Wood and Wood Products [1 ed.] 9781620819869, 9781620819739

Citation preview

Copyright © 2012. Nova Science Publishers, Incorporated. All rights reserved. Wood and Wood Products, Nova Science Publishers, Incorporated, 2012. ProQuest Ebook Central,

Copyright © 2012. Nova Science Publishers, Incorporated. All rights reserved. Wood and Wood Products, Nova Science Publishers, Incorporated, 2012. ProQuest Ebook Central,

MATERIALS AND MANUFACTURING TECHNOLOGY

Copyright © 2012. Nova Science Publishers, Incorporated. All rights reserved.

WOOD AND WOOD PRODUCTS

No part of this digital document may be reproduced, stored in a retrieval system or transmitted in any form or by any means. The publisher has taken reasonable care in the preparation of this digital document, but makes no expressed or implied warranty of any kind and assumes no responsibility for any errors or omissions. No liability is assumed for incidental or consequential damages in connection with or arising out of information contained herein. This digital document is sold with the clear understanding that the publisher is not engaged in Wood and Wood Products, Nova Science Publishers, Incorporated, 2012. ProQuest Ebook Central, rendering legal, medical or any other professional services.

MATERIALS AND MANUFACTURING TECHNOLOGY J. PAULO DAVIM - SERIES EDITOR UNIVERSITY. OF AVEIRO, AVEIRO, PORTUGAL

Copyright © 2012. Nova Science Publishers, Incorporated. All rights reserved.

Drilling of Composite Materials J. Paulo Davim (Editor) 2009. ISBN: 978-1-60741-163-5 (Hardcover) 2009. ISBN: 978-1-60876-584-3 (E-book) Artificial Intelligence in Manufacturing Research J. Paulo Davim (Editor) 2010. ISBN: 978-1-60876-214-9 (Hardcover) 2011. ISBN: 978-1-61761-564-1 (E-book) Metal Cutting: Research Advances J. Paulo Davim (Editor) 2010. ISBN: 978-1-60876-207-1 (Hardcover) 2010. ISBN: 978-1-61122-573-0 (E-book) Tribology Research Advances J. Paulo Davim (Editor) 2011. ISBN: 978-1-60692-885-1 (Hardcover)

Tribology of Composite Materials J. Paulo Davim (Editor) 2010. ISBN: 978-1-61668-319-1 (Hardcover) 2012. ISBN: 978-1-62100-999-3 (Softcover) 2010. ISBN: 978-1-61324-772-3 (E-book) Micro and Nanomanufacturing Research J. Paulo Davim (Editor) 2010. ISBN: 978-1-61668-488-4 (Hardcover) 2012. ISBN: 978-1-61942-003-8 (Softcover) 2010. ISBN: 978-1-61324-366-4 (E-book) Medical Device Manufacturing Mark J. Jackson and J. Paulo Davim (Editors) 2012. ISBN: 978-1-61209-715-2 (Hardcover)

Wood and Wood Products, Nova Science Publishers, Incorporated, 2012. ProQuest Ebook Central,

Metal Matrix Composites J. Paulo Davim (Editor) 2012. ISBN: 978-1-61209-771-8 (Hardcover) 2012. ISBN: 978-1-62257-614-2 (E-book)

Wood and Wood Products J. Paulo Davim (Editor) 2012. ISBN: 978-1-62081-973-9 (Softcover) 2012. ISBN: 978-1-62081-986-9 (E-book)

Copyright © 2012. Nova Science Publishers, Incorporated. All rights reserved.

Biomedical Tribology J. Paulo Davim (Editor) 2011. ISBN: 978-1-61470-056-2 (Hardcover) 2011. ISBN: 978-1-61470-153-8 (E-book)

Wood and Wood Products, Nova Science Publishers, Incorporated, 2012. ProQuest Ebook Central,

Copyright © 2012. Nova Science Publishers, Incorporated. All rights reserved. Wood and Wood Products, Nova Science Publishers, Incorporated, 2012. ProQuest Ebook Central,

MATERIALS AND MANUFACTURING TECHNOLOGY

Copyright © 2012. Nova Science Publishers, Incorporated. All rights reserved.

WOOD AND WOOD PRODUCTS

J. PAULO DAVIM EDITOR

New York

Wood and Wood Products, Nova Science Publishers, Incorporated, 2012. ProQuest Ebook Central,

Copyright © 2012 by Nova Science Publishers, Inc. All rights reserved. No part of this book may be reproduced, stored in a retrieval system or transmitted in any form or by any means: electronic, electrostatic, magnetic, tape, mechanical photocopying, recording or otherwise without the written permission of the Publisher. For permission to use material from this book please contact us: Telephone 631-231-7269; Fax 631-231-8175 Web Site: http://www.novapublishers.com NOTICE TO THE READER The Publisher has taken reasonable care in the preparation of this book, but makes no expressed or implied warranty of any kind and assumes no responsibility for any errors or omissions. No liability is assumed for incidental or consequential damages in connection with or arising out of information contained in this book. The Publisher shall not be liable for any special, consequential, or exemplary damages resulting, in whole or in part, from the readers‟ use of, or reliance upon, this material. Any parts of this book based on government reports are so indicated and copyright is claimed for those parts to the extent applicable to compilations of such works.

Copyright © 2012. Nova Science Publishers, Incorporated. All rights reserved.

Independent verification should be sought for any data, advice or recommendations contained in this book. In addition, no responsibility is assumed by the publisher for any injury and/or damage to persons or property arising from any methods, products, instructions, ideas or otherwise contained in this publication. This publication is designed to provide accurate and authoritative information with regard to the subject matter covered herein. It is sold with the clear understanding that the Publisher is not engaged in rendering legal or any other professional services. If legal or any other expert assistance is required, the services of a competent person should be sought. FROM A DECLARATION OF PARTICIPANTS JOINTLY ADOPTED BY A COMMITTEE OF THE AMERICAN BAR ASSOCIATION AND A COMMITTEE OF PUBLISHERS. Additional color graphics may be available in the e-book version of this book. LIBRARY OF CONGRESS CATALOGING-IN-PUBLICATION DATA Wood and wood products / editor, J. Paulo Davim. p. cm. Includes bibliographical references and index. ISBN:  (eBook) 1. Wood products. I. Davim, J. Paulo. TS843.W66 2011 674'.8--dc23 2012017844

 New York

Wood and Wood Products, Nova Science Publishers, Incorporated, 2012. ProQuest Ebook Central,

CONTENTS Preface

Copyright © 2012. Nova Science Publishers, Incorporated. All rights reserved.

Chapter 1

ix A Quantitative Method to Measure the Surface Roughness of Sanded Wood Products Lidia Gurau, Hugh-Mansfield Williams and Mark Irle

Chapter 2

Moisture Regain in Kiln-Dried Wood Minghui Zhang and Rado Gazo

Chapter 3

Cutting Energy on Wood and Wood Products Machining Alfredo Aguilera and Pierre-Jean Méausoone

Chapter 4

Chapter 5

Investigation of Optimum Parameters for Multiple Performance Characteristics in Drilling Wood Composites (MDF) Using Grey –Taguchi Method K. Palanikumar, S. Prakash and J. Paulo Davim Electrical Resistance vs Moisture Content in Four Different Types of Cork Products José R. González-Adrados, Florentino González-Hernández, Juan I. Fernández-Golfin, José L. García De Ceca, María Conde García and Francisco García Fernández

Index

Wood and Wood Products, Nova Science Publishers, Incorporated, 2012. ProQuest Ebook Central,

1

25

57

87

109

125

Copyright © 2012. Nova Science Publishers, Incorporated. All rights reserved. Wood and Wood Products, Nova Science Publishers, Incorporated, 2012. ProQuest Ebook Central,

Copyright © 2012. Nova Science Publishers, Incorporated. All rights reserved.

PREFACE Wood as an engineering material can be technically defined “as a hygroscopic, orthotropic, biological, and permeable material having extreme chemical diversity and physical complexity with structures, that vary extensively in their shape, size, properties and function”. Therefore, using wood to its best advantage and most efficiency in engineering and technology applications, specific characteristics or chemical, physical and mechanical properties must be considered. Currently, it is usual to divide the products in two classes, solid wood and composite wood products. Solid wood include, for example, applications in furniture and cabinets, flooring, shipbuilding, bridges, mine timbers and posts. Composite wood products include, for example, plywood, insulation board, oriented strand board, hard-board and particle board. However, wood has also others related products with great potential interest, namely, cork products. This book aims to provide the research and review studies on wood and wood products with special emphasis in machining aspects. The first chapter provide information on a quantitative method to measure the surface roughness of sanded wood products. Chapter 2 is focused on moisture regain in kiln-dried wood. Chapter 3 discuss cutting energy on wood and wood products machining. Subsequently, Chapter 4 deals with the delamination in drilling of wood composite boards and investigation of optimum parameters for multiple characteristics in drilling wood composites (MDF) using GreyTaguchi method. Finally, the chapter 5 is focused on electrical resistance versus moisture in four different types of cork products. The present research book can be used for final undergraduate engineering course (for example, wood, materials, mechanical, manufacturing, etc) or as a subject on wood and wood products at the postgraduate level. Also, this book can serve as a useful reference for academics, wood researchers, mechanical,

Wood and Wood Products, Nova Science Publishers, Incorporated, 2012. ProQuest Ebook Central,

x

J. Paulo Davim

manufacturing and materials engineers, professionals in areas related to the wood and wood products. The Editor acknowledge their gratitude to Nova Publishers for this opportunity and for their professional support. Finally, I would like to thank all the chapter authors for their availability for this work.

Copyright © 2012. Nova Science Publishers, Incorporated. All rights reserved.

Aveiro, Portugal J. Paulo Davim April 2012

Wood and Wood Products, Nova Science Publishers, Incorporated, 2012. ProQuest Ebook Central,

In: Wood and Wood Products Editor: J. Paulo Davim

ISBN: 978-1-62081-973-9 © 2012 Nova Science Publishers, Inc.

Chapter 1

A QUANTITATIVE METHOD TO MEASURE THE SURFACE ROUGHNESS OF SANDED WOOD PRODUCTS Lidia Gurau1, Hugh-Mansfield Williams2 and Mark Irle3 Copyright © 2012. Nova Science Publishers, Incorporated. All rights reserved.

1

Transilvania University of Brasov, Faculty of Wood Engineering, Wood Processing and Wood Products Design Department, Brasov, Romania 2 Trada Technology Ltd, Chiltern House, Stocking Lane, Hughenden Valley,High Wycombe, Buckinghamshire, UK 3 L'unam Université, Ecole Supérieure du Bois, Laboratoire Matériaux et Composites Bois, Nantes, France

ABSTRACT No agreed guidelines exist in wood surface metrology on how to objectively measure and evaluate the surface quality of a sanded surface. As a consequence, evaluation most commonly relies on subjective human perception rather than on quantitative measurements. This chapter presents a review of a method developed for sanded wood surfaces, which covers the choice of instrument type, the measuring resolution, the 

E-mail: [email protected].

Wood and Wood Products, Nova Science Publishers, Incorporated, 2012. ProQuest Ebook Central,

2

Lidia Gurau, Hugh-Mansfield Williams and Mark Irle minimum evaluation length, aspects of filtering and the separation of processing roughness from anatomical irregularities. Compared with previous studies in the literature, processing roughness parameters calculated in this study are not influenced by wood anatomy and therefore provide objective, quantitative references for the quality of sanding. The method was tested in this chapter to evaluate the influence of various species on roughness parameters when sanding with P120 grit size paper. Further, the method can be used to evaluate roughness parameters for different combinations of sanding variables and species to optimise the sanding process.

Copyright © 2012. Nova Science Publishers, Incorporated. All rights reserved.

1. INTRODUCTION The quality of sanding is important because it determines the final quality of a finished wood surface and influences the finishing costs. In spite of this, opinions on the quality of sanding commonly rely on human perception by visual and tactile assessments based on experience rather than on quantitative measurements (Sandak and Tanaka, 2003). Both are subjective methods, which may bias the evaluation of the quality of sanding, to the extent that some rough surfaces may be judged smooth or a smoothly sanded surface may be considered rough. An objective, quantitative evaluation of surface quality requires the use of measuring instruments to collect data from the surface, followed by a series of filtering procedures and finally, by a numerical evaluation of the surface roughness. A numerical evaluation implies the calculation of standard roughness parameters that allow comparisons to be made between different surface textures. The principal measure of the quality of sanding is the surface roughness, so a greater understanding of the effect of process parameters on surface roughness would encourage the optimisation of sanding operations. Although methods for measuring surface roughness have been standardised for homogenous materials, they are not applicable to wood, and no other specific guidelines have been developed (Krish and Csiha, 1999). Roughness represents the finer irregularities of the surface texture that are inherent in a machining process (ASME B46.1: 2009). However, profile data from any nominally flat surface contains not only roughness, but also form errors and waviness that do not characterise the processing. Form errors and waviness should be excluded from any assessment of the surface roughness. Form errors constitute large deviations from the nominal shape of the

Wood and Wood Products, Nova Science Publishers, Incorporated, 2012. ProQuest Ebook Central,

Copyright © 2012. Nova Science Publishers, Incorporated. All rights reserved.

