University of West Hungary Sopron DOCTORAL (Ph.D) THESIS STUDY OF WOOD SURFACE ROUGHNESS ON P AND R PROFILE FOCUSED ON BIG POROUS SPECIES Csiha Csia Sopron 4
Graduate Schoo: József Cziráki Doctora Schoo of Wood Sciences and Technoogy (Head: Winker András DSc.) Program: Discipine: Wood Technoogy (Head: Dr. Kovács Zsot) Science of Materia Engineering and Technoogy Tutor: Dr. Szabó Imre
I. The object and aim of the research When anayzing surface roughness of different wood species there appears a specia probem deriving from the nature of the materia that unti some species have an even anatomica structure, free of big opened vesses, ti than other species have bigger-smaer crosscut vesses. The presence and number of big pores has a serious infuence on resuts by distortion of roughness parameters, makes impossibe the adjudication of machining quaity. In course of surface roughness evauation for big porous species there is the probem that even in case of a machining quaity supposed to be good the different roughness parameters obtained as resut of various statistica evauations suggest a bad machining quaity. This fact raises the necessity that for an objective adjudication of machining quaity the vesse grooves to be fitered out. The probem is characteristic for measurement and evauation of big porous species where crosscut vesses appear aong the measuring ength (acacia, oak, ash, birch). Because this phenomenon affects ony a group of species fe out for ong time from the focus of attention for a ong time. With raising quaity expectations there is a need for changes in this fied too. By fitering the disturbing presence of vesses are not soved a the probems which characterize the surface roughness measurement of wood. In practice surface quaifying is carried out by touch by hand (tactie measurement). Beyond the vesse remova quick and objective measurement woud be carried out if there woud be avaiabe a roughness parameter to describe one parametric the rea estate of the surface. In this contest further investigations were carried out to decide woud the primary P profie and its P z parameter - figured out upon the R z parameter - fufi the expectation. The actuaity of the research was underined by cases when the producer and the procurer coudn t agree about the roughness convenience of oak tabe surfaces in absence of measuring method and parameter. The measurements performed for this PhD work have a regrettabe imitation introduced by the styus tip measuring instrument, which isn t suitabe for inine measurements. Taking into account the expected ransom of the instrument the method deveoped for vesses to be removed is indifferent from the way the profie points were coected. The method focuses on vesse fitration, making possibe a wider game of wooden species to be estimated for machining quaity reasons meanwhie deivers some unsuitabe aspects of the traditiona measuring technique. Concusions based on iterature: 1. Reviewing the referring iterature it shown up squarey that the big voumen pubications appear mainy in German speech area and most of them dea with paining and mouding and their correations. Ony a few papers dea with sanding.. From papers marks out that for wood surfaces first of a styus tip instruments are used. Most authors eaning on meta industry standards and instruments use the disposabe measuring equipment for wood surfaces without any critique. Ony a few artices make questionabe the suitabiity for wood surface measurements of the measuring instrument and method and suggest the need of review. 3. Some authors have the common opinion that based on Abbot curves parameters the anatomica and machining quaity can be separated in form of Rvk and Rpk, Rk. In front of these considerations in many papers we meet the observation that the presence of big vesses makes the evauation impossibe. It s noticeabe that the evauation probems caused by presence of big vesses generay are handed by the authors in two different ways: on one hand at the seection of the species to be investigated the big porous ones are et apart or on the other hand these species are recaed from the range of sampes by the time of evauation. There is a need for a method to be evauated for a better surface quaity evauation of big porous species. 4. Reated to the direction of measurements we can resume that the authors aways choose the measuring direction perpendicuar to the fiber. Their decision sometimes is supported by comparison measurements conducted parae with fiber, other times it isn t. On the measured profie perpendicuar to the fiber they define the waviness and roughness profie. In the same time some authors mention the waviness ike a so-caed chinematic roughness measured aong the fiber. This means that in case of wood makes no sense to define a wave deriving from machining perpendicuar to the fiber. The measurements were performed with the Mahr instrument with the imitation that the measuring and computing instrument doesn t aow the cacuation of roughness parameters without deriving waviness and roughness from the primary P profie. In this context the assessment of waviness perpendicuar to the fiber is a imitation attached to the measuring instrument. 3 4
