The temperature gradients of each step mohmby the vacuum. Thus, thrwgh dynamicaUy and each time period can be obtained through solving mmdng the moisture, the pressing time and "wash- Equation 133 along with boundary and initial con&- out" effects wodd be reduoed. Little heat traasficr oc- curs from B to C in the padel-to-faa direction due The perforations in the platens will be separated by to~vacuum.ThevacuumwillbereducedatCand an equal distance. Xu modeling, a grid will be vkdly stoppedwhen thepressst~utstureleasetheboar& supesimpoeed over the board domain, Each hole will Theheatis~frPmCtolYbyd~n.Asbe located at the center of one element area in the sbowninFQtm 14 thevaporpressm(thedashed grid. A finite difference method will be used to sitnu- line) after evacuation will be much less than the late the variation of temperatuxel vapor pm,and pel's IB strength, adthus blows ar delaminatfon MC in the pressing mat Since the evacuation is un- would not We1 su- the af steady state mass transfer, the time diffmmtial wlll be hesit transfer- in a typical pressing cycle. smallmalltohave an acemate simulatioa The turbulent Four kinds of flaw can occur in woad (Siau 1995): lIlodelingwillstartfmmtheisothcrmalstageandwill 1. viscous or *, model each element volume to save amputmpro- 2. Wndent? cesshg time. The solution of one dement will be the 3. -,and lwnmby and initial cxditions of surrPunding ele- ments. Heattranskin vacuum period is 4. ILW~~slip flow or Kwtdscn diibh. the assumed Thelaminarflow~wkthe~olds's~kto be c#mdudan. isbebwabaut2,000,atwhichpoiman~~ causesasmooth,streamiinedflowwithrtohrrbu- ~In~~tion~f~co~tidpressingThe exprbmtal mrbibles of this study included process, a laminar flow was assumed and Daxy's law thme levels of platen hole number (PHN) (no holes is wasusedtoddbethemass~feP.Whenthe PHNI; low nmhberof hdes is PHN2; and Mgh num- ~lds'snumber exceeds the critical dues, lami- ber of holes is PHN3) and five levels of £lake MC (2%- and each nar flow begins to break down d eddies and dhu- 8%, 12%, 17%, 20%). For PHN 14,the bancesocc~r.TheThe~~lmdesthbe~~lditi(~is~-holes were udomdy distributed thxwghout the aIu- lentk.Whmadcalvanuunisappliedtothe minum platens, which wcm 0.32 cm (118 h)in thick- ness. FigPrt 3 ilhstrates one design of the platens. ~t~~~pdessingsbqplairandvapor an?eraa;latedfr-om~piky-pon?reghesandtufbu- Southern (EM sp.) flakes wem obtained ~fbwupatiansareneededto~bethemass fnom a kdOSB company in centd Louisiana The flakesu~nftutherhammemdedjatosmaIlerBaka and heat tradk Under the tddent am&ions, an usingarnotormiUer(lndustrhlPhsman~ exponential rehionship exists between the &ow rate by Briggs per- andpnessurediffkmn~wthertbantheliwar& & Strattcm Gorp.. 4 133412) at 8 tionshipdes&bedbyDaFcy'skw,asshownin~ cent MC. The @cles wee screened on a 6-cm mesh on tkmC61. Thegovemingequationforturbulentflowis screen, and particles the sueen were then uni- given below (Kuiko 1%6, Siau 1995): fomdy divided into five gmups. Eve mobtam treat- mewere randody applied to the five groups of particles with each gmup dvingone treatment. Theg~iupwlthatargetedMCof2peroentwas o~ed.Theotherfourgnxlpsofparticles~ condmanod*. acconhg to the taqeted MC of the group in a conditioning *bec The PF resin used

k = pamabWty of wood mmposites (rd), AP = &?a=sm-(Pa), Stawnstant, q = v-hxmsityof vapor (lg m-1s-1).