A Quantitative Method to Measure the Surface Roughness …

3

workpiece. They may be due to internal stresses in the wood or inaccuracies in the machine/tool/workpiece system. Form errors are removed by a least-squares regression to obtain what is called the primary profile (ISO 3274: 1996). Waviness is caused by incidental variables such as machining vibration or differential shrinkage within the growth ring. Waviness is removed by numerical filtering of the primary profile. Filters are categorised by their wavelength cut-off value , which separates the wavelengths that are within the range of interest for a particular feature from those that are not (ISO 11562: 1996). The line corresponding to the wavelength suppressed by the profile filter is called the mean line. Filtering wood surface data is complex because wood contains specific anatomical structures that create a surface texture independent of any processing. When this anatomical roughness is greater than the roughness due to processing, it creates distortions when processing data with filters from standards most commonly cited in the literature, since they only anticipate roughness due to a machining process. Such standard profile filters in ISO 11562 (1996), ISO 13565-1 (1996) and ASME B46.1 (1995) were found to introduce a type of distortion known as “push-up” (Krish and Csiha, 1999), especially in areas with grouped pores (Gurau et al., 2005), as well as end effects in the first and last half cut-off lengths of the profile (Figure 1). The distorted profile may be compared with a profile with no distortion (Figure 10). Irrespective of the distortion, anatomical irregularities can obscure the pattern of the processing roughness, particularly where a fine grit size has been used. A proper evaluation of the quality of sanding requires anatomical irregularities to be excluded from the roughness data (Westkämper and Riegel, 1993; Gurau et al., 2007). This separation leads to the processing roughness profile. Roughness parameters can be calculated from processing roughness profiles that allow comparisons to be made between different surfaces. If these parameters are to be useful, they must be repeatable, which implies some standardisation of factors affecting their measurement and calculation. Such factors include the measuring instruments, measuring and filtering methods and the choice of standard or non-standard parameters. A detailed set of recommendations for accurately measuring and evaluating the processing roughness of sanded wood surfaces was developed by Gurau (2004). This chapter contains a review of the proposed method and the effect on sanding roughness of varying the species.

Wood and Wood Products, Nova Science Publishers, Incorporated, 2012. ProQuest Ebook Central,

4

Lidia Gurau, Hugh-Mansfield Williams and Mark Irle

Copyright © 2012. Nova Science Publishers, Incorporated. All rights reserved.

Figure 1. Roughness profile with “push-up” distortions and “end effects” introduced by the Gaussian filter in ISO 11652: 1996, from oak sanded with P 1000. Vertical dashed lines mark the levels of the first and last half cut-off length of the filter. (Gurau et al., 2005), (Fig. 5 © 2004, Springer Science and Business Media).

2. METHODOLOGY The method described here addresses the choice of measuring instrument, the recommended measuring direction, measuring resolution and evaluation length. For data evaluation, the method describes ways of obtaining profiles free of distortions, and separating the processing roughness from anatomical irregularities.

2.1. Measurement Variables 2.1.1. The Choice Of Instrument Type Taylor Hobson instrument, TALYSCAN 150, was used that could apply two of the most common measuring techniques, laser triangulation and stylus scanning, with a single handling of the specimen. Since only the scanning head was changed, this instrument offered the advantage of inspecting exactly the same area with both methods. Their suitability for wood surfaces was evaluated in terms of their repeatability and their ability to detect peaks and valleys.

Wood and Wood Products, Nova Science Publishers, Incorporated, 2012. ProQuest Ebook Central,

A Quantitative Method to Measure the Surface Roughness …

5

Copyright © 2012. Nova Science Publishers, Incorporated. All rights reserved.

Figure 2. Comparison of stylus and laser surface images on an oak area of 1 mm x 1 mm.

The stylus was better able to detect surface irregularities and was more accurate than the laser triangulation device, which has a tendency to smooth the surface profile irregularities (Figure 2). Consecutive measurements of the same profile showed that stylus is more repeatable than the laser (Figure 3). Those characteristics made the stylus more reliable in meeting the objectives of this research in that it was better able to separate processing and anatomical irregularities (Gurau et al., 2001). However, the stylus was significantly slower than the laser triangulation, and so this may preclude it from in-line quality control.

2.1.2. The Choice of Measuring Direction The influence of measuring direction was examined on the roughness parameters Ra, Rq, Rz and Rt from ISO 4287: 1997, calculated from an oak surface sanded along the grain with P60 grit. The roughness parameters were adapted for wood in that they were calculated over the entire evaluation length rather than shorter sampling lengths. The evaluation length is restricted by the capacity of the measuring instrument, so its division into sampling lengths, as instructed by ISO 4287, leads to data sets that do not represent the variation of the wood surface (Gurau, 2004). The surface was measured with sequential scans across the sanding marks, which is more representative and more reproducible than along the marks, (Richter et al., 1995; Faust and Rice, 1986; Hiziroglu, 1999), but the roughness parameters were evaluated both along and across the sanding marks.

Wood and Wood Products, Nova Science Publishers, Incorporated, 2012. ProQuest Ebook Central,

6

Lidia Gurau, Hugh-Mansfield Williams and Mark Irle

Copyright © 2012. Nova Science Publishers, Incorporated. All rights reserved.

Figure 3. Repeatability of measurements with laser and stylus evaluated with area mean and height parameters (ASME B46.1: 2009) on a spruce surface sanded with P60.

Wood surfaces should be measured in the direction that gives the maximum values of the irregularities (Triboulot, 1984). Parameters across the grain were higher and had lower standard deviations than those along the grain. In a single factor ANOVA test, the roughness parameters were significantly higher across the grain than along the grain at the 1 % significance level. This is in agreement with the results obtained by Huang and Chen (1992). The variation among profiles along the grain is due to the variable depth of the anatomical features and of the grit marks. A measuring direction across the sanding marks was more meaningful, and it was also suitable for further separation of sanding marks from wood anatomy.

2.1.3. The Choice of Measuring Resolution A high resolution provides a very detailed data set that can be filtered later, but the time for scanning and data processing is significantly higher than for a low resolution. The best resolution is the lowest resolution that still allows an accurate evaluation of roughness parameters. The effect of varying the resolution was investigated on beech and spruce specimens sanded with P1000 grit size and oak specimens sanded with P1000 and P120 grit size, scanned at 1 m resolution. It was assumed this resolution captured all the anatomical and processing details, subject to the limitations caused by the geometry of the stylus and the precision of the instrument.

Wood and Wood Products, Nova Science Publishers, Incorporated, 2012. ProQuest Ebook Central,

A Quantitative Method to Measure the Surface Roughness …

7

Copyright © 2012. Nova Science Publishers, Incorporated. All rights reserved.

Figure 4. Comparison of roughness parameters for profiles measured along and across the sanding direction, oak sanded with P60.

Lower resolutions of 2, 5, 10, 20, 50 and 100 m were obtained as subsets of the original data. A visual comparison of decreasing the resolution and the consequent distortion is given in Figure 5 and Figure 6 for oak sanded with P120.Since the datasets were from the same surfaces and differed only in their resolution, the effect on the roughness parameters of choosing different resolutions could be clearly observed. For resolutions lower than 1 m the error for each parameter was calculated in percentage terms. The parameters Ra, Rq, Rsk, Rku were calculated as in ISO 4287:1997, but over the evaluation length. RSm from the standard was modified and is renamed RSmw in this paper. RSmw differs from RSm in that the minimum height and spacing requirements for a profile element are disregarded. If they are not, then the width and depth of the anatomical features can obscure the processing features. The other calculated parameters were Rk, Rpk and Rvk from ISO 13565-2: 1996. The roughness parameters were also presented as figures for each grit and species combination. Figure 7 shows an example of oak sanded with P120. The data have been normalised on the y axis relative to the parameter Ra and the x axis has a logarithmic scale. This method of presentation allows an overview of the variation of the parameters. The value of any roughness parameter R on the y axis is given in [1] as a function of Ra measured at a resolution of 1 m. Figure 8 shows variation of RSmw, the spacing parameter.

Wood and Wood Products, Nova Science Publishers, Incorporated, 2012. ProQuest Ebook Central,

8

Lidia Gurau, Hugh-Mansfield Williams and Mark Irle

Copyright © 2012. Nova Science Publishers, Incorporated. All rights reserved.

Figure 5. The effect of decreasing the resolution to 5 m; oak sanded with P120 (dotted line - resolution 1 m, solid line - resolution 5 m).

Figure 6. The effect of decreasing the resolution to 50 m; oak sanded with P120 (dotted line - resolution 1 m, solid line - resolution 50 m). Wood and Wood Products, Nova Science Publishers, Incorporated, 2012. ProQuest Ebook Central,

Copyright © 2012. Nova Science Publishers, Incorporated. All rights reserved.

A Quantitative Method to Measure the Surface Roughness …

9

Figure 7. The influence of measuring resolution of an oak surface sanded with P120 on normalised roughness parameters (x-logarithmic scale, y-normalised data; Ra1-value of Ra at resolution 1 m). All values are in m.

R  f Ra 1   Ra 1  i  1  R1  Ra1f (Ra1)R1 i-

[1]

Ra parameter measured at a resolution of 1 m. value of a roughness parameter R normalised relative to Ra1. roughness parameter R measured at a resolution of 1 m. rank marking the value of resolution

It was found that although the resolution was sensitive to the grit size, a value of 5 m gave the smallest percentage error and was reliable enough to be recommended for measuring wood surfaces sanded with commercial grit sizes (Gurau, 2004).

Wood and Wood Products, Nova Science Publishers, Incorporated, 2012. ProQuest Ebook Central,

Copyright © 2012. Nova Science Publishers, Incorporated. All rights reserved.

10

Lidia Gurau, Hugh-Mansfield Williams and Mark Irle

Figure 8. The influence of measuring resolution of an oak surface sanded with P120 on the spacing parameter RSmw (x-logarithmic scale). All values are in m.

2.1.4.The Choice of Evaluation Length The reliability of the evaluation of any roughness parameter depends on the length of the profile that is evaluated. A long evaluation length increases the reliability of the roughness parameters (ISO 4288: 1996) since it increases the probability of recording a profile that contains the variation of the surface. The maximum evaluation length depends on the capacity of the measuring instrument. It was found that wood does not comply with the evaluation length requirements of the general standard ISO 4288 (1996) because of its variable anatomy. The evaluation length is associated with the cut-off value of the filter, which is the same as the sampling length (ISO 4287: 1997). According to this standard, a minimum of five sampling lengths is required within an evaluation length. Therefore, ISO 4288: 1996 recommends an evaluation length of 12.5 mm for a cut-off length of 2.5 mm. Gurau et al. (2006) found

Wood and Wood Products, Nova Science Publishers, Incorporated, 2012. ProQuest Ebook Central,

A Quantitative Method to Measure the Surface Roughness …

11

Copyright © 2012. Nova Science Publishers, Incorporated. All rights reserved.

2.5 mm as a suitable cut-off length for common grit sizes, but, the evaluation length of 12.5 mm does not seem to be long enough for wood. The sensitivity of the roughness parameters Ra, Rk and RSmw, calculated as above, to the evaluation length was investigated on profiles from tangential surfaces of oak and spruce sanded with P120 grit. The roughness parameters were initially calculated over a 5 mm length, taken as the first 5 mm of the profile. The evaluation length was gradually increased to 50 mm. Figure 9 shows that all roughness parameters for spruce were unstable at the standard evaluation length and tend to stabilize towards a value of 50 mm. Similar observations were made for oak. Note that the ordinate values for RSmw and Rk are not their real values; to ease the comparison they were normalised relative to Ra. A value of 50 mm seems reasonable for wood because the amount of variation of the roughness parameters stabilised. This value was also recommended by Ostman (1983) and Richter et al. (1995), but without any published justification.

Figure 9. The influence of the evaluation length on the roughness parameters Ra, Rk and RSmw; spruce sanded with P120. In the equation f (Ra5) - the representation of the roughness parameter on the y axis; Ra5 - value of Ra at 5 mm evaluation length; R5 value of any roughness parameter R at 5 mm evaluation length; i - rank marking the length of evaluation.

Wood and Wood Products, Nova Science Publishers, Incorporated, 2012. ProQuest Ebook Central,

12

Lidia Gurau, Hugh-Mansfield Williams and Mark Irle

2.2. Wood Surface Evaluation 2.2.1. Form Error Removal According to ISO 3274 (1996), form errors can be removed by fitting a polynomial regression through the original data. The primary profile is obtained by subtracting the regression line from the original data. For wood surfaces it was found that the regression was adversely affected by the presence of deep pores under a smooth plateau, and particularly, by grouped pores (Gurau et al., 2009). This shortcoming can be avoided by removing the deep pores prior to applying the regression. However, assuming that the regression is the best fit of the total profile, the standard method in ISO 3274: 1996 may be an appropriate method of obtaining the primary profile for further processing because computationally it is less expensive than the method with pores removal, accepting that a small error will remain at the ends of the roughness profile (Gurau et al., 2009).

Copyright © 2012. Nova Science Publishers, Incorporated. All rights reserved.

2.2.2. Filtering the Primary Profile 2.2.2.1. The Choice of the Filter The “push-up” and end effects produced by ISO 11562 (1996) and ISO 13565-1 (1996) can be fully corrected, as Figure 10 shows when compared with Figure 1.

Figure 10. Roughness profile with no distortions, oak sanded with P1000. (Gurau et al., 2005), (Fig. 10 © 2004, Springer Science and Business Media). Wood and Wood Products, Nova Science Publishers, Incorporated, 2012. ProQuest Ebook Central,

A Quantitative Method to Measure the Surface Roughness …

13

Copyright © 2012. Nova Science Publishers, Incorporated. All rights reserved.

A number of profile filters were examined and the one that introduced the least distortion was the Robust Gaussian Regression Filter (RGRF), described in a new standard ISO/TS 16610-31 (2010). RGRF is a modification of the Gaussian filter from ISO 11562 (1996) and is applied iteratively to a data set until a convergence condition is met. This was tested and found suitable for wood surfaces when the standard was in draft (Fujiwara et al., 2004, Gurau, 2004). Since then, the RGRF was successfully applied to wood surfaces by other researchers including Hendarto et al. (2005) and Tan et al. (2010).