5. The R parameters are computed ony for the R profie athough in case of wood this means a distortion in comparison with the rea profie estate. There is a need for roughness parameters to be defined for the primary profie and the contro of their suitabiity for to be quaifying parameters. In the mirror of the above considerations the present work marks out the confirmation of the theoretica considerations the eucidation of contradictions and the evauation of a new method based on the remova of vesses from the measured profie to make possibe the evauation of surface roughness of big porous species. First for measurements resuting primary P profie the estabishment of roughness parameters was necessary based on the R parameters. The estabishment and examination of P parameters is justified by the fact that the practica measuring method by touch by hand was to be interocated. The adjacent profie to the tactie evauation is the primary P profie. In order to compare the P parameters with R ones the remova of vesses and waviness fitration were performed. This way both the P z and R z parameters were investigated for the same profie and their correation meanwhie the waviness was separated by means of a so caed robust fiter to avoid the distortions on R profie caused by the traditiona Gauss fiter. For the equiong profies the W z parameter was cacuated estabished in the same way as Rz. There was examined the dependence of R z and P z from the sanding grain size for beech sampes and a simiar dependence was expected for other species too after the vesses were removed. The P z /R z quotient and its reation to W z were investigated in search for a genera context. Further more their reation was investigated for different type of sanding machines on acacia sampes. To characterize the P profies P z parameter the P =W z +R z sum was investigated and their reation. Taking into account that the measurements need to be carried out for severa times ti the reative sope is minimized the effect of more measurements carried out in the same trace was investigated to eucidate the destruction effect of the tip. The surface roughness of big porous species for different grain size was estabished after vesse remova to justify that after vesse fitration the roughness of the ground tissue (Monár, 1999) is in good correation with the grain size. Contrary no correation can be justified! II. The method of research short summary The investigations were performed mainy in abor circumstances and ony some auxiiary tests were performed in factory. Sanding of sampes took pace in the training workshop of the University of West Hungary by portabe sanding machine with SIA brand sanding canvas 75 x 533 mm, fufiing the ISO 4586- with different grain size: 6, 8, 1, 1, 15, 18,, 4, 8, 3, 4, 5 and 6. Before use they were conditioned 7 hours on 3± o C and 45% reative humidity. The korund grit was fixed eectrostatic and by means of resins on the highy fexibe canvas. The dust from the sampes was brushed away by a fine haired brush connected to a vacuum ceaner. For measurements performed in the same trace: For a species sampes sanded with 15 and 18 were prepared being the most frequent used grain sizes in the practice. In the same trace measurement repeated by 1 times were performed to examine how the roughness changes. Based on practica experience after 3-4 measurements the reative sope can be eiminated. On acacia repeated measurements were conducted by times. For the comparison of R z and P z sampes produced on portabe and contact sanding machines were prepared 5 sampes for each 1, 15 and 18 grain size. Wood Grain size species 6 8 1 1 15 18 4 8 3 4 5 6 Acacia 5 5 5 5 5 5 5 5 5 5 5 5 5 Oak 5 5 5 5 5 5 5 5 5 5 5 5 5 Ash 5 5 5 5 5 5 5 5 5 5 5 5 5 Birch 5 5 5 5 5 5 5 5 5 5 5 5 5 Beech 5 5 5 5 5 5 5 5 5 5 5 5 5 The chosen measuring instrument: Perthometer S3P is a styus tip instrument with 5 m radius. Connecting the measuring instrument with a computer by RSC 3 cabe it makes possibe the data export by means of an auxiiary ASCII program. The auxiiary program was needed because the measuring instrument transmits the R roughness profie printed on paper, the computed parameters in digita form on monitor so as the unit doesn t aows the access to the measured data. The vesse fitering Curve Cut program: was written in Borand Dephi. The routines were made by Dr.Apár Tibor based on agorithms defined by the author. After caing the data the program makes possibe to figure the roughness profie, the frequency curve near different dissociations can be marked the portions to be removed and in 5 6