r 2 dkncterofaca.piby(m),and P = d-wdvapor(kgm4). The first term on the right of the Equation is the flow & in the qdhy-pornus regions formed between parti~andthedtermontherightisthe~8pil- usedittthemmr~of~(a)aOQUIph Euy flow rate inside particles. and (b)&&n ofa particular hole a the cad was obtained fnm Borden Chemical Company The Figme 4 can be divided into three parts. The first pat% solid content of the resin was 51 percent. is the Ztntar iwmw part of the temperatme--time Pdwem made with a miin conkof 4.5 per- Unes. The water in the bo9t.d core began vaporhhg at cent. Resin was applied to the particles in a drum the end of this period, which lasted 1 to 2 rnirmtts af- blender by an ait.atonMng nozzle. The targetsd den- ter press dosum, dqxading on mat MC. Vaporizing sity of the boards was 0.65 @an3. he target thickness tem~~)ofPHNlworethehig3lestatthr: ofthebaardswas 12.7~1nthemiddleofthefonn- five flake MC levels and all pattrthan those of other hgoperation, tbmdcouple wlr;#s were buried in PHNlevels.TheVBof~~withPHN3 the geometric center of the randomly formed mats to (i.e., the highest hole density) were the lowest and measure the temperature chmgm of the mat core were almost egual to 100% for the five rnoistum lev- during hat After hmhg, one of the treat- els. Since the vapor pmssure is proportional to tem- ment platens (no holes, low PHN, high PHN)was mi- peramthe vapor pressure inside the mat farPHN3 do& selected and put on the top of the formed mat level was similar to that of ambient conditiotls, as theupperplaten.'Rw,boardswenmadeforcach One h&tof a high MC is the @ck heat transfkr combination of MC and PHN variables. For each fromboavd~to~re,orthe"~ock"efFect ~,thetempera~ofthecomwene~im-(29). To analyve the effects of MC on heat tamdkr mediately after the pressure reached the maxh~urn fmmboardfkcetowre,hear~on~ and kept neconbg until the opening of the press. wore conducted for the data that repmsented the lin- The baards were tested for MOR, MOE, and inter- ear part of the tempmituretimc lines in Figum 4. nal bnd (XB) stmngth, water absorption OVA), ad The slopes ("CIsec.)of the mpssion lines m summa- thickness swelling (TS) according to the ASTM rizedin~b2.Asshownin~~2,theholesonthe D 1037-98. platens slowed the heat transfer from panel face to Taperatun changes ofthe board core with press- coreatafhkeMCbelow13.7~torndhadUttleef- ing time forthe five MCs and thrree PHNlevels are pre- feet after this MC point. However, the theheat transfer sented in Hgnm 4. Each temperature-time line in rates indwith the inuease of mat MC for dl