2.2.2.2. The Choice of the Filtering Cut-Off Length The cut-off length has a significant effect on the calculation of roughness parameters of sanded wood surfaces, so it is important to choose the correct cut-off length, which may vary with grit size and species. The general standard ISO 4288:1996 provides recommendations for selecting the cut-off length of the roughness filter in conjunction with the estimated values of the parameters Ra and Rz. For a homogeneous surface, the range of roughness parameters in this standard normally reflects the processing roughness. For a wood surface, which contains both processing and anatomical irregularities, the selection of a suitable cut-off length becomes difficult, because the processing roughness is often obscured by larger irregularities characterising the anatomy. Gurau et al. (2006) tested forty three cut-off values ranging from 0.025 mm to 40 mm on various sanded species. The cut-off length of 2.5 mm seemed suitable when filtering profiles surfaces sanded with commercial finishing grit sizes with the RGRF. Finer processing may require larger cut-off lengths to overcome the distorting effect of deep pores (Gurau et al., 2006). 2.2.3. Separating the Processing Roughness from the Wood Anatomy Once a roughness profile is free of any distortions, the Abbot-curve defined in ISO 13565-2 (1996) is the most appropriate starting point for devising a separation method since it is a straightforward tool for calculating the distribution of the profile heights (Gurau, 2004). The method of separating the processing roughness from wood anatomy was described in detail by Gurau et al. (2005). In Figure 11, the Abbot-curve is constructed by sorting the profile data in descending order. Statistically outlying peaks and valleys appear as non-linear regions in the Abbot-curve, and can be excluded. The upper and lower points of abrupt change in the local curvature of the Abbot-curve were identified by monitoring the variation of its second derivatives (Figure 11). These points

Wood and Wood Products, Nova Science Publishers, Incorporated, 2012. ProQuest Ebook Central,

14

Lidia Gurau, Hugh-Mansfield Williams and Mark Irle

were taken to mark the thresholds for the core data (Figure 12). Note that in Figure 11, since the data is evenly spaced along the 50 mm evaluation length, 50000 µm is 100% of the data and 25000 µm (for example) is 50% of the data.

Copyright © 2012. Nova Science Publishers, Incorporated. All rights reserved.

Figure 11. Detection of the lower and upper thresholds in the Abbot curve, oak sanded with P1000. The y-axis shows the profile data, while the x-axis is the length of the profile. (Gurau et al., 2005), (Fig. 12 © 2004, Springer Science and Business Media).

Figure 12. Separation of the core roughness profile from wood anatomy, oak sanded with P1000. (Gurau et al., 2005), (Fig. 13 © 2004, Springer Science and Business Media).

Wood and Wood Products, Nova Science Publishers, Incorporated, 2012. ProQuest Ebook Central,

A Quantitative Method to Measure the Surface Roughness …

15

Copyright © 2012. Nova Science Publishers, Incorporated. All rights reserved.

Figure 13. Processing roughness of an oak profile sanded with P1000. (Gurau et al., 2005), (Fig. 14 © 2004, Springer Science and Business Media).

Processing roughness was defined as the core roughness of a profile where the outlying peaks and valleys have been replaced with zeros. The anatomical roughness was taken as the valleys below the lower threshold, while the peaks above the upper threshold represent the fuzziness. Fuzziness is caused by groups of fibres that are attached to the surface at only one end; it varies with species, density and moisture content and to a lesser extent with processing. Figure 13 shows the processing roughness of an oak profile sanded with P1000, after the separation. Note the small range on the y axis for this fine grit size.

2.2.4. Calculation of Processing Roughness Parameters The general standards ISO 4287 (1997) and ISO 13565-2 (1996) give a variety of quantitative measures of surface roughness. A single value of these parameters is defined on a nominal interval called the sampling length. The length used for assessing the profile is called the evaluation length, which in general should contain five sampling lengths. However, given the variability in wood anatomy, roughness parameters calculated over the evaluation length are recommended for a wood surface as they were found to be more reliable than those defined on sampling lengths (Gurau, 2004).

3. EXPERIMENTAL METHOD The method reviewed above was previously used to test the effect of various grit sizes on oak surfaces (Gurau et al., 2007). In this chapter, this

Wood and Wood Products, Nova Science Publishers, Incorporated, 2012. ProQuest Ebook Central,

Copyright © 2012. Nova Science Publishers, Incorporated. All rights reserved.

16

Lidia Gurau, Hugh-Mansfield Williams and Mark Irle

method is applied to species of different anatomies and densities to see how they influence the surface roughness. Two softwoods were selected: spruce (Picea abies L.) as a low density species and Scots pine (Pinus sylvestris L.) as a high density softwood. The selection of hardwoods contained oak (Quercus robur L.) and ash (Fraxinus excelsior L.) as ring porous species, beech (Fagus sylvatica L.) as a diffuse porous species and mahogany (Swietenia macrophylla King) as a tropical species with seasonal growth increments. The specimens were conditioned to a stable moisture content of approximately 12 % by storage in a climate controlled environment. Two replicates of each species were prepared from different boards. The ash and pine specimens were prepared from radial faced boards, while the specimens of all other species were tangential faced. The specimens were pre-sanded parallel to the grain with a wide belt sander, firstly with a P60 grit size followed by P80, to remove the irregularities from previous sawing and planing operations. Then the specimens were cut to surface dimensions of 100 mm × 90 mm, suitable for final sanding with P120 paper. Final sanding was carried out in the laboratory on a Makita 9402 portable belt sander. The sander was inverted and mounted on a solid base, and a stiff frame was constructed around it. The specimen was held rigidly at all times on the belt. The sanding was performed with aluminium oxide closed-coated cloth belts measuring 600 x 100 mm. The processing was conducted at a constant contact pressure of 0.0032 N/mm2 and a belt speed of 5 m/s, the fastest speed on this machine. Before the specimens were sanded, the new sanding belts were dulled by continuous sanding for 30 minutes, to remove the initial sharpness of the abrasive grits. Fresh belts result in high roughness values that are not representative of the process. The surface measurements were carried out on the TALYSCAN 150 at 3M, Atherstone, UK. The scanning head was a stylus with 2.5 m tip radius and 90º tip angle, which moved across the surface perpendicular to the sanding marks at a speed of 1000 m/s. To analyse the influence of species, six areas of 2.5 mm × 50 mm were randomly selected from the surfaces of the two specimen replicates. Each area contained 5 profiles scanned on a length of 50 mm, which made a total of 30 profiles for a specific sanding variable investigated. Each profile was recorded at a resolution of 5 m, while the gap between profiles was 500 m.

Wood and Wood Products, Nova Science Publishers, Incorporated, 2012. ProQuest Ebook Central,

Copyright © 2012. Nova Science Publishers, Incorporated. All rights reserved.

A Quantitative Method to Measure the Surface Roughness …

17

Data was stored in ASCII format and processed with algorithms written in MathCad™. Form errors were removed with a 2nd order polynomial regression, which proved to be the best fit of the initial data. The roughness profiles were obtained by filtering the surface with the Robust Gaussian Regression Filter with a cut-off length of 2.5 mm, which produced undistorted profiles. The processing roughness was separated from the other irregularities of the surface as described above. Peaks and valleys that were not part of the processing roughness were replaced with zeros, which were excluded in the calculation of roughness parameters. The processing roughness was evaluated with various roughness parameters, including Ra, Rq and Rt from ISO 4287 (1997), and Rk from ISO 13565-2 (1996). The species density was determined from measurements taken at zero moisture content to allow comparisons and interpretation between species. Grit particles penetrate wood as in a hardness test perpendicular to the grain. The hardness test was not performed for the set of specimens, however, Brinell hardness values in Table 2 were taken from Wagenführ (2000) for comparison with the surface roughness parameters. The size of the Brinell indenter is more suitable for this comparison than the larger Janka indenter commonly used for wood hardness.

4. RESULTS AND DISCUSSION From Figure 14 it can be seen that the roughness values Ra, Rq, Rk and Rt for pine and mahogany were similar as were those for oak and beech. It was noted that their densities had a similar pattern since pine had a density 0 = 585 kg/m3, that was very close to mahogany with 0 = 554 kg/ m3; oak had a density 0 = 632 kg/m3, which was close to beech with 0 = 697 kg/ m3 (Table 1). Rq, Ra and Rk are parameters that depend on the core roughness rather than peaks and valleys. They show in Figure 14 that the spruce was the roughest species and it also had the lowest density 0 = 396 kg/m3 (Table 1). It was followed by pine and then mahogany. Oak and beech were smoother than mahogany and had very close values. The species with the smallest roughness was ash, which also had the highest density, 0 = 703 kg/m3.

Wood and Wood Products, Nova Science Publishers, Incorporated, 2012. ProQuest Ebook Central,

18

Lidia Gurau, Hugh-Mansfield Williams and Mark Irle

Copyright © 2012. Nova Science Publishers, Incorporated. All rights reserved.

Figure 14. Roughness parameters for various wood species. Values represent mean roughness parameters (m).

Table 1. The absolute density of tested species. Values represent kg/m3 Parameter Absolute density

Spruce 396

Pine 585

Species Mahogany Oak 554 632

Beech 697

Ash 703

From an anatomical point of view, species with large and grouped pores and wide medullary rays maybe expected to suffer more frequent micro fractures than homogeneous species, considering that these areas have low resistance to sanding stresses. On the other hand, low density species, however homogeneous, will have more fuzziness on their surfaces than high density species. Oak and ash had similar anatomies being both ring porous, but ash, which had a higher density, was smoother. Ash also appears harder than oak (Table 2).

Wood and Wood Products, Nova Science Publishers, Incorporated, 2012. ProQuest Ebook Central,

A Quantitative Method to Measure the Surface Roughness …

19

Table 2. Brinell hardness perpendicular to the grain from Wagenführ (2000). Values represent N/mm2 Parameter

Copyright © 2012. Nova Science Publishers, Incorporated. All rights reserved.

Brinell hardness

Spruce 12

Pine 13-19-24

Species Mahogany 18

Oak 34

Beech 34

Ash 37-41

Figure 15. Correlation of sanding roughness parameters with species density.

Similarly, spruce and pine, were influenced by their difference in density, as pine yielded a smoother surface than spruce. This is also supported by species differences in hardness. Mahogany and pine have different anatomies, but because they had similar densities, their roughness parameter values were very close. As far as hardness is concerned, the values for mahogany fall in the middle of the range for pine (Table 2). Similarly, beech and oak have different anatomies, but their density was close and so were the roughness parameter values. The hardness was similar for oak and beech (Table 2).

Wood and Wood Products, Nova Science Publishers, Incorporated, 2012. ProQuest Ebook Central,

20

Lidia Gurau, Hugh-Mansfield Williams and Mark Irle

Copyright © 2012. Nova Science Publishers, Incorporated. All rights reserved.

Mahogany as a tropical species with seasonal growth increments differs in anatomy from ring porous species. Although oak and ash have grouped pores that are larger than those in mahogany, which can lead to fuzziness, mahogany was rougher perhaps due to its lower density and hardness. Although the density of ash was only slightly higher than that of beech, the harder surface of ash may have caused its much lower roughness parameters. From the observations above, species density plays a much more important role in surface roughness than species anatomy. Figure 15 shows that there are good second order polynomial correlations of roughness parameters with density. The coefficient of determination, R2, was greater than 0.8 for all the parameters investigated. The Rk parameter, which measures the depth of the core profile, had the highest value R2. Previous research found Rk to be the most useful indicator of processing roughness (Gurau, 2004). Although not directly tested on the studied specimens, hardness may play as important a role as density. An absolute differentiation of surface quality based on species criteria would be difficult, because of density and hardness variation within the same species.

CONCLUSION A set of recommendations for wood surface metrology developed by Gurau (2004) contains the choice of instrument type, the measuring resolution, the minimum evaluation length, aspects of filtering and a method for the separation of processing roughness from anatomical irregularities. This was applied on various wood species sanded with P120. Compared with previous studies in the literature, processing parameters calculated in this chapter excluded the influence of wood anatomy and represent objective references for the quality of sanding. It appeared that the surface roughness of different species depends on species density rather than on their variable anatomy. Softwoods with their generally lower densities will be likely to display rougher surfaces than commercial hardwoods. Although not measured in this study, wood hardness may also be an important criterion for differentiating the roughness of different species sanded in the same conditions. The method can be further used to evaluate roughness parameters for different combinations of sanding variables and species to optimise the sanding process.

Wood and Wood Products, Nova Science Publishers, Incorporated, 2012. ProQuest Ebook Central,

A Quantitative Method to Measure the Surface Roughness …

21

ACKNOWLEDGMENT Authors would like to acknowledge Springer Science and Business Media for its kind permission for reusing figures from a previously published material (Gurau et al., 2005).

Copyright © 2012. Nova Science Publishers, Incorporated. All rights reserved.

REFERENCES ASME B46.1. (1995). Surface texture. (Surface Roughness, Waviness, Lay), ASME B46 Committee. ASME B46.1. (2009). Surface Texture. (Surface Roughness, Waviness, Lay), ASME B46 Committee. Faust, T.D., Rice, J.T. (1986). Characterizing the Roughness of Southern Pine Veneer Surfaces. Forest Products Journal 37(11/12), p.75-81. Fujiwara, Y., Fujii, Y., Sawada, Y., Okumura, S. (2004). “Assessment of wood surface roughness: A comparison between tactile roughness and three-dimensional parameters derived using a robust gaussian regression filter”. Journal of Wood Science, vol. 50 (1), p.35-40. Gurau, L., Mansfield-Williams, H., Irle, M. (2001). A Comparison of Laser Triangulation and Stylus Scanning for Measuring the Roughness of Sanded Wood Surfaces. In: B Bučar (Ed): Proc. of the 5th International Conference on the Development of Wood Science, Wood Technology and Forestry. 5th – 7th September 2001, Ljubljana. Slovenia. p.299-310. Gurau, L., Mansfield-Williams, H., Irle, M. (2005). Processing Roughness of Sanded Wood Surfaces. Holz als Roh und Werkstoff. 63(1), p.43-52. Gurau, L. (2004). The Roughness of Sanded Wood Surfaces. Doctoral thesis. Forest Products Research Centre. Buckinghamshire Chilterns University College. Brunel University. Gurau L., Mansfield-Williams H., Irle M. (2006). Filtering the roughness of a sanded wood surface. Holz als Roh und Werkstoff, vol. 64(5), p. 363-371. Gurau, L., Mansfield-Williams, H., Irle, M. (2007). Separation of Processing Roughness from Anatomical Irregularities and Fuzziness to Evaluate the Effect of Grit Size on Sanded European Oak. Forest Products Journal. 57 (1-2). p.110-116. Gurau L., Mansfield-Williams H., Irle M., Cionca M. (2009). Form error removal of sanded wood surfaces. European Journal of Wood and Wood Products (Holz als Roh und Werkstoff), vol. 67(2), p.219-227.