eective way can be setted the repacement vaues. The program draws the roughness profie and the new frequency curve after the vesses are removed and offers the possibiity for data to be saved in a convenient form. In a further optimation the program anayses in an inequaity reation the data deriving from the waviness profie and makes possibe the vesse stumps to be removed. After vesse fitration the new regression ine is computed and the edges of the removed vesses are contracted. Henceforth with the use of a robust Gaussian fiter the waviness is separated from the P profie and an R profie is generated. Whie performing the tentative measurements the traditiona Gauss fiter was ony used for demonstrative reasons. The robust Gaussian fiter did the separation of wave in order to eiminate the unfavorabe accompanying phenomenon of traditiona Gauss fiters. The unfavorabe accompanying phenomenon of the traditiona Gauss fiter: When measuring big porous species in the neighborhood of big vesses we obtain a big ampitude waviness profie due to the functioning principe of the Gauss fiter. In comparison with the rea state the roughness profie is pushed up in the neighborhood of big pores. The phenomenon appears for both waveengths but for a shorter waveength the distortion is more remarkabe. The measure of the unpeasant mutation basicay depends on vesses depth and diameter but the number of pores and their arrangement aong the measuring ength infuences it too. Fitering by means of Gauss fiter the big porous species push ups, virtua his appear on the R profie in positive direction reative to mean ine. The R profie is a distortion of the rea profie that shows materia there where in reaity doesn t exist. The different measuring instruments compute the roughness parameters based on 5-5 -1 5-5 -1 Gauss fiter (cut off,8 mm) Gauss fiter (cut off,5 mm) 7 P profi R profi W profi P profi R profi W profi R profie it s questioned if the R profie derived from the traditiona gauss fiter is suitabe for the quantitative evauation of big porous species. The distortion of R profie is caused by the waviness fitration. Needs to be investigated the necessity of fitering and the interocution possibiity of the traditiona Gauss fiter. The pushed up vesse edges appear in a positive direction. The machining faiure appears in positive direction too. In this sense the presence of vesses on profies to be fitered with traditiona Gauss fiter worsen consideraby the roughness parameter of machining. The pushed up vesse edges do appear on the frequency curve as we. Comparing the frequency curve of R and P profies its conspicuous that in the positive range there are his on the eg because of the pushed up vesses that are missing from the P profies curve. The robust Gaussian regression fiter of the ground tissue Describing the waviness of the ground tissue is equa with the probem to find the wave that doesn t press cose to vesses but maintains the characteristics of the ground tissue. To describe the waviness component the ISO/CTS 1661-31draft standard was used as forwarded to some experts at present time is under comment and review. He standard contains a method proposed by Brinkman that generates the wave by means of a robust Gaussian fiter. The term robust covers the fact that the fiter due to some theoretica considerations isn t sensibe to the so-caed outier data. The deep pores of big porous species don t contribute to the characterization of processing quaity and from this point of view they may be considered outiers. The robust Gaussian regression fiter can be defined by the faowing genera arrangement: n ( z wk ) = 1 1 75 5 5 R profies frequency curve -1-8 -6-4 - 4 δ m s k x min 8 1 75 5 5 P profies frequency curve -1-8 -6-4 - 4
With the weighting function: π x π exp n λ n ( k ) s k = x λ z a profie height in the unfitered profie w k waviness vaue at index k k the index of the position reative to the centre of the averaging function the index of the profie points s k the weighting function m - the additiona vertica weight m the index of the iteration step x the samping interva In the first step when m= the wave is assessed by the means of the traditiona Gaussian fiter and the additiona weight () =1 is appied to each data point. In subsequent iterations the vaue of is given by the condition: ( m) z ( m) w 1 ( ) if: z w c B m ( ) cb m δ = The robust agorithm appies a regression fiter iterativey to a dataset unti the mean ine is satisfactory. The robust fiter extends by an additiona vertica weighting at each data point. can take vaues between zero and one. c B is a threshod vaue given by: c otherwise ( m) ( m) B = 4. 