UC: 2.1% HC: 6.79b Mi2 13.m MC: 17.0% MC: 19.7% 1 r

M MC (96) 2.0 8.0 13.0 17.0 20.0 --Actual MC (%) 2.1 8.7 13.7 17.0 19.7 PHNl"" 2.08 (0.066) 2.89 (0.105) 2.77 (0.038) 327 (0.066) 411(0.430) PHNZ 2.08 (0.032) 2.85 (0.133) 2.48 (0.095) 3.15 (0.501) 4.19 (0.522) PHN3- 2.05--- (0.054) 2.66 (0.275)- 2.54 (0.142) 3.70 (0.145) 453 (0.2971 'O~fapueskes~thb~~ ~f~nprrsentspa*DwbaZe-~~~l-mr~~~~a-hbaklllmtkr~~nd~3-~bols~bcr~~ platen number levels. The amage temperatwe Mi cent even though total mat MC was about 22 parent mtes were 3.M0, 2.9S0, ad3.109: per second for for the fifth group furnish. Although boards pn?ssed PHN1, PHN2, and PHN3, mspective137; for difkmnt with PNNl did not fail, the bonding between flakes MC ld,the average rates wen 2.07*?2.809 2.6E was weak. Boa& with 19.7 percent moisture level 3.37*, and 4.31% per d for 2.1, 8.7, 13.7, 17.0, were delaminated except for one board of F'HN3. and 19.7 mntflake MC, reqedvel~ A-values and standard deviations of the phy- The second part of the temperahue-time lines was sical and mechanical prop&= of tested flakelmad frwn the time when vaporization began to the dme specimens are sdzedin We3. Most boards when core ternperam indagain. Moisture in at 19.7 pement ffake MC were delaminated, and the the mat was vaporized in this period. This period data for these boards was not available. The VDPs of lastedhmsecondsfor2.1 pe~mtMCtornorr?thanZ Merent MC and PHN Iwls are shown in Figum 5. minutes for the 17 percent MC for both PHN2 and Moisture bad significant effectson flexutd proper- PHN3. It is noted that the second part never came to ties. Figure 6 shows that both MOR and MOE anendintheprwsingcycleforPHN1 ktheflakeMC mukd the highest level when the flake MC was of 17pemmtand 19.7pe~ent.Thetimeusedinthis about 9 percent. It is evident that the fldpmper- period indwith kcmasing MC and decnasad ties of boards at 2.1 mtftake MC we-or to withindgPNN. those of bodat other MC levels except MOR'at the The rest of the temperatme-time line is the third 19.7 percent level. Both MOR and MOE demwd put. For the last two Bake moisture Id,the third with an increase in MC after the 9 percent £lake MC. part never came to the PHN1 (17% and 19.7%) be Both MOR adMOE of PHN3 were lower tban the cause the press was opened before the ternperam other two PHN levels. However, PHNl and PHN2 lost bmxsed. The highest temperature mached at the more men& than PHN3 as the MC indfrom end of gmssing cycles derreased with increasing flake 8.7 to 17.0 percent. MCs. fBvahresofboardsatd@erentPHNandMCleveIs After the boards were pressed? no blow or de- are shown in Table 3. As a high MC deep hination was found with furnish NIC up to 17 per- ened the VDP cvves for platens without holes; while

spedfic -# tFfW MOR MOE IB TSb WA~ NTS"

- -PHN3 ~LTn PHNl

Pion. Sb.and H.w 4 37 FiguFt6.-~sctsOfPH~ and MC on (a) MOR and (b) MOE of southern pine fZakeboard