Wood and Wood Products, Nova Science Publishers, Incorporated, 2012. ProQuest Ebook Central,

Copyright © 2012. Nova Science Publishers, Incorporated. All rights reserved.

22

Lidia Gurau, Hugh-Mansfield Williams and Mark Irle

Hendarto B., Shayan E., Ozarska B., Carr, R. (2005). Analysis of roughness of a sanded wood surface. International Journal of Advanced Manufacturing Technology, 28 (7/8), p.775-780. Huang Y., Chen S. (1992). Belt sanding of parquetted veneer-overlay board. Research Report Quarterly of Institute of Forestry (translation from Chinese), 7(2), p. 139-148. Hiziroglu, S. (1999). Surface Roughness Evaluation of Medium Density Fibreboard Manufactured in Malaysia. Journal of Tropical Forest Products, 5(1), p.93-97. ISO 11562 (1996 + Cor 1: 1998). Geometrical product specifications (GPS) – Surface texture: profile method. Metrological characteristics of phase correct filters. International Organization for Standardization. ISO 13565-1. (1996 + Cor 1: 1998). Geometrical product specifications (GPS) – Surface texture. Profile method. Surfaces having stratified functional properties. Part 1: Filtering and general measurement conditions, International Organization for Standardization. ISO 13565-2. (1996+ Cor 1: 1998). Geometrical product specifications (GPS) – Surface texture: Profile method. Surfaces having stratified functional properties. Part 2: Height characterisation using the linear material ratio curve, International Organization for Standardization. ISO 3274. (1996 + Cor 1: 1998), Geometrical product specifications (GPS) – Surface texture. Profile method. Nominal characteristics of contact (stylus) instruments. International Organization for Standardization. ISO 4287. (1997 + Amd1: 2009), Geometrical product specifications (GPS). Surface texture. Profile method. Terms. Definitions and surface texture parameters. International Organization for Standardization ISO 4288 (1996 + Cor 1: 1998). Geometrical product specifications (GPS) – Surface texture. Profile method. Rules and procedures for the assessment of surface texture. International Organization for Standardization. ISO/TS 16610-31. (2010). Geometrical product specification (GPS) – Filtration. Part 31: Robust profile filters. Gaussian regression filters, International Standards Organisation. Krish, J., Csiha, C. (1999). Analysing Wood Surface Roughness Using an S3P Perthometer and Computer Based Data Processing. In:Proc. XIII Sesja Naukowa “Badania dla Meblarstwa”. Poland. p.145-154. Ostman, B.A.L. (1983). Surface Roughness of Wood-based Panels after Ageing. Forest Products Journal 33(7/8), p.35-42.

Wood and Wood Products, Nova Science Publishers, Incorporated, 2012. ProQuest Ebook Central,

A Quantitative Method to Measure the Surface Roughness …

23

Copyright © 2012. Nova Science Publishers, Incorporated. All rights reserved.

Richter, K., Feist, C., Knaebe, M.T. (1995). The Effect of Surface Roughness on the Performance of Finishes. Part 1. Roughness Characterisation and Stain Performance. Forest Products Journal 45(7/8), p.91-97. Sandak, J., Tanaka C. (2003). Evaluation of surface smoothness by laser displacement sensor 1: effect of wood species. Journal of Wood Science, 49, p.305-311. Tan, P.L, Sharif, S., Sudin, I. (2010). Roughness models for sanded wood surface. Wood Science and Technology, Published on-line: 6 Oct. DOI: 10.1007/s00226-010-0382-y. Triboulot, P. (1984). Reflexions sur les Surfaces et Mesures des États de Surface du Bois. Annales des Sciences Forestières 41(3), p.335-354. Wagenführ, R. (2000). Holzatlas. Carl Hanser ed. Fachbuchverlag Leipzig. ISBN 3-446-21390-2. Westkämper, E., Riegel, A. (1993). Qualitätskriterien fur Geschlieffene Massivholzoberflächen. Holz als Roh- und Werkstoff. 51(2), p.121-125.

Wood and Wood Products, Nova Science Publishers, Incorporated, 2012. ProQuest Ebook Central,

Copyright © 2012. Nova Science Publishers, Incorporated. All rights reserved. Wood and Wood Products, Nova Science Publishers, Incorporated, 2012. ProQuest Ebook Central,

In: Wood and Wood Products Editor: J. Paulo Davim

ISBN: 978-1-62081-973-9 © 2012 Nova Science Publishers, Inc.

Chapter 2

MOISTURE REGAIN IN KILN-DRIED WOOD Minghui Zhang and Rado Gazo Dept. of Forestry and Natural Resources, Purdue University, US

Copyright © 2012. Nova Science Publishers, Incorporated. All rights reserved.

ABSTRACT Although the moisture content of wood has been studied for about 100 years, researchers mainly focused on the moisture content change during wood drying. There are few studies of the relationship between ambient environment, time and moisture content changes in wood after drying. With small dried wood samples and a commercial-size package of kiln-dried dimensional lumber adsorption study, we developed data and provide guidance to help minimize further losses related to moisture regain in wood. Additionally, investigation of wood moisture on micro scale is introduced using time domain nuclear magnetic resonance technique.

2.1. WATER VAPOR ADSORPTION IN KILN-DRIED WOOD 2.1.1. Introduction The moisture content of dried lumber fluctuates over time with temperature and relative humidity changes. While many studies have been 

E-mail: [email protected].

Wood and Wood Products, Nova Science Publishers, Incorporated, 2012. ProQuest Ebook Central,

26

Minghui Zhang and Rado Gazo

conducted to investigate sorption equilibria, the overall sorption rate in wood from three directions at the same time has not been studied sufficiently. In this section, the adsorption of moisture by small kiln dried red oak samples over time at room temperature (21.1 0C) and high relative humidity (80%) is studied and a model is developed to predict the amount of water vapor adsorption over time for three dimensional kiln dried red oak samples.

X.1.2. Literature Review

Copyright © 2012. Nova Science Publishers, Incorporated. All rights reserved.

Water vapor sorption velocity in cellulose and its derivatives was first investigated by Sheppard and Newsome (1930). They explored Fick‟s diffusion law dealing with the adsorption process and described the relationship between time and water vapor adsorption accounting for maximum sorption weight. Stamm (1956) found a linear relationship between moisture change and square root of time up to two-thirds of the final change in uncoated cellophane. Similar results were obtained in wood samples filled with molten metal (Stamm 1959, 1960a). The rate of sorption of water vapor by small specimens of klinkii pine was studied in the absence of air by Christensen and Kelsey (1959). They developed the formula ((

t log( p f / pi )  constant, where t is the time required for half the change in moisture content to occur; pi and pf are the initial and final vapor pressures respectively) to approximately predict the sorption rate. Comstock (1963) investigated moisture diffusion coefficients in yellow-poplar from adsorption, desorption and steady state data. He plotted the square of fractional moisture change vs. t/L (where t is time and L is the specimen thickness) for free specimens (i.e. specimens that were allowed to hang freely in the humidity chamber, diffusion occurring from two faces instead of just one as for the cup specimens) and found apparently linear portion of the curve. He concluded that this linear relationship is true for all cases where the final equilibrium relative vapor pressure was 0.50 or less. Skarr (1972, 1988) provided several sorption theories applicable to wood. Several researchers also studied water vapor or bound water diffusion in wood. Siau (1995) summarized and developed a series of formulas both for steady-state and unsteady state moisture diffusion. Liu et al. (2001) developed an inverse moisture diffusion algorithm for determining diffusion coefficients. Baronas and Ivanauskas (2002) used a diffusion model to investigate moisture movement in porous solids in a numerical way. These studies were based on

Wood and Wood Products, Nova Science Publishers, Incorporated, 2012. ProQuest Ebook Central,

Moisture Regain in Kiln-Dried Wood

27

Fick‟s first and second laws of diffusion and the diffusion coefficient is obtained by solving differential equations with boundary conditions. Although the time factor is included in most models to calculate moisture diffusion coefficient, the relationship between sorption and time cannot be easily obtained due to the complex solution of a differential equation. In addition, some investigators doubt the notion that Fick‟s laws can be applied to water vapor or bound water diffusion in wood (Comstock 1963, Nakano 1994 and 1995). Wadsö (1994) believes that water vapor adsorption process in wood also includes non-Fickian stage due to the slow sorption process.

X.1.3. Mathematical Model

Copyright © 2012. Nova Science Publishers, Incorporated. All rights reserved.

Sheppard and Newsome (1930) applied the following integrated form of Fick‟s equation for diffusion in a slab or sheet to describe water vapor sorption by cellulose:

Q 8 Dt  1 Dt  1 Dt   1  2 [exp(( 2 )( ) 2 )  exp(9( 2 )( ) 2 )  exp(25( 2 )( ) 2 )  ...] Q1  a 2 9 a 2 25 a 2

(1) where: Q1 = maximum adsorption

Q = adsorption at time t a = ½ thickness of sheet D = diffusion coefficient Equation (1) can be transformed into: Ce  C 8 Dt  1 Dt  1 Dt   [exp(( 2 )( ) 2 )  exp(9( 2 )( ) 2 )  exp(25( 2 )( ) 2 )  ...] Ce  C0  2 a 2 9 a 2 25 a 2

(2) where: Ce - final equilibrium moisture content, %; C0 - initial equilibrium moisture content, %; C - average moisture content at time t, %; Although wood is an anisotropic media, we can still apply Fick‟s law in three dimensions as if solving an isotropic problem (Crank, 1975), so

Wood and Wood Products, Nova Science Publishers, Incorporated, 2012. ProQuest Ebook Central,

Minghui Zhang and Rado Gazo

28

C  2C  2C  2C  D( 2  2  2 ) t x y z

(3)

where: C- moisture content; t- time; D- diffusion coefficient; x- wood sample thickness (0, 2a); y-wood sample width (0, 2b); z- wood sample length (0, 2c) The solution for the diffusion equation (3) was given by Newman (1931) with diffusion calculations in drying porous solids as following:

Ce  C  A* B *C Ce  C0

(4)

where: Dt  2 1 Dt  1 Dt  )( ) )  exp( 9( 2 )( ) 2 )  exp( 25( 2 )( ) 2 )  ...] 2  a 2 9 a 2 25 a 2 8 Dt  2 1 Dt  2 1 Dt  2 B  2 [exp(( 2 )( ) )  exp( 9( 2 )( ) )  exp( 25( 2 )( ) )  ...]  b 2 9 b 2 25 b 2 8 Dt  2 1 Dt  2 1 Dt  C  2 [exp( ( 2 )( ) )  exp( 9( 2 )( ) )  exp( 25( 2 )( ) 2 )  ...]  c 2 9 c 2 25 c 2

Copyright © 2012. Nova Science Publishers, Incorporated. All rights reserved.

A

8

2

[exp((

Ce - final equilibrium moisture content, %; C0 - initial equilibrium moisture content, %; C - average moisture content at time t, %;

Q / Q  0.36

1 When the value of , all terms but the first may be neglected in equation (2) (Sheppard and Newsome, 1930). So the Equation (4) can be

Ce  C  0.363  0.0466 simplified as follows if Ce  C0 for three dimensional sorption problem:

Wood and Wood Products, Nova Science Publishers, Incorporated, 2012. ProQuest Ebook Central,

Moisture Regain in Kiln-Dried Wood

Ce  C 83  exp( t ) Ce  C0  6

29

(5)

The Equation (5) can be rearranged to:

C  Ce 

83

6

(Ce  C0 ) exp( t )

(6)

The Equation (6) can be then expressed in a more general format as:

Copyright © 2012. Nova Science Publishers, Incorporated. All rights reserved.

C  a  b exp(  t  )

(7)

where: a, b, α and β are parameters. When β = 1, the Equation (7) has the same format as Equation (6). The form of Equation (7) is similar to an empirical equation used by Hall (1957) to calculate the equilibrium moisture content for farm crops. This empirical equation can fit S-shaped or sigmoid isothermal adsorption curve when β > 1. Thus equation (7) has universal significance for data fitting.

X.1.4. Materials and Methods A clear, kiln dried red oak board was selected and planed to the thickness of 18 mm. Table 1. Kiln dried red oak sample dimension

Dimension (W * L * T)a

Group 1

Group 2

Group 3

Group 4

Group 5

Group 6

1 * 0.71b

1*1* 0.71

1*2* 0.71

1*3* 0.71

2*1* 0.71

2*2* 0.71

Zhang et al. 2007. a. W- width (inch), L- length (inch), T- thickness (inch). b. This sample group is cylinder shape. The diameter is one inch and thickness is 0.71 inches.

Wood and Wood Products, Nova Science Publishers, Incorporated, 2012. ProQuest Ebook Central,

30

Minghui Zhang and Rado Gazo

Copyright © 2012. Nova Science Publishers, Incorporated. All rights reserved.

Six groups of specimen were cut from this board within a one foot length in order to eliminate factors affecting the sorption isotherm as a result of different specimen history (Wengert and Mitchell, 1979). Each group of specimen had the same thickness but different width and length in order to see whether there is influence on sorption curve due to width and length. The specimen dimensions for each group are given in Table 1. All specimens were conditioned to the equilibrium moisture content of 7% in an environmental chamber set to 45% relative humidity and 21.1 0C. The specimens were then put in plastic bags to keep them from picking up moisture. The relative humidity in the environmental chamber was increased to 80% while temperature was kept constant at 21.1 0C. When the conditions in the environmental chamber stabilized, the specimens were removed from the plastic bags. Pins were used to support the specimens (Figure 1). During adsorption, specimens were periodically weighed until there was no apparent weight change. Moisture content was obtained by the oven-dried method.

Zhang et al. 2007. Figure 1. Specimen arrangement in the environmental chamber.

Wood and Wood Products, Nova Science Publishers, Incorporated, 2012. ProQuest Ebook Central,

Moisture Regain in Kiln-Dried Wood

31

Moisture Content vs. Time 13.5

12.5

Moisture Content (%)

11.5

10.5

9.5

8.5

Copyright © 2012. Nova Science Publishers, Incorporated. All rights reserved.