4478 median z w Profie heights ying cose to the mean ine estabished in the previous iteration are mutipied by vaues of cose to one, so amost their fu vaue is incuded in a modified averaging function. Zero mutipies a the profie points further than c B from the mean ine. The number of the iteration steps isn t more than six due to the prescriptions of the standard. In the present 9 work three steps were performed. The waviness parameter of the robust regression arrangement: w n s k ( m+ 1) = 1 k = n = 1 s z k δ δ ( m) ( m) The introduction of the P profie in examination There are a ot of standardized parameters for the study of roughness meanwhie a part of wood surfaces can not be quaified by means of surface roughness measurement. The main goa of the present work was to get coser to the quaifying of wood processing by means of roughness measurements with the obtained resuts. Taking into account that in practice quaifying of fine worked wooden surfaces is sti done by sensing by hand the introduction of the primary P profie is justified being the cosest one to the rea surface. Anaysis is needed if the parameters of the P profie are suitabe as quaifying parameters for the surface adjudicated by hand touch. The comparison is made upon the R z and P z parameters of the same profie. The measuring instrument computes the standardized R z parameter on five evauation engths. To make the resuts comparabe the computation of P z parameters was made on five evauation engths contrary to the seven possibe. Arguments for vesse grooves to be removed Even in case of a good machining quaity may occur that depending from the depth and number of vesses unfavorabe roughness parameters resut, making impossibe the objective adjudication of the processing quaity. The diameter and depth of the vesses ater with some order of magnitude from the characteristic vaues of the ground tissue and worsen consideraby the roughness parameters computed. Due to the geometry of the styus tip the vesse was with the incination smaer than 45 o can t be assessed reaistic. The shape of the vesses visuaized on the roughness profie differs significanty from the rea one. It is casua at which ange a vesse is cut through and where is positioned reative to the surface. In this sense considering them computing with them at the evauation won t provide usefu information. It is casua the diameter and deepness and the number of vesses appearing aong the trace. This resuts that the roughness parameters can t correate with the processing parameters. 1
After vesse fitration the weaker quaity of machining can be identified as R pk in the positive range on the Abbot curve. Without vesse remova this vaue is occasionay mixed with the virtuay pushed up vesse edges. The presence of vesses affects the regression ine too. The theoretica considerations of vesse remova The vesse fitration is carried out upon the observation that the vesses appear in the negative range reative to the mean ine. This way the vesses can be identified on the frequency curve under the form of a characteristic hiy eg in the range of negative vaues. On the frequency curve the processed surface of the ground tissue is identica with the most numerous roughness parameters. The Curve Cut program for vesse remova aows us to mark and repace with convenient vaues a the data of the range to be removed. Zero repaces the data beonging to vesses to avoid the appearance of fase vaues. Frequency (pc) The frequency curve of Acacia sampe, sanded with 18 before vesse remova 1 1 8 6 4-6 -4 - Roughness (ym) The repacement of data The frequency curve of Acacia sampe, sanded with 18 after vesse remova 1 1 8 6 4-6 -4 - Roughness (ym) If the number of the data between the two consecutive vesse edges woud be maintained as originay and woud be repaced with the oca vaues of the wave this woudn t resut the remova but the fi up of vesses. The action woud resut in an unfavorabe change in Abbot curves shape, showing materia where in reaity doesn t exist. The repacement of vesse data to the oca vaues of the wave by maintaining the origina number of data woud resut a surface with too smooth portions on vesse paces. Such a repacement woud distort the vaue of Ra. We can concude that when removing the vesses it s unsuitabe to maintain the origina number of data. Frequency (pc) Taking into account that the machining quaity can be evauated on the ground tissue whie performing the fitration the data beonging to vesses are removed ti the oca vaues of the wave characterizing the ground tissue than the meeting point of the wave and vesse are unified. This way the phenomena appearing around the vesse edges, their rea push up, tear or other deterioration contributes to the characterization of processing quaity. The profies generated after vesse remova: P and R The P profie has as mean ine the regression