there was Iittle in the VDP hrPHN3 In the ing hot pnessing, the "wash-in"and "wash-out*effects Eaur MC levels. XB stmngth of the boards inawmd were dudat high MC levels. When the flake MC with the number of holes on the platens To compare was 13.7 percent, IB strength increased by 12.9 the dlikence in IB values among three PWNs at each percent and 30.8 percent as the PHN increased from MC level, pirwise cornprisons were made and the 1 to 2 and from 1 to 3, nspedvely. The same per- resultingp-values of t-tests arelisted in Table 4. At 2.1 centages of increase were found when the £lake MC pemznt MC level, PHN had no effects on the bonding was17pemnt. pl.opertiesof~board.Sincesomemois~wasre- TS and WA after 48-hour soaking and non- leased from the board mat once it was vaporized dur- rec~vcrablleTS after 21 &hour soakfng aru! jmsatd changed the heat and mass pmpertiesesof boardmatmuinghot~Fkxuralandbonding proprtiEsofllakdmd~fiegativelycodted with~fhkeMCwhentheflakeMCwas greater than 8.7 pacent. IB strength of boolrds pressed at modemte to high flake MC was signjfi- dyimproved with increasing; PHN Iev&. Dimen- sional stability in thickness was negatively correlated with flake MC and impmved with incmasing PHN lev- els.Ourcumntandfirtundinclude.. 1. establish a heat and mass traask database and BoPrQinthisgroupbkw~~an~k Values amsignificant at the 5 pe~cent&&cant led. finite di&rence modas that quantitatively sim- W a pressing cycle of particleboard, OSB, and MDF for both conventional and dynamic pressing%-=# 2. ascertain the gas permeability of pqrticle bods, OSB,andMDFasafkmdbydensity and MC, 3. detedqe the thdco-vity of partide- budand OSB as affected by d-, MC,tern- PHNZ 0.13 (0.02) 023 (0.06) 0.31 (0.02) 0.28 (0.03) and furnish sizes, and PHN3 0.17 027 (0.02) (0.04) 0.26 (0.03) 0.33 (0.03) 4. dedopaldynamicpressingsystemtooptimk- val~esinpaFmtbeses9nJtandardarrre the pmprties of wood composites. PHN npnsen~~platenhdenumbcc PHNl-m~es:PIINZ- low hole mrmber~d);PHN~ -high hole nrmba (bb1d). mrslmm Amekan Society for Rsting and Mat- (ASTM). 1998. Standardnsethodsofevat~the~ofwood- base fiber and particle panel materials. D1037-98.Philade- in Table 3. Flake MC had signi6cant e&cb on TS, WA,and non-nxmverableTS (NTS). Both 'IS and NTS Bolton, AJ. and RE. Humphrey 1988. The hot pabg of imseased as MC increased from 8.7 to 13.7 pemxnt. dy-famad~ann~PeatLAdmoftbc FHN had no efkctson TS d WA,but hid significant li~,idenntiEyingtbepsfmazyphysidp.ocessesand the nature d their interaction. HoIzforschung. 42(6): effiects on NTS. NlSs were significantly reduced 403406. when the boards were pressed wit31 perkated plat- Bolton, AJ., PK Humpany, ,and PX. Kavwms. 1- Tha ens mable 3). hor~ofmy-~waod-~~teb.PaIt The Linear dationship between WA and TS was m.Pnaictedvepor~andtemperatlnr?variadon witla time, armpared with experimental data for hhatory fMlnn by fahmann (1978). A semktionsbip be- txlardn 43(4b199-206. tween WA and TS was fdin this study. The slope Bdton, AJ., RE. Humphrey, and PX. lbwmms. 1989b. The duesofthisfinear&-mF&~+~~tedinS trot~ofdrp-~woOd-~~~lgpoeites.PartW. &5.Slopevaluesiucm~edasflakeMCin~ ~variationofmat~eontentwithtime. mt. that 43(3~265-274. from 2.1 to 13.7 Thisindicates boards Bowen, ME1970.Heattransferinflak&odduringpless- ~ata~flakeMCsvvdldnwrreaftep~ing.Ph9.nle&.~SEeteUnia unitpexenhgeofwaterthantheboardspressedata athaaJRS. 1959. Heat pmetraticm in tbe presshg ofply- hfiakeMC. PHN had no eflktontheslapevalues. wwd FOR?& Res Bull. 44. HMSO., Lodon. Clmong, EX, KO. Tesaso, &ud F.G. Mamilk 1974. Pame- a~oftw\?nty.twoddiameteP~growingon scut&m piae sites. Wood Fiber 6(1)91-101. The pr$iminary msu.lts show that the -ted W,C., C. MI, and B, Waag. 2001 b. M- of OSI) hot ~hadthefunctionaf~1110istured~~porinspmcbses..'I;echnicalreport,F~~Carp. 18 fiomthematdurh.lghotpnssing.h*aurrat PP- hi, C.C., W. Wasykiw, and J. Ti. U)Ola hdngbehavior of +tar design. boards can be pmsd without Mows ddstraw~tebaerds.zn:Proc.oftheutaira- up to 17 pemxmt flake MC. The perforated caul tion d Agriculd and Fmestq Reddues Symp. Oct. 31- Nm 3,2003. Jhgm, China, Sdance & Rdmiqw Litem- ICayhm, F. and JA. Johnson. 