7.5

6.5 0

125

250

375

500

625

750

875

1000

1125

1250

Time (hour) group1

group2

group3

group4

group5

group6

Zhang et al. 2007. Figure 2. Moisture content vs. time scatter plot.

X.1.5. Results and Discussion Adsorption Curve Shape The amount of adsorption over time for all six groups is shown in Figure 2. The scatter plots show similar shape: at the beginning of adsorption, wood picks up moisture at a constant rate. This period is short compared to the whole adsorption process. Following the constant rate, the slope of adsorption curve decreases gradually while the moisture content increases. The vapor adsorption rate slows down as specimens pick up more and more moisture from the environment.

Wood and Wood Products, Nova Science Publishers, Incorporated, 2012. ProQuest Ebook Central,

Minghui Zhang and Rado Gazo

32

1 0.9 0.8

Fractional sorption

0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 Copyright © 2012. Nova Science Publishers, Incorporated. All rights reserved.

0

10

20

30

40

Time (square root hour) group1

group2

group3

group4

group5

group6

Zhang et al. 2007. Figure 3. Water vapor fractional sorption vs. square root time scatter plot.

After this period of slowing down, the water vapor adsorption continues at low, but approximately constant rate until the equilibrium moisture content is reached, ending the sorption process.

Water Vapor Fractional Adsorption Vs. Square Root Time Comparison A comparison of water vapor fractional adsorption E 

Ce  C vs. Ce  C0

square root of time scatter plot (Figure 3) with results from other studies (Christensen, 1960; Stamm, 1959, 1960a, 1960b; and Wadsö, 1994) shows that all of the plots have something in common, although different wood species, dimensions, air circulation speed, temperature, relative humidity,

Wood and Wood Products, Nova Science Publishers, Incorporated, 2012. ProQuest Ebook Central,

Moisture Regain in Kiln-Dried Wood

33

methods and water vapor or bound water migration direction were used. The similarities include: (1) A linear relationship between E and square root time at the beginning of the adsorption process. (2) Adsorption rate decreases sharply when samples approach the equilibrium moisture content. (3) Within the same species, the samples with higher initial adsorption rate slow down the adsorption rate when they approach the equilibrium moisture content.

Model Parameters Solution Equation (7) was used in fitting the experimental adsorption data. This was accomplished by performing nonlinear regression analyses using the NLIN procedure in the Statistical Analysis System software (SAS version 9.1).

Copyright © 2012. Nova Science Publishers, Incorporated. All rights reserved.

Table 2. Parameters fitted by SAS Sample No 1

alpha

beta

a

b

a-b 6.63

initial MC 7.10

final MC 12.61

0.1474

0.5093

12.65

6.02

2

0.1189

0.5366

12.81

3

0.1340

0.5237

12.66

4

0.0956

0.5673

5

0.1484

6

0.2020

7

6.04

6.76

7.19

12.80

(-0.06, 0.13)

5.89

6.76

7.14

12.63

( -0.13, 0.12)

13.16

5.50

7.66

7.83

13.26

( -0.18, 0.14)

0.5082

12.88

6.17

6.71

7.22

12.81

( -0.13, 0.15)

0.4627

12.27

6.06

6.21

7.06

12.29

(-0.21, 0.19)

0.0449

0.6692

12.92

5.91

7.01

7.01

12.96

(-0.12, 0.12)

8

0.0521

0.653

12.95

6.00

6.95

7.05

13.00

( -0.14, 0.08)

9

0.0593

0.6269

12.49

5.59

6.90

6.98

12.50

( -0.11, 0.07)

10

0.0758

0.5876

12.27

5.36

6.91

7.05

12.32

(-0.12, 0.11)

11

0.0379

0.659

12.81

5.85

6.96

6.91

12.69

(-0.05, 0.05)

12

0.0427

0.6508

12.80

5.87

6.93

6.95

12.79

( -0.07, 0.06)

13

0.0416

0.6488

12.53

5.51

7.02

6.98

12.49

( -0.08, 0.07)

14

0.0462

0.6363

12.28

5.23

7.04

7.04

12.28

(-0.07, 0.07)

15

0.1230

0.5450

12.84

6.13

6.71

7.21

12.96

( -0.11, 0.14)

16

0.1569

0.5094

12.50

5.87

6.63

7.17

12.59

( -0.11, 0.12)

17

0.0485

0.6527

12.70

5.73

6.97

7.04

12.70

( -0.07, 0.05)

18

0.0627

0.6124

12.53

5.53

7.00

7.12

12.50

( -0.10, 0.09)

Zhang et al. 2007.

Wood and Wood Products, Nova Science Publishers, Incorporated, 2012. ProQuest Ebook Central,

residue range (-0.14, 0.15)

Minghui Zhang and Rado Gazo

34

All samples‟ initial and final moisture content, parameters and residual range are tabulated in Table 2. Equation (7) does fit the experimental data well according to Table 2. In addition, the final moisture content is close to parameter a, and the initial moisture content is close to value a – b. Therefore, Equation (7) can be expressed as:

C  Ce  (Ce  C0 ) exp( t  )

(8)

The correlated relation (Figure 4) between parameters α and β obtained through regression analysis (adjusted R-square 0.966) shows a linear relation as follows: β = -1.29148α + 0.70412

(9)

Parameters Affecting Value of Factors in the Model In Equation (8), only parameters α and β are used.

0.7 0.65 0.6 beta

Copyright © 2012. Nova Science Publishers, Incorporated. All rights reserved.

The equilibrium moisture content can be calculated from a formula by Eckelman (1998) or Simpson (1998). The equilibrium moisture content for kiln dried lumber has a lower value than calculated due to hysteresis lag during adsorption process.

0.55 0.5 0.45 0.4 0

0.07

0.14 alpha

Zhang et al. 2007. Figure 4. Scatter plot between beta and alpha. Wood and Wood Products, Nova Science Publishers, Incorporated, 2012. ProQuest Ebook Central,

0.21

Moisture Regain in Kiln-Dried Wood

35

So it seems that it is not necessary to consider specimen dimensions as when using Fick‟s law or diffusion coefficients. The presented model can easily predict water vapor adsorption amount over time. Parameter α and β can be obtained through inverse method. However, due to the way in which Equation (4) was simplified into Equation (5), it is clear that parameters α and β are directly related to diffusion coefficient, wood species, dimensions, air circulation speed, temperature, relative humidity, and water vapor or bound water migration direction. The general format of Equation (7) gives time as a new scale with parameters α and β. This format states that water vapor adsorption is not completely controlled by Fick‟s law if the diffusion coefficient is treated as a constant value for the whole sorption process. The reason comes from the fact that water vapor adsorption in wood includes both diffusion and swelling stress relaxation. Alfrey et al. (1966) classified stages of moisture diffusion and polymer relaxation according to their relative rates as follows:

Copyright © 2012. Nova Science Publishers, Incorporated. All rights reserved.

(1) The rate of diffusion is much lower than the rate of stress relaxation; (2) The rate of diffusion is much higher than the rate of stress relaxation; (3) The diffusion and relaxation rates are comparable. Water vapor adsorption in wood includes all of the above stages because water vapor adsorption over time curve does not show any apparent two-stage behavior. The curve follows an exponential function and has a smooth and continuous shape. The adsorption rate is determined mainly by the slower process of diffusion or relaxation (Christensen, 1960). Time scale of parameters α and β combines these two processes and makes the model fit the experimental data well. When parameters α and β are strongly correlated it is possible that α or β is the function of diffusion and swelling stress relaxation. In addition, the diffusion and swelling stress relaxation are affected by factors such as temperature, relative humidity, wood species, and air circulation. These factors were outside the scope of this study.

X.1.6. Conclusion This study showed that Fick‟s second law has limitations when applied to water vapor adsorption in wood. A new time scale concept is introduced due to two processes involved in water vapor adsorption in wood – diffusion and

Wood and Wood Products, Nova Science Publishers, Incorporated, 2012. ProQuest Ebook Central,

36

Minghui Zhang and Rado Gazo

stress relaxation. Water vapor adsorption over time in wood follows an exponential function. In this research, parameters in the model were obtained by the inverse method. Parameters α and β as the function of wood species, dimensions, temperature, relative humidity and air circulation speed need to be further studied.

X.2. MOISTURE DISTRIBUTION IN A COMMERCIAL-SIZED LUMBER PACKAGE STORED IN A HIGH HUMIDITY ENVIRONMENT

Copyright © 2012. Nova Science Publishers, Incorporated. All rights reserved.

X.2.1. Introduction Most secondary wood products manufacturers use kiln dried lumber in their production. Many of them, especially small to medium sized companies, do not have environmentally controlled storage space for this lumber. Under certain conditions, kiln-dried lumber can pick up moisture from its surroundings, resulting in numerous problems during wood products manufacture and use. While a wealth of scientific information related to rate of movement of moisture in small clear specimen of wood exists, there is no reliable information about the rate of movement of moisture in a commercialsized package of kiln dried, rough, full length lumber. This is mostly due to the difficulties of obtaining repeated accurate moisture content readings from within the package of lumber over extended periods of time using traditional means such as kiln samples or moisture content meters. Recent development of wireless radio frequency moisture content sensors makes such an undertaking possible.

X.2.2. Literature Review Lumber Storage Several authors suggest that to prevent kiln or air dried lumber from regaining moisture while in storage it should be kept in closed and heated storage (Simpson, 1991). Aside from increased costs, lumber producers and furniture manufactures often do not have sufficient space for storing kiln-dried lumber in a heated shed. Other, less costly ways such as outdoor lumber storage, open-shed storage and closed, unheated storage are often used

Wood and Wood Products, Nova Science Publishers, Incorporated, 2012. ProQuest Ebook Central,

Copyright © 2012. Nova Science Publishers, Incorporated. All rights reserved.

Moisture Regain in Kiln-Dried Wood

37

(Simpson et al., 1999). These storage solutions result in various moisture adsorption rates. For example, Rietz (1978) concluded that in open sheds, lumber regains about 1 percent of moisture content per month while in closed sheds about 0.3 percent per month. Therefore, it is not necessary for dried lumber to be stored in closed sheds if storage periods are short. For longer periods of lumber storage in an unheated closed shed, however, kiln dried lumber will gradually regain moisture and may not meet final moisture content requirements. In his work, Rietz gave only the average speed of moisture content change in a package of dried lumber and did not describe moisture content distribution over time in different storage situations. Bois (1978) emphasized that woodworking and furniture plants need to pay attention to plant conditions since dried lumber may regain moisture in several days. Bois, however, did not give any particulars. Mathewson (1930) recorded changes in moisture content in different parts of a typical pile of softwood lumber during air drying. He found that moisture content varied over time throughout the pile. Regaining of moisture in dried lumber, however, is the inverse problem and the processes are different. Jenkins (1934) studied the absorption of moisture in kiln-dried lumber. He reported data on moisture content change over time for both the outside and the center of a lumber package. However, there is no information for the regions between the outside and the center of the lumber package. Shubnyi (1975) studied the distribution of moisture content in pine lumber during storage in wrapped packages. Subsequently Dobrynin and Kulakova (1980) monitored moisture content changes in lumber packages wrapped in waterproof paper and provided data on the variability of moisture content over time. Additionally, Kulakova (1983) tabulated monthly data (based on yearround measurements in storage) on the rate of increase or decrease in the moisture content of the lower layers of lumber packages. The drawback of this study is that it did not consider moisture content distribution in the entire package. A study by James et al. (1984) concluded that the moisture gradient decreased greatly within 60 hours after the material was taken out of the kiln, but the average moisture content did not change significantly. Baronas et al. (2001) developed a model of moisture movement in wood stored outdoors. This model was used to predict the moisture content in sawn boards during long term storage under outdoor climatic conditions and was tested experimentally. The sawn boards were processed into surfaced specimens 50 by 270 by 35 mm; experimental moisture content values were determined as the average of the individual specimens at the same time during

Wood and Wood Products, Nova Science Publishers, Incorporated, 2012. ProQuest Ebook Central,

38

Minghui Zhang and Rado Gazo

Copyright © 2012. Nova Science Publishers, Incorporated. All rights reserved.

storage. Thus, this model cannot be easily applied to the whole lumber package.

Moisture Content Meters In order to study the relationship between moisture content distribution and time, it is necessary to collect real-time moisture content data in a whole lumber package. Therefore, the method for measuring moisture content should be considered carefully. The American Society for Testing and Materials (ASTM, 2002) describes three methods for measuring wood moisture content, namely, moisture meters, oven drying, and distillation. Distillation mainly applies to wood treated with chemicals, and the oven-dry method while accurate throughout the whole range of moisture content, is time consuming. Moisture meters are popular due to their ease of use. There are three kinds of electric moisture meters (i.e. resistance type, power-loss type and capacitance type) used widely in the wood industries. James (1988) gave a detailed description of their operation. Wilson (1999) and Quarles (2004) concluded that these moisture meters generally give reasonable readings but are not entirely accurate. Radiofrequency and microwave moisture meters also have been developed (James, et al. 1988, Gu, 1997). Because these moisture meters need to either make contact or be in proximity to the sample to be measured, they are not practical for measuring the moisture content inside a lumber package. Consequently, radio frequency sensors for measuring moisture content remotely were developed (Dai and Ahmet, 2001, Carll and Wolde, 1996). These sensors make possible real-time remote moisture content monitoring inside a lumber package.

X.2.3. Objectives Although the moisture content of lumber has been studied for about 100 years, researchers mainly focused on the moisture content change during lumber drying. There are few studies of the relationship between ambient environment, time and moisture content changes in a lumber package after drying. Generally, losses from lumber degrade in drying range between $9-54 per thousand board feet. It is our goal to develop data and provide guidance to help minimize further losses related to moisture regain in packages after the lumber has been dried. Specific objectives are as follows:

Wood and Wood Products, Nova Science Publishers, Incorporated, 2012. ProQuest Ebook Central,

Moisture Regain in Kiln-Dried Wood

39

1) Obtain a moisture content distribution over time diagram for a package of kiln dried, red oak lumber. 2) Develop a model to describe moisture content distribution over time for a package of kiln dried, red oak lumber.