ine. In the first step the position and the incination of the regression ine is function of number and position of deep vesses. As the roughness vaues are reated to zero needs to be defined the regression ine of the vesse fitered profie and has to be recacuated the new Y coordinates of the profie points in order to reduce the incination to zero. Ony the roughness parameters recacuated upon the newy assessed regression ine are equa with the parameters of the profie characterizing the ground tissue without big pores. The P z parameters were cacuated after the regression ine was recacuated and the profie data reinterpreted. After obtaining the vesse-removed profie the R profie was assessed by means of the robust Gaussian waviness fiter. III. The resuts of the research On beach sampes without arge opened pores between P z and R z vaues and the grain size y = y a o + n hyperboic reation was found. Contrary for x oak containing big opened pores neither P z nor R z correated with the grain size of the sanding paper. In order to make possibe the objective adjudication of processing quaity a method was deveoped for the quick and precise remova of vesses: The advantages of the method: - it s objective and quicker than the existing ones - functions for anyhow data points - it s indifferent from the way the data points were coected - the program aows us the assessment and evauation of non standardized parameters too - the program shows its advantages for ong traces 11 1
Looking for a one parametric quaifying parameter the P z, W z and P z /R z parameters and their reations were investigated. After vesse remova the P z and R z showed y = y a o + Hyperboic A kác R z értékei reation és between a csiszoószem the R z cse vaues finom and sága the n között grain fennáó size hiperboikus on acacia sampe összefüggés x hyperboic dependence from the a = 1339,548 6 grain size. This way by vesse b = 1,446 remova the expected reation was y = 3,374 4 R =,9941 obtained between P z, R z and the R =,9883 grain size. The imit roughness of im y = x the investigated sampes was im y = 3,374 x around 5 m for P z and about 3-4 m for R z. Based upon the P z /R 4 6 z A csiszoószemcse The grain size finomsága reation investigations were conducted on acacia sampes sanded on portabe and contact sanding machine to compare roughness characteristics. In the P/R reation there is a 5% difference being a characteristic vaue. The P z /R z vaue cacuated after the vesses were removed provides information about the waviness of the profie but doesn t provide any information about the roughness associated. Theoretic profies with differing roughness may produce the same P/R. To provide further information about the reation of waviness and P/R the W z parameter of waviness was cacuated. For the (P z /R z,w z )point pairs the ine with y = 1 + a x equation can be fit. The P/R reation can t be interpreted for vaues ower than 1, this resuts that the ine has the border (,1) on the eft. Growing waviness resuted a ineary grooving P/R reation for each of the investigated sampes. The content of P z was further anayzed in order to evauate the difference between the assessed P z and the cacuated P z being the sum between R z and W z (P z =R z +W z ). For the (P z /R z,w z ) point pairs the ines with the equation y = 1 + a x can be fit. The difference equas the vaue of 1 m at 1-1 m waviness vaues. The difference resuts from the definition of P z ; R z and W z. The Akác The minták reation P P z ; R z and W z are cacuated as z between '/R z értékei P z /R a huámosság z vaues and Wthe z értékei W z parameters függvényében on acacia sampes 1,8 average for 5 samping engths. This resuts that for growing Pz /Rz 1,6 waviness grows the probabiity for Pz/Rz a peak vaue of R z to appear in a 1,4 such a position on descending section of the wave that doesn t 1, contribute to the cacuation of P zi. 1, By defining the samping engths 1 3 4 5 The A Wz profiok waviness W in more than five sections the z huámparaméterei parameter of profies (ym) (m) correation between of measured and cacuated vaues woud be better. Roughness Érdesség (µm) (m) The A Pz /Rz profiok vaue P' z /R of z értékei profies 13 Looking for one parametric characteristic, investigating the P z ; R z and W z parameters the concusion was found that there is no parameter suitabe to characterize aone the state of the profie. The characterization of roughness needs two parameters (P z /R z ; R z ). Near the R z parameter that provides information about the roughness of the processed surface the P/R is an exceent parameter expressing sensua the measure in which the profie aters from the idea R profie due to waviness and expresses in the same time the rea state of the surface by the presence of the P z parameter. IV.Thesis summary of new resuts 1. A method was