1H3. Heat and maisame move ~Press.405pp. mentinwoad~materiels~thepreJsingopa Eh$uad P. and R.M. Nussbaum. 2000. Mon- in Scots atioa-Ashp.WedmoBJ.In:Nmdcdnaahodsiohsat ~amdN~spmazandthct~w~Mlndry- W. 2511-531. RW. Lewis, K Morga~~andBk ing. Hakhdnmg. 54(5)&-456, Schref&r;= G~u%J. 1977. ZLWSvug mit EX-- Lcnth, C.A. and FA Kamke~.1996. InHssdepationsof fhkehd gie-ModelIzurEerecbnuagdesT~in mat-oa.PartL-the-- vliemitte bei der HsiBpItssung. Holz Rob Werksto& a~.Wood FikSci. 28(2):153-167. 3!5(5)183-188. Luitov,A.W 1966. Heat admass transfca in Geima B.L and J.H. Kwoa 1999. flakeboart- d- bdcs. lk-@monResr,Loadon. ing. Part II. Pmdmneptal vnse ta Maku, I: d R. Hamada. 1955. Sadies on &@bad II steam injeaion pltesring. Wood FiiSci. 31(1): 15-27. swt~lingp~ptrtiesofchtp~.WoodRes.KyotoUnix Gainer,R.L., J.H. Kwon, and J. Bolton. 1998.~boardtbick- 15:53-59. ness swelling. Part I. Stress ralaxaton in a flakebod mat Maloney, T.M. 1977. Madent Pkekebowrl and Dry-- pi- Wood Wber Sci 30(4):326-338. berbod~~FonstRaducrsSodery,~- Ciraastslom,KUH)3.~~0f~rmdVOCs Jon, WI. duringdryingof sa~fnaspoutedbcd.ForestlProd.J. Madan, JD. 1942. The rate of Fl?mpaaaarc change in wood 53(10):48-55. panclsheatedbet~eenhot~.USDAF~Products H~~~I~~~,TZ..F.G.W~~~QP.H.W,RD.M,PXWLab. Report 1299. Reu 1960. and D.S. Ladd 1987. A model to pl.edla dedcdty profile PPtardy,RD,B.A~,~.~,AD.Williams,d~ of padckbod. Wood Fiber Sd 19(1):81-92. Laufenh 1989. Pressiag of wood coanposite pads at Hse. C.Y., ILL. Geima, W.E. Hse, aud RC. lbg. 1991. Efktd modaPte tempaatun and high moiaup CQ- Forest Prod J. 39(4).27-32. Rta, J.T. 1960.The~ofselectedvnriaMw6ntb~- NOK 19-21, Seattle, WA, C.Y. Hse. B. Tomita, and S. J. ties of flatp.essed fhkthds. M.S. Thesis, Dept of woad -, Ed. and ScL. North CarobState Udu, Raleigh, NC. Hmqhql RE. and AJ. Bdtan, 1989. Tbe hot paesdng uf Shenl LC. 1973. Steaxn-pws proces for cur@ phendb &-fonned wood-based compc&es. Part II. A sirnulation bonded Farest Pmd. 3.23(3)21-29. modelfoaheatandmoisarretransfrs,and~msuk Smith, D. 1980. Cod- in prem design frrr !mucod Holzf- 43(3):199-206. boa&. In: F~Lof the 14th International Symposium on Johnson, S.E. d FA Umke. 1994. Cb ' 'sties of phe -, Aprii 2-5, Washingtoa State Univ., Rrllman. d-fonndd&ydeadhesiveboMkinsteamiajcctbnpes+ WAT.Malortey,rn ed &keWWood Fiber Sci. 26(2)259-269. Sokunbi. O.K. 1978. Aspects of we KamkeI FA. and U.CaPey. 19888. Gas pmsum emd tempera M.Sc. Thesis, Univ. ofoft\iglea, US tltninthe mat during BaleeboomlmanhForest Rod, !hmh, MJ. 1974. Untem~hmgdea Elastizitat d Wrhm- J. 38(3):4 1-43. Mtder Spdiese bei & S-m. Hdz- Kamke, FAand W.Caaey. 1988b. Fundament& dflaLsboard 15(1h45-48. manufactwe: Inter~I-matconditioas. Forest M. 3. Sgdta; H. 1999. Steam has podto reduce thiclmess swell 38(6)..3&44. but oimtdes must be cmmmme. W WorM, Sept. pp. Kade, FA. and !3.E. Joh.1991. Phcncdic nesin interaction 49-5 1. dwhg~-~~m~~f~.En:~.aBStridder, MD. 1959. Efbof pn?ss &es and MC on paper- tbeAdhesivesandBolldedWoodSymp.Nw.19-21,Seattle. ik ofDt~&- flalreboard FmM Y. 9(7)*303-215. WA, Chung. Y. Hse, Bunichiro Tomita, and Susan J. Suchalrumd, 0.1962. Tbe density distrhtion of -, Ed. Michigan Agri. Ex* Sta. Quiutdy BuUetiu,- hkhigan Kamke, FA and M.P. Wolm 1991. Fmhma&& of &b State Unix 45(1):104-121. board Itlmw: woad-- Ielatie. wood Ward. I.M. 1983. lbkdmical pmpeathof solid poIymas, 2nd Sci Tech 2557-71. ed. Wilqr-In-. New Y& NY. KdljM.W. 1977. CriticalU~Icevlsnvof~dpsbe- Wolcaet. ME?,FA Krrmlte, and DA Diilard 1990. Fudmmi- -w=-k3~~~~prapatiesof mlsof---*oftbe piarticleboards. Publ. No. 5065. N. CardmAgd Exp. Sta., wood component. Wood Fiber Sd. 22(4):345-361. Raleigh, NC. zomb0ri.B.G. 2001.Modelingthe~ientEffectscfiaiagthc Kelly, S.S.. T.G. Rids, and W.G. Gbsec 1987. Rdaxation be Hot-pIesdng of wood-- Composites. Ph.D. Diss., Wr- haviaroftbeamorphrusann~tsofd3. MaeSci. ginia Polytechnic Institute and State Univ.. Blah* VA. 223617-624. pp. 212.