X.2.4. Materials and Methods Materials One thousand board feet of 8-foot long, FAS grade, red oak lumber dried and equalized to a moisture content of 7% was used for this study. The lumber was end coated prior to kiln-drying.

Copyright © 2012. Nova Science Publishers, Incorporated. All rights reserved.

Moisture Content Equalization Prior to the beginning of the experiment, all lumber was placed inside a 500 ft3 environmental chamber. Lumber was stacked using ¾” thick stickers and 24” spacing between stickers. Five four-inch square bolsters were placed underneath the package. The conditions in the chamber were set to an EMC of 7% for dried red oak (21.1 0C and 45% relative humidity). These conditions were maintained for three months. Moisture Content Measurement A wireless moisture content probe system manufactured by Lignomat USA, Ltd composed of sensors, transmitters and receivers was used in the following manner. At a preset time interval, the sensor measured the moisture content; transmitter converted the analog sensor signal to a digital reading and sent it to a receiver via radio frequency. The data was automatically received by a personal computer and saved in a database. Prior to the experiment, all sensors were calibrated using small specimens and the oven-dried method. Sensor Arrangement After the three month equalization phase, the stickers were removed and the lumber was arranged into a package that was 8-foot long, 44 inches wide and 34 layers high, resembling a standard way in which the lumber is shipped. At the same time, thirty nine sensors and transmitters were arranged inside the package. Small pockets had to be cut out from adjacent boards to facilitate the placement of sensors and transmitters. The coating was removed from one end of each board by cutting off 1” sections. The thirty nine sensors were divided into three groups of thirteen each.

Wood and Wood Products, Nova Science Publishers, Incorporated, 2012. ProQuest Ebook Central,

40

Minghui Zhang and Rado Gazo

.

Zhang et al. 2006.

Copyright © 2012. Nova Science Publishers, Incorporated. All rights reserved.

Figure 5. Cross section of a sensor arrangement (numbers represent sensor positions).

One group of sensors was placed 6 inches from one end of the package (end coated), another group of sensors was placed 6 inches from the other end of the package (end coating removed) and the last group of sensors was placed in the middle of the package, 4 feet from either end. The cross-sectional arrangement within the package was the same for all three groups and it is shown in Figure 5. Sensors were placed so as to include boards which have two, one or none of the surfaces exposed to the environment. Some of the sensors were in the top layer, some on the sides of the package, some just below the surface, and some were several layers deep.

Moisture Adsorption After the package was assembled, the conditions in the chamber were kept at 21.1 0C and 45% relative humidity (EMC of 7% for dried red oak) for an additional 10 days to monitor the performance of the sensors. After the ten days, conditions in the chamber were changed to 21.1 0C and 80% relative humidity. These conditions, according to the Wood Handbook (USDA Forest Service, 1999) would result in a wood EMC of 16% during moisture desorption (drying). However, due to the moisture sorption hysteresis (Skaar 1979, Peralta 1995), these same conditions resulted in an EMC of about 12% during moisture adsorption (regaining of moisture below fiber saturation point). The wireless probe system collected and recorded readings from all 39 sensors once every hour for nineteen weeks.

Wood and Wood Products, Nova Science Publishers, Incorporated, 2012. ProQuest Ebook Central,

Moisture Regain in Kiln-Dried Wood

41

X.2.5. Results and Discussion Moisture Content Distribution throughout the Package After 19 weeks and about 125,000 readings, the highest measured moisture content was in the top layer of the package. The average moisture content was 10.9% (the maximum was 11.8% and the minimum was 10.2%). This layer almost reached the EMC for the set condition. The bottom layer with an average moisture content of 9.9% (the maximum was 11.2% and the minimum was 9.1%) had the second highest MC. 11.5

Moisture Content (%)

9.5

8.5

7.5

Time (hour) 1st layer

7th layer

10th layer

13th layer

18th layer

23rd layer

Zhang et al. 2006. Figure 6. Moisture content change by layer. Wood and Wood Products, Nova Science Publishers, Incorporated, 2012. ProQuest Ebook Central,

28th layer

34th layer

3192

3024

2856

2688

2520

2352

2184

2016

1848

1680

1512

1344

1176

1008

840

672

504

336

0

6.5 168

Copyright © 2012. Nova Science Publishers, Incorporated. All rights reserved.

10.5

Minghui Zhang and Rado Gazo

42

Moisture Content (%)

12 11 10 9 8 7

3192

3024

2856

2688

2520

2352

2184

2016

1848

1680

1512

1344

1176

840

1008

672

504

336

168

0

6

Time (hour) Top edge

Top middle

Zhang et al. 2006. Figure 7. Moisture content comparison between middle and edge boards of the 1st layer.

Moisture Content (%)

11 10 9 8 7

3192

3024

2856

2688

2520

2352

2184

2016

1848

1680

1512

1344

1176

1008

840

672

504

336

168

6

0

Copyright © 2012. Nova Science Publishers, Incorporated. All rights reserved.

12

Time (hour) 1st

18th

34th

Zhang et al. 2006. Figure 8. Moisture content comparison between outside boards in the 1st, 18th, and 34th layers.

The lowest moisture content was measured in the 23rd and 28th layers. The average moisture content of these layers was 8.5%. Interestingly, the driest area was not in the center of the package, but rather about 1/3 up from the bottom. The center of the package (18th layer) had an average moisture content of 9.6% (the maximum was 10.8% and the minimum was 8.4%). The mean moisture content change over time by layer is shown in Figure 6. The mean for each layer was calculated as average of readings from all sensors in that layer. It can be observed that the first (top) layer had a

Wood and Wood Products, Nova Science Publishers, Incorporated, 2012. ProQuest Ebook Central,

43

significantly different adsorption rate when compared to other layers. The mean moisture content change in this layer was 1.25% in first 24 hours. This layer‟s moisture adsorption curve follows exponential function and is very similar to a free standing wood sample moisture adsorption curve (Comstock, 1962). For all other layers, the rate of moisture adsorption was not significantly different (alpha = 0.05 level) during first 24 hours. After 72 hours, the rate of adsorption seemed to be divided into 4 groups from high to low: 1st layer > 34th layer > 7th, 10th, 13th, and 18th layer > 23rd and 28th layer. Figure 7 shows differences in moisture adsorption between edge and middle boards in the 1st layer. The edge boards had both face and edge exposed to the environment, whereas the middle board had only a face exposed. Figure 8 shows differences in moisture adsorption between boards in the 1st, 18th, and 34th layers that had a surface (either a face or an edge) exposed to the environment. There was a statistical significant difference (alpha = 0.05) in adsorption rate during 19 weeks between all three layers, except during the first week between layers 18 and 34. Figure 9 shows differences in moisture adsorption between outside and inside boards. The outside boards had either a face or an edge exposed to the environment, whereas the inside boards did not have any surface directly exposed to the environment. The mean difference between the outside and inside readings increased during the first week and then held constant for the duration of the experiment.

11

Moisture Content (%)

10 9 8 7

Time (hour) Outside

Inside

Zhang et al. 2006. Figure 9. Moisture content comparison between outside and inside boards. Wood and Wood Products, Nova Science Publishers, Incorporated, 2012. ProQuest Ebook Central,

3192

3024

2856

2688

2520

2352

2184

2016

1848

1680

1512

1344

1176

840

1008

672

504

336

168

6 0

Copyright © 2012. Nova Science Publishers, Incorporated. All rights reserved.

Moisture Regain in Kiln-Dried Wood

Minghui Zhang and Rado Gazo

44

Effect of End Coating End coating is commonly applied to the ends of boards prior to drying. Its purpose is to slow down the loss of moisture through the ends of boards during drying and thereby minimize lumber degrade caused by end checks and splits. During the experiment, the factory applied end coating was left on one end of each board and removed from the other end. This allowed any differences during the water adsorption process to be observed. The means of moisture content readings collected from sensors at specific times for both coated and non-coated ends were compared using a paired t-test as described by Montgomery (2001). The difference estimates (coated uncoated) and the confidence intervals ( d  2S E ) over the 19 weeks are

and uncoated ends developed after 72 hours. The d gradually increased during first four weeks to approximately 0.6% and then kept constant (Figure 10). 1.6 1.4

MC Difference (%)

1.2 1 0.8 0.6 0.4 0.2 0 3192

3024

2856

2688

2520

2352

2184

2016

1848

1680

1512

1344

1176

1008

840

672

504

336

0

-0.2 168

Copyright © 2012. Nova Science Publishers, Incorporated. All rights reserved.

plotted in Figure 10. Only the inside boards not directly exposed to the environment were used for this comparison. There was no significant difference (alpha = 0.05) in moisture content change between the coated and uncoated ends during the first 72 hours. This was due to the fact that sensors were located 6 inches from the end of the board and it took about 72 hours for the moisture to migrate to this location. However, significant differences in moisture content change between coated

Time (hour)

Zhang et al. 2006. Figure 10. Difference in moisture content between coated and uncoated ends of inside boards (dashed line represents mean difference, dotted lines represent lower and upper bounds).

Wood and Wood Products, Nova Science Publishers, Incorporated, 2012. ProQuest Ebook Central,

Moisture Regain in Kiln-Dried Wood

45

The ends from which the end coating was removed had reduced rate of moisture adsorption. This result was surprising and warrants more study.

Modeling Moisture Gain over Time Potentially, wood exhibits all three diffusion stages as described by Alfrey et al. (1966) during the water vapor adsorption process. Although it may be possible to divide the adsorption process into its three stages as defined above and develop mathematical formulas to predict moisture gain over time, this would involve finding solutions to complex diffusion equations. To our knowledge, there is no mathematical model that can successfully predict the adsorption process for all three stages (Crank, 1975). It is possible, however, to fit experimental data by using statistical software such as SAS to obtain an empirical formula. Andrews and Johnston (1924) developed an integrated form of Fick‟s equation for diffusion into a slab or sheet during the study of absorption of water by rubber as follows:  2 kt

Copyright © 2012. Nova Science Publishers, Incorporated. All rights reserved.

Q 8   1  2 (e 4 Q1 

a

2

1   e 9

9 2 kt 4 a2

1   e 25

25 2 kt 4 a2

 ...) (10)

where: Q1= maximum absorption Q= absorption at time t a= 0.5 * thickness of sheet k= diffusivity constant According to Andrews and Johnston (1924), all terms but the first may be neglected when the values of Q/ Q1>0.36. Therefore, the formula (10) can be simplified and generalized to predict moisture content over time as follows: MC = a – b * EXP (-α * t^β)

(11)

where: MC= moisture content at time t a, b, α and β are parameters Formula (11) was used in fitting the experimental adsorption data. This was accomplished by performing nonlinear regression analyses using the NLIN procedure in the Statistical Analysis System package (SAS version 9.1). The residue ranges are within -0.23 and 0.22, except for the top layer. The larger residue ranges from the sensors in the top layer all occurred during the first day of water vapor adsorption.

Wood and Wood Products, Nova Science Publishers, Incorporated, 2012. ProQuest Ebook Central,

Minghui Zhang and Rado Gazo

46

Table 3. Parameters for Formula (2) of coated package end (x, y, and z are distances from the center of the package; x lengthwise, y widthwise and z heightwise)

Position

α

β

a

b

residual range

x

y

z

1

0.1633

0.3122

12.5

5.5

(-1.03, 0.20)

-42

-18

16.5

2

0.0402

0.3933

12.5

5.5

(-0.50, 0.10)

-42

0.5

16.5

3

0.0559

0.4025

12.5

5.5

(-0.51, 0.11)

-42

18

16.5

4

0.0292

0.4931

11.8114

5.3619

(-0.11, 0.14)

-42

-12

10.5

5

0.0101

0.6115

12.3477

5.5348

(-0.14, 0.11)

-42

0.5

7.5

6

0.0340

0.4413

12.6752

5.9116

(-0.14, 0.08)

-42

-6

4.5

7

0.00909

0.6185

9.7953

3.3653

(-0.10, 0.11)

-42

-18

-0.5

8

0.00536

0.7093

11.6549

4.6728

(-0.13, 0.11)

-42

-9

-0.5

9

0.0165

0.5169

12.9364

6.1413

(-0.18, 0.10)

-42

0.5

-0.5

10

0.00117

0.8460

9.5529

2.7595

(-0.07, 0.07)

-42

6

-5.5

11

0.000449

0.9524

8.9407

2.2564

(-0.06, 0.08)

-42

12

-10.5

12

0.0122

0.5946

11.6422

5.3035

(-0.18, 0.20)

-42

-18

-16.5

13

0.0209

0.6234

11.4644

5.0875

(-0.23, 0.16)

-42

18

-16.5

Copyright © 2012. Nova Science Publishers, Incorporated. All rights reserved.

Zhang et al. 2006.

The relationship between the moisture content and the square root of time for the initial adsorption is linear. This corresponds to findings of Stamm (1959) and Comstock (1962). Therefore, Formula (11) along with the parameters from Tables 3, 4 and 5 (calculated by NLIN procedure in SAS) can be used to predict moisture content over time.

Other Factors Affecting Water Vapor Adsorption in Wood In this chapter, only one species of wood under constant temperature and relative humidity was considered. The effects of wood species, air circulation, cycling temperature and relative humidity during storage should be further studied. All of these factors were outside the scope of this chapter.

X.2.6. Conclusion Based on our observations, the following conclusions can be made about moisture content changes in kiln-dried package of lumber stored in high humidity environment:

Wood and Wood Products, Nova Science Publishers, Incorporated, 2012. ProQuest Ebook Central,

Moisture Regain in Kiln-Dried Wood

47

1. The highest rate of moisture adsorption occurs in the top and bottom layers of the lumber package. 2. There is a linear relationship between the moisture content and the square root of time during first 24 hours for the top layer. 3. During first 72 hours, maximum moisture content increase is about 0.4% for the boards which are not in the top and bottom layers. 4. The most moisture adsorption (about 2% increase in MC) occurs during first 8 weeks. 5. The rate of moisture adsorption decreases with time and follows exponential function. 6. The lowest rate of moisture adsorption is concentrated in the lower third of the lumber package.