deveoped for the vesses to be marked and removed from the profie ti the oca vaues of the wave characterizing the ground tissue. The method is indifferent from the measuring instrument the data were coected with and from the measuring ength. Its avaiabiity for quick evauation is principay shown for ong traces. The method makes possibe the objective measurement and evauation of processing roughness of big porous species.. Searching for a quick and objective quaifying the P profie was chosen for further examination. A suggestion was made for a new regression ine to be assessed on the data remaining after vesse fitration. Ony the roughness parameters recacuated upon the newy assessed regression ine are equa with the parameters of the profie characterizing the ground tissue without big pores. It s unfavorabe to maintain the origina number of data instead the removed vesses the vesse edges shoud be unified. Investigating the P z /R z quotient in function of the W z parameter inear dependence was found. The point-pairs fit we the y = 1 + a x ine. The sope of the ine a based on my investigations is of,1 m order and depends from the components of the ground tissue and in this sense from the wood species too. 3. To the same R z vaues may beong different surface geometries so the standardized R z one parametric characteristic doesn t describe fu-scae the state of the profie. Investigating the Pz, Rz and Wz parameters I ve found that there isn t one parametric characteristic for quaifying to fufi the characterization of the state of the profie by itsef. I suggested the description of roughness to be done by two parameters: (P z /R z ; R z ). 14
4. Derives from the way of evauation that increasing the number of samping engths the P z vaue converges to P z =R z +P z. The difference between the measured P z and the cacuated P z investigated in reation with W z concuded that increasing wave resuts increasing difference, the context is inear by y = 1 + a x. The characterization of the profie ameiorates if the number of samping engths is setted in more than seven sections. 5. Using the evauated vesse fitering method on acacia, oak, ash and birch, four big porous species hyperboic dependence was found between the roughness of the ground tissue and the sanding grain size on P and R profies as we. Rising the grain size to infinite, on different species the roughness on P profie arises around 5,1 and 6,3 m, for R profie the threshod vaue arises around 3,3 and 5 m. The above vaues provide information about the anatomica roughness of the species ater from the ast one ony in measure in which the styus tip instrument approximates the rea surfaces characteristics. V. Pubications connected to the subject 1. Csiha, Cs. (1998): Feüetkeze technoógiai sorok összehasonító eemzése KIPA módszerre. (Doktori szigorat). Csiha, Cs. (1999): Wood surface evauation differenciation between vesses and other forms of roughness. Proceedings of the Fourth Internationa Conference on the deveopement of Wood Science, Wood Technoogy and Forestry Missenden Abbey. 3. Krisch, J. Csiha, Cs. (): Anaysing wood surface roughness using an S3P Perthometer and computer based data processing, Badania da mebarstwa XIII, Poznan, pp. 145-155, 4. Csiha, Cs. - Krisch, J. (): Vesse fitration a method for anaisyng wood surface roughness of arge porous species, Drevarsky Vyskum 45(1): 13-, 5. Dr. Szabó, I. Csiha, Cs. (3): Ragasztási és feüetkezeési foyamatok. Egyetemi jegyzet. 6. Csiha, Cs. Apár, T. (3): Nagyedényes fafajok feüeti érdességének értékeése. Faipar LI. Évf. 3/1., pp.:11-16. 7. Csiha Cs. (4): Measurement of wood surface roughness of big porous species. COST E18 Coatings on wood, Symposium on measurement methods, Coating Consutancy Proceedings, Copenhagen Ora presentations: 1. Csiha Cs. (1996): Bevonatok repedezése és kivátó okai. Szakmai napok Székeyudvarhey, 1996. Csiha Cs. (1997): Feüeti repedések kiértékeése számítógépes képeemz módszerre. Szakmai napok Marosvásárhey, 1997 Fina report not paced in ibrary: 1.. Síkvidéki nyár és fenyütetvények termesztésének és a faanyag minségének összefüggései. OTKA, Faanyagismerettan Tanszék (1994-1997):. High quaity products from Back Locust, OTKA, TGYI - Zóyomi Egyetem (1998-1999) 3. Piacképes bútorcsaád önáó kifejesztése a hagyományos feüetkezeés és kézi festés, patinázás és öregít technoógiák együttes akamazásáva. OTKA - Hubertusz (1999-) 4. Bevonatok minsít rendszerének kidogozása, NKFP, (1999-1) 15 16
5. Betéri termékek fejesztése hazai faanyag bázison (ragasztási, színhomogenizáási, feüetkezeési kíséretek aapján), tömörfa parketták, fronteemek eáítása, NKFP, TGYI (-3) 6. Kütéri bútorcsaád eáítása tömörfábó (akácbó) technoógiai, szerkezet- és formatervezési kérdések megodása. Küönös figyeemme a színhomogenizáásra és a környezetbarát feüetkezeésre. (-3) 17 18