40 6 Recent Dtmkpmm in Wood Companites

Recent Developments in Wood Composites

Edited by ToddE Shupe ~zte~cpsor; Louisian~Fopwf Mum DeveIspment Cenrer; Schoor of lbmvable Natuml~~ LouisiQna Stilte univm- Center; Batan Rouge, Louisianq USA

Forest Produds Socfety 2801 Marsha Court Madfson, WI 53705-229s phone: 608-231-1 361 Fa: 60%-231-2152 www.forestpmLorg Table of Contents

Assessment of Screw Holding and Other Property Variations of Canadian Furniture-Grade Particleboards by EmmdK Sackey, Kde E. Semple, Je Jan Park, and Gregoty D. Smith .1

Properties Survey d FurnittursGrade Partic1elxmrck MS and M2 Grade Comparison and a Practical In-Situ Test for Internal Bond Strength by Kate E. Wpk,Emd Sacw, Hehn Park., adGegory D. Smith. ... .9 fldProperties, ][ntemdBond Strengthf and Dimensional Stability of Medium Density Fiberboad Panels Made From Hybrid Poplar Clones by Jun fi &if S.X Zhang,Wd Riedl, and GikBwtte...... 7

Dynamic Contd of Moisture During Hot Ressing of Wood Comtes byChengPiaoIToddF.Shupe,MdCluurgY.61se...... 3 27 CompdonBehavior of Resinless Oriented Strandbard Mat: Effect of Internal Mat Errvironrnent and Void Volume Fraction by Jong N. Lee, LapT. Watsotz, and Frederick A Kbmke ... 4 1

Properties of Bio-Based Medium Density ~~ bySa@Lee, TdF.SIrupe, dhy.Hse...... 0 89J a Manufamng Particleboard from Under-Utilized Spedes in Oklahoma by Sdim Hizi~~glu...... 59

Development of Binderless Particleboard hmKenaf Core Using Steam-Injection Pressing by f ianying Xu, Ragil Wyourini Gwn- Han, ED. Wmg, Yuzo Okudawa, and Shuichi Kawai ...... 65 Optimizing Natural FibReinfod Polymer Composites Through Experimental Design by Rupert Wmmer, Andm Ganz, Cunter Wdad Rudoffk&r. . . . . 73

Characterizing Wet-Proass Hardbr>ard Man- by the Use of Eixpaimental Design by Thom~sHuttq Rupert Wiynmer, adRudolfksler ...... 8 1

Heterogeneous Nucleation of a Semiaystdline Polymer on Fiber SurFaces bySangyeob~,ToddEShu~e,LesIieH.Gnw,m,dChungKHse.... 6