Copyright © 2012. Nova Science Publishers, Incorporated. All rights reserved.

X.3. INTRODUCTION OF TIME-DOMAIN NMR APPLIED TO MOISTURE IN WOOD Nuclear magnetic Resonance (NMR) has been a powerful measurement tool in the last decades. Besides assessing compositional and structural features, elucidating molecular structure, resolving micro-structural arrangement, it is also a quantitative analytical tool in materials science. Time-domain (TD) pulsed NMR instruments are widely used due to their relatively low cost, ease of operation and ability to provide quantitative information on molecular dynamics in materials. Since they are designed to detect receptive nuclei like hydrogen, fluorine, and phosphorus, they can be used to study water (which contains hydrogen nuclei) in wood qualitatively and quantitatively. Green wood generally contains water in three forms: water vapor, free water and bound water (Skaar 1988). Usually bound and free water in wood are studied separately since it is difficult to identify these two water forms simultaneously with traditional methods. Time-domain nuclear magnetic resonance (TD-NMR) technique can easily distinguish water states according to relaxation time and give more quantitative information on water in wood than any other method (Araujo et al., 1992). Hsi et al. (1977) found two T2 (spin-spin relaxation time) components of water in white-cedar wood powder, suggesting the existence of two different water states in their wood samples; Riggin et al. (1979) showed the possibility of at least two different relaxation times for water in wood for

Wood and Wood Products, Nova Science Publishers, Incorporated, 2012. ProQuest Ebook Central,

Minghui Zhang and Rado Gazo

48

Copyright © 2012. Nova Science Publishers, Incorporated. All rights reserved.

hydrogen nuclei of water in white spruce sapwood with moisture content from 5 to 176%; Menon et al. (1987) collected the CPMG data from Douglas fir and western red cedar wood sample until the signal was down to less than 1% of the initial amplitude and got the third T2 component. Nanassy (1976) got the linear relationship between the integrated absorption intensity of the narrow component (bound water) and the moisture content up to 22% with continuous wave NMR; Sharp et al. (1978) measured wood moisture content with Pulsed NMR techniques in the moisture content range from 0 to 176%; Merela et al. (2009) demonstrated that wood moisture content can be determined instantaneously on the basis of its mass and the amplitude of its NMR free-induction-decay (FID) signal. All the above studies show TD-NMR technique is an ideal tool to investigate moisture in wood in micro-scale. Time-domain NMR application in wood is usually through FID, T2 and T1 (spin-lattice relaxation time) measurements. FID and T2 measurements are important methods in studying moisture in wood. Figure 11 shows a typical ash heartwood FID curves with different moisture content. Table 4. Parameters for Formula (2) of uncoated package end (x, y, and z are distances from the center of the package; x lengthwise, y widthwise and z heightwise)

Position

α

β

a

b

residual range

x

y

z

1

0.0562

0.4158

12.5

5.5

(-0.60, 0.17)

42

-18

16.5

2

0.0547

0.3535

12.5

5.5

(-0.55, 0.08)

42

0.5

16.5

3

0.115

0.3185

12.5

5.5

(-0.75, 0.31)

42

18

16.5

4

0.00418

0.5903

14.16

7.657

(-0.10, 0.17)

42

-12

10.5

5

0.000982

0.7904

11.3872

4.6296

(-0.07, 0.06)

42

0.5

7.5

6

0.000229

1.0112

9.7546

2.9281

(-0.08, 0.07)

42

-6

4.5

7

0.00282

0.7259

9.5723

3.1045

(-0.09, 0.07)

42

-18

-0.5

8

0.00141

0.8336

10.0998

3.327

(-0.11, 0.10)

42

-9

-0.5

9

0.000763

0.9404

10.204

3.3685

(-0.09, 0.09)

42

0.5

-0.5

10

0.00787

0.6325

10.7968

4.1985

(-0.17, 0.22)

42

6

-5.5

11

0.00733

0.6919

10.5816

3.9706

(-0.12, 0.15)

42

12

-10.5

12

0.00441

0.7338

10.596

4.1513

(-0.19, 0.11)

42

-18

-16.5

0.6832

9.6385

2.9824

(-0.22, 0.18)

42

18

-16.5

13 0.0096 Zhang et al. 2006.

Wood and Wood Products, Nova Science Publishers, Incorporated, 2012. ProQuest Ebook Central,

Moisture Regain in Kiln-Dried Wood

49

Table 5. Parameters for Formula (2) of package middle (x, y, and z are distances from the center of the package; x lengthwise, y widthwise and z heightwise)

Position

α

β

a

b

residual range

x

y

z

1

0.0302

0.4679

12.5

5.5

(-0.48, 0.13)

0.5

-18

16.5

2

0.0604

0.328

12.5

5.5

(-0.66, 0.11)

0.5

0.5

16.5

3

0.1022

0.263

12.5

5.5

(-0.75, 0.11)

0.5

18

16.5

4

0.000826

0.8954

9.9233

3.1231

(-0.08, 0.05)

0.5

-12

10.5

5

0.000211

0.997

11.6838

4.7932

(-0.14, 0.09)

0.5

0.5

7.5

6

0.00142

0.7092

13.4659

6.5513

(-0.10, 0.07)

0.5

-6

4.5

7

0.00991

0.5486

10.4414

3.8965

(-0.10, 0.06)

0.5

-18

-0.5

8

0.0121

0.4715

13.6805

6.9948

(-0.12, 0.11)

0.5

-9

-0.5

9

0.00725

0.6297

11.5119

4.7094

(-0.15, 0.10)

0.5

0.5

-0.5

10

7.37E-07

1.3875

15.4928

8.7678

(-0.09, 0.06)

0.5

6

-5.5

11

4.77E-07

1.689

8.0884

1.3414

(-0.07, 0.06)

0.5

12

-10.5

12

0.0135

0.5191

12.0071

5.6965

(-0.13, 0.08)

0.5

-18

-16.5

0.9549

9.9144

3.5287

(-0.12, 0.13)

0.5

18

-16.5

S0

80 70

S1 60

Intensity/%

Copyright © 2012. Nova Science Publishers, Incorporated. All rights reserved.

13 0.000648 Zhang et al. 2006.

50 MC 37.5%

40 30 20

MC 5.7%

10 0 0

0.05

0.1

0.15 Time/ms

0.2

0.25

0.3

Figure 11. NMR FID curves for ash.

The FID shape can easily distinguish between solid wood and water components of the sample since wood protons signal decays quickly while water protons signal slowly. The first part of FID (signal intensity) is directly

Wood and Wood Products, Nova Science Publishers, Incorporated, 2012. ProQuest Ebook Central,

Minghui Zhang and Rado Gazo

related to the number of protons in the sample (more NMR signal means more protons in the sample). S0 is the total signal from both solid wood and water proton components. The solid matter decays to zero within 35μs (Nanassy, 1973) and the water component starts to decay from 60μs (Xu et al. 1996). The signal intensity (S1) at 60μs includes all water states. Figure 12 shows the relationship between the FID signal intensity and the moisture content of red oak, walnut and ash species. From the graph, we can get their linear relationship. The R-squares are all above 99%. The Table 6 is tabulated with comparison between oven-dried mass of the three different wood species by gravimetric method and calculated from linear regression equations in Figure 12. The results show that the calculated mass values are smaller than ones obtained with gravimetric method. The reason for that is that oven-dried treatment cannot remove all bound water in wood. Generally the T2 relaxation signal is measured by Carr-Purcell-MeiboomGill (CPMG) sequence. The relaxation curves obtained from the CPMG sequence can best be represented as a continuous distribution of relaxation times (Labbe et al. 2006). Different relaxation times mean different water states in wood. Figure 13 shows continuous distribution of the spin-spin relaxation times through CPMG experiment for yellow poplar sample with moisture content of 63%. Six states of water in different structural environments can be observed. Researchers believe slow relaxation component (T2 of 100ms or more) is free water and fast relaxation component (T2 of about 10ms) is bound water. 90

Red oak

y = 84.309x + 34.637 2 R = 0.9954

Walnut

y = 64.913x + 29.728 2 R = 0.9954

Ash

y = 88.184x + 39.842 2 R = 0.9919

80

NMR FID S0

Copyright © 2012. Nova Science Publishers, Incorporated. All rights reserved.

50

70 60 50 40 30 60%

50%

40%

30% MC

20%

10%

0%

Figure 12. Linear relationship between the FID signal So of the NMR and the moisture content of Red oak, Walnut and Ash during the desorption.

Wood and Wood Products, Nova Science Publishers, Incorporated, 2012. ProQuest Ebook Central,

Moisture Regain in Kiln-Dried Wood

51

T2 Distribution for Yellow Poplar

Signal Intensity (AU)

35000 30000 25000 20000 15000 10000 5000 0 0.01

0.1

1 10 100 Relaxation Time (ms)

1000

10000

Copyright © 2012. Nova Science Publishers, Incorporated. All rights reserved.

Figure 13. T2 distribution for yellow poplar with MC 63%.

Table 6. Comparison between oven-dried mass of the three different wood species by gravimetric method and calculated from linear regression equation species walnut Red oak ash

oven-dried mass 1.333 1.648 1.518

Calculated mass 1.31 1.632 1.51

Table 7. T2 distribution for yellow poplar with MC 63% (ms) Peak No. 1 2 3 4 5 6

Begin time 0.06 0.560806 3.15406 16.0253 54.2321 506.895

Peak time 0.122179 0.760634 5.80223 26.6324 81.4218 506.895

End time 0.183435 1.2641 14.4772 39.9847 99.766 561.099

Wood and Wood Products, Nova Science Publishers, Incorporated, 2012. ProQuest Ebook Central,

Integral value 2173.7325 4990.0576 131490.304 27283.7 72251.6 34391.37

52

Minghui Zhang and Rado Gazo

From results in Table 7, water in wood can be not only qualitatively separated into different states but also quantitatively calculated according to T2 distribution curve. This technique makes the study of moisture in wood more accurate and scientific. TD-NMR technique also has other applications in wood science. It is known that T2 and/or T1 are related to wood density, water exchange in wood, and porosity. Time-Domain Nuclear Magnetic Resonance will likely play an important role in wood science and technology research in the future.

ACKNOWLEDGMENT

Copyright © 2012. Nova Science Publishers, Incorporated. All rights reserved.

Authors would like to acknowledge the Forest Products Society and the Society of Wood Science and Technology for their kind permission to reuse text, figures and tables from previously published material (Zhang et al. 2006 and 2007) and National Natural Science Foundation of China for supporting part of this research.

REFERENCES Alfery, T., E. F. Gurnee and W. G. Lloyd. 1966. Diffusion in glassy polymers. J. Polym. Sci. C12: 249-261. American Society for Testing and Materials (ASTM). 2002. Standard test methods for direct moisture content measurement of wood and wood-base materials. ASTM D 4442-92. In: ASTM Annual Book of Standards. ASTM. West Conshohocken, PA. Andrews, D. H. and J. Johnston. 1924. The rate of absorption of water by rubber. Journal of The American Chemical Society. 46(3): 640-650. Araujo, C. D., A. L. Mackay, J. R. T. Hailey, K. P. Whittall and H. Le. 1992. Proton Magnetic-Resonance Techniques for Characterization of Water in Wood - Application to White Spruce. Wood Science and Technology 26(2): 101-113. Baronas, R. and F. Ivanauskas 2002. Numerical investigation of moisture movement in porous solid using a diffusion model. 15th Nordic Seminar on Computational Mechanics, Aalborg, Denmark.

Wood and Wood Products, Nova Science Publishers, Incorporated, 2012. ProQuest Ebook Central,

Copyright © 2012. Nova Science Publishers, Incorporated. All rights reserved.

Moisture Regain in Kiln-Dried Wood

53

Baronas, R.,F. Ivanauskas, I. Juodeikiene and A. Kajalavicius. 2001. Modelling of moisture movement in wood during outdoor storage. Nonlinear Analysis: Modelling and Control. 6(2): 3-14. Bois, P. J. 1978. Handing, Drying, and Storing Heavy Oak Lumber. Forest Prod. Utilization. Technical Report. No. 8. USDA Forest Serv., Forest Prod. Lab., Madison, WI. Carll, C. and A. T. Wolde. 1996. Accuracy of Wood Resistance Sensors for Measurement of Humidity. Journal of Testing and Evaluation. 24(3): 154160. Christensen, G. N. 1960. Kinetics sorption of water vapor by wood. Australian Journal of Applied Science. 11: 294-304. Christensen, G. N. and K. E. Kelsey 1959. The rate of sorption of water vapor by wood. Holz als roh-und werkstoff 17: 178-188. Comstock, G. L. 1962. Moisture diffusion coefficients in wood as calculated from adsorption, desorption, and steady state data. Master thesis. North Carolina State College, Raleigh, N.C. Comstock, G. L. 1963. Moisture diffusion coefficients in wood as calculated from adsorption, desorption, and steady state data. Forest Prod. J. 13(3): 97-103. Crank, J. 1975. The mathematics of diffusion. 2nd ed. Oxford: Clarendon Press. Dai, G. and K. Ahmet. 2001. Long-term monitoring of timber moisture content below the fiber saturation point using wood resistance sensors. Forest Prod. J. 51(5): 52-57. Dobrynin and Kulakova. 1980. Storage of dry sawn timber under natural conditions. Derevoobrabatyvayushchaya-Promyshlennost’. No. 2, 3-4. Eckelman, C. A. 1998. The Shrinking and Swelling of Wood and Its Effect on Furniture. FNR 163. 26pp. Gu, H. 1997. Moisture gradient during kiln drying of red oak. Master thesis. Dept. of Wood Science and Forest Products, VPI and SU, Blacksburg, VA. 5 pp. Hall, C. W. 1957. Drying farm crops. Agricultural Consulting Associates, Inc. 18-19pp. Hsi, E., Hossfield, R. and Bryant, R. G. 1977. Nuclear magnetic resonance relaxation study of water absorbed on milled Northern white cedar. J. Colloid Interface Sci. 62: 389-395. James, W. L. 1988. Electric Moisture Meters for Wood. FPL-GTR-6, USDA Forest Serv., Forest Prod. Lab., Madison, WI.

Wood and Wood Products, Nova Science Publishers, Incorporated, 2012. ProQuest Ebook Central,

Copyright © 2012. Nova Science Publishers, Incorporated. All rights reserved.

54

Minghui Zhang and Rado Gazo

James, W. L., E.T. Choong, D.G. Arganbright, and D.K. Doucet. 1984. Moisture levels and gradients in commercial softwood dimension lumber shortly after kiln-drying. Forest Prod. J. 34(11/12): 59-64. James, W. L., Y. You-Hsin and R. J. King. 1988. A Microwave method for measuring moisture content, density, and grain angle of wood. Res. Note FPL-0250. USDA Forest Serv., Forest Prod. Lab., Madison, WI. Jenkins, J. H. 1934. Report on the absorption of moisture by kiln-dried lumber. Forest Service Circular 23. Department of The Interior, Canada. Kulakova. 1983. Changes in the moisture content of packaged seasoned lumber in storage. Derevoobrabatyvayushchaya-Promyshlennost. No. 12, 5-6. Labbe, N., B. De Jeso, J.C. Lartigue, G. Daude, M. Petraud, M.Ratier. 2006. Time-domain H-1 NMR characterization of the liquid phase in greenwood. Holzforschung 60(3): 265-270. Liu, J. Y., W. T. Simpson and S. P. Verrill. 2001. An inverse moisture diffusion algorithm for the determination of diffusion coefficient. Drying Technology 19(8): 1555-1568. Mathewson, J. S. 1930. The Air Seasoning of Wood. Technical Bulletin No. 174. United States Department of Agriculture. Washington, D.C. 47-49pp. Menon, R. S., A. L. Mackay, J.R.T. Hailey, M. Bloom, A.E. Burgess, J.S. Swanson. 1987. A Nmr Determination of the Physiological Water Distribution in Wood during Drying. Journal of Applied Polymer Science 33(4): 1141-1155. Merela, M., P. Oven, I. Sersa, U. Mikac. 2009. A single point NMR method for an instantaneous determination of the moisture content of wood. Holzforschung: International Journal of the Biology, Chemistry, Physics, and Technology of Wood 63(3): 4p. Montgomery, D.C. 2001. Design and analysis of experiments. New York: John Wiley. Nakano, T. 1994. Non-steady state water adsorption of wood Part I. a formulation for water adsorption. Wood Sci. and Tech. 28(5): 359-363. Nakano, T. 1995. Note on a formulation of a water adsorption process of wood. Wood Sci. and Tech. 29(3): 231-233. Nanassy. A. J. 1973. Use of wide line NMR for measurement of moisture content in wood. Wood Science 5(3):187-193 Nanassy, A. J. 1976. True dry-mass and moisture content of wood by NMR. Wood Science 9(2): 104-109. Newman, A. B. 1931. The drying of porous solids: diffusion calculations. American Institute of Chemical Engineers 27: 311-333.

Wood and Wood Products, Nova Science Publishers, Incorporated, 2012. ProQuest Ebook Central,

Copyright © 2012. Nova Science Publishers, Incorporated. All rights reserved.

Moisture Regain in Kiln-Dried Wood

55

Peralta, P. N. 1995. Sorption of moisture by wood within a limited range of relative humidities. Wood and Fiber Science. 27(1):13-21. Quarles, S. L. 2004. Physical Limitations of Moisture Meters. http://www.colostate.edu/programs/cowood/New_site/Publications/Article s/moisturemeter.pdf Rietz, R. C. 1978. Storage of Lumber. Agri Handb. 531:4. USDA Forest Serv., Forest Prod. Lab., Madison, WI. 4 pp. Riggin, M. T., A. R. Sharp, R. Kaiser. 1979. Transverse NMR relaxation of water in wood. Journal of Applied Polymer Science 23(11): 3147-3154. Sharp, A. R., M. T. Riggin, R. Kaiser, M. Schneider. 1978. Determination of moisture content of wood by pulsed nuclear magnetic resonance. Wood and Fiber 10(2): 74-81. Sheppard, S. E. and P. T. Newsome. 1930. The sorption of water vapor by cellulose and derivatives. Part II. The kinetics of sorption. J. of Physical Chemistry. 34(6): 1158-1165. Shubnyi. 1975. Distribution of moisture in lumber during storage in wrapped packages. Derevoobrabatyvayushchaya-Promyshlennost’. No.4, 9-11. Siau, J. F. 1995. Wood: influence of moisture on physical properties. Dept. of Wood Science and Forest Products, VPI and SU, Blacksburg, VA. 219pp. Simpson, W. T. 1998. Equilibrium moisture content of wood in outdoor locations in the United States and worldwide. Research Note. FPL-NR268. USDA Forest Serv., Forest Prod. Lab., Madison, WI. 11 pp. Simpson, W. T., ed. 1991. Dry Kiln Operator‟s Manual. USDA Forest Serv., Forest Prod. Lab., Madison, WI. 229 pp. Simpson, W. T., J. Tschernitz, and J. Fuller. 1999. Air drying of lumber. Gen. Tech. Rept. FPL-GTR-117. USDA Forest Serv., Forest Prod. Lab., Madison, WI. 51-53 pp. Skarr, C. 1972. Water in Wood. Syracuse, Syracuse University Press. 218pp. Skaar, C. 1979. Moisture sorption hysteresis in wood. In: Symposium on wood moisture content-temperature and humidity relationships. VPI and SU, Blacksburg, Virginia. 23-35 pp. Skarr, C. 1988. Wood-Water Relations, Springer-Verlag Berlin Heidelberg. 283pp. Stamm, A. J. 1956. Diffusion of water into uncoated cellophane. I. From rates of water vapor adsorption, and liquid water absorption. J. of Physical Chemistry 60(1): 76-82. Stamm, A. J. 1959. Bound-water diffusion into wood in the fiber direction. Forest Prod. J. 9(1): 27-32.

Wood and Wood Products, Nova Science Publishers, Incorporated, 2012. ProQuest Ebook Central,

Copyright © 2012. Nova Science Publishers, Incorporated. All rights reserved.

56

Minghui Zhang and Rado Gazo

Stamm, A. J. 1960a. Bound-water diffusion into wood in across-the-fiber directions. Forest Prod. J. 10(10): 524-528. Stamm, A. J. 1960b. Combined bound-water and water-vapor diffusion into Sitka Spruce. Forest Prod. J. 10(12): 644-648. USDA Forest Service, Forest Products Laboratory. 1999. Wood Handbook: Wood as an Engineering Material. Gen. Tech. Rept. FPL-GTR-113. USDA Forest Serv., Forest Prod. Lab., Madison, WI. 3-7pp. Wadsö, L. 1994. Unsteady-state water vapor adsorption in wood: an experimental study. Wood and Fiber Sci. 26(1): 36-50. Wengert, E. M. and P. H. Mitchell. 1979. Psychrometric relationships and equilibrium moisture content of wood at temperatures below 212 0F. Symposium on wood moisture content - temperature and humidity relationships. Blacksburg, Virginia. Wilson, P. J. 1999. Accuracy of a Capacitance-type and Three Resistance-type Pin Meters for Measuring Wood Moisture Content. Forest Prod. J. 49(9): 29-32. Xu Y., Araujo C.D., MacKay A.L., Whittall K.P. 1996. Proton spin-lattice relaxation in wood – T1 related to local specific gravity using a fastexchange model. Journal of Magnetic Resonance, Series B 110, 55–64. Zhang, M., Gazo, R., Cassens, D. and J. Xie. 2006. Moisture Distribution in a Dried Red Oak Lumber Package Stored in a High Humidity Environment. Forest Products Journal. 56(4):75-80. Zhang, M., Gazo, R., Cassens, D. and J. Xie. 2007. Isothermal Water Vapor Adsorption Process in Kiln Dried Red Oak. Wood and Fiber Science. 39(3):397-403.

Wood and Wood Products, Nova Science Publishers, Incorporated, 2012. ProQuest Ebook Central,

In: Wood and Wood Products Editor: J. Paulo Davim

ISBN: 978-1-62081-973-9 © 2012 Nova Science Publishers, Inc.

Chapter 3

CUTTING ENERGY ON WOOD AND WOOD PRODUCTS MACHINING Alfredo Aguilera1, and Pierre-Jean Méausoone2 1

Copyright © 2012. Nova Science Publishers, Incorporated. All rights reserved.

Forest Science and Natural Resources Faculty, Universidad Austral de Chile, Valdivia, Chile 2 École Nationale Supérieure des Technologies et Industries du Bois, Lorraine University, Épinal, France

ABSTRACT The importance of enclosing the cutting conditions within defined limits is associated with the type of material to be machined – solid wood or wood based material – , since it determines the behavior of the cutting process. The cutting conditions must be set also considering machine and tool limits, with productivity and quality being the key points of the process. A complete evaluation of the raw material (wood), their proper selection and maximum size tolerance should be taken into account as a priority since they involve higher costs throughout the manufacturing process. All machining variables and parameters, including the particular characteristics of the wood material will influence the performance of the process in terms of manufacturing quality and cutting precision, tool wear and finally on the cutting forces and cutting power, becoming useful for 

Corresponding author, E-mail: [email protected].

Wood and Wood Products, Nova Science Publishers, Incorporated, 2012. ProQuest Ebook Central,

58

Alfredo Aguilera and Pierre-Jean Méausoone the online control of the performance limits of the cutting tool under a given operating condition. This chapter aims to provide a general understanding of the cutting kinematics elements such as chip thickness, cutting geometry and chip formation. Cutting energy – power is introduced for cutting planes, general concepts on tool wear and main factors affecting this parameter. Finally, some examples and findings are shown to illustrate cutting power trends on solid wood and wood panels.

Copyright © 2012. Nova Science Publishers, Incorporated. All rights reserved.

1. INTRODUCTION The evolution of means of production and the international competitiveness in the wood industry currently demand a higher profitability of the processes. The development of new technologies and cutting materials become the key points in the productivity of the manufacturing process. In an industrial context, the use of wood is very diverse given the variety of companies, with highly specialized machinery and processing methods, depending on the product to be manufactured. Thus, each machine is associated with a given product, making it necessary to control the machining conditions due to the very high cost of machine maintenance. Under optimized conditions, the loss of raw material and frequency of tool changes are minimized and the surface roughness is maximized, as well as the production costs associated with each piece. Therefore, the profitability of the machine is improved. The importance of enclosing the cutting conditions within certain limits is related to the type of material to be machined, since it determines the behavior of the cut. Therefore, the interaction between the tool and the material become a subject of high interest in order to establish a database that would allow to identify the cutting conditions that are accepTable for a particular material and for each tool, making it possible to define the working conditions of the tool to maximize its tool life while keeping a proper surface finish, in compliance with safety rules. A main problem of the wood industry is the lack of knowledge about the range of cutting conditions for an optimized machining, a problem particularly caused by the heterogeneity and the anisotropy of the raw material, where it is only possible to select average values of cutting conditions. It has been observed that the processing methods are often defined by the product according to its material and shape. Therefore, a given process is associated with a particular machine. The processes are quasi-linear and

Wood and Wood Products, Nova Science Publishers, Incorporated, 2012. ProQuest Ebook Central,

Cutting Energy on Wood and Wood Products Machining

59

Copyright © 2012. Nova Science Publishers, Incorporated. All rights reserved.

require many machines. Each one performs a specific operation, depending on the chosen processing entity. In addition, each entity is associated with a cutting tool that performs a specific operation. The settings and cutting parameters are a function of the material to be machined. The problem is the determination of cutting conditions that are appropriate for each entity (Aguilera et al. 2003) (Figure 1). The cutting conditions are set depending on the limits of the machine and tool, with productivity and quality objectives. With regard to metal working, there is plenty of technical information available. For example, if certain material needs to be machined, the machine settings are changed to reach a specific result in terms of surface quality or tool life. Several actors are involved in the industrial wood machining process, which must be in perfect harmony since the neglect of one of them leads to operational problems that affect the process performance, costs, product quality or compliance with customer terms. These actors are:     

The machine, The tool, The operator training, The production orders, The raw material. Machining subject Object form

Link with other surfaces

Choice of the spindle, the tool and its trajectory

Material and cutting depth

Precision

Cutting conditions determination

Choice of positioning axes

Choice of machine kinematics

Choice of actuators and movement processors

Specification of the control unit

Detailed design

Figure 1. Problem determination in wood cutting conditions determination. Wood and Wood Products, Nova Science Publishers, Incorporated, 2012. ProQuest Ebook Central,

Alfredo Aguilera and Pierre-Jean Méausoone

60

Machine set-up

Wood raw material selection

Wood based product

Production orders

Tool choice Operator skill

Copyright © 2012. Nova Science Publishers, Incorporated. All rights reserved.

Figure 2. Cause – effect diagram from production orders to final wood based product.

The lack of control over one or more of these components is critical, making it necessary to know the exact limits, characteristics and constraints of each actor in order to predict how far it is possible to get without affecting the expected and required results of the final product at some stage of production (Figure 2). Once the production order is set considering the needs and constraints of the client, it is necessary to establish the required volume of raw materials (wood, adhesive, coatings, other materials as needed, etc.). In the case of wood material, the following aspects should be taken into account:          

Wood density, Moisture content, Wood hardness, Abrasiveness, Growth ring width, Reaction wood, Juvenile or mature wood, Fiber direction, Resin pockets, Knots, type and size.

A comprehensive review of the raw materials, proper selection and size tolerances, should be taken into account as a priority, since this component has the higher costs throughout the manufacturing process. The interaction between machine set-up and tool choice allows adjusting the cutting conditions for done wood or wood based material that will be subjected to a specific cutting process, in other words, a wood base component will be

Wood and Wood Products, Nova Science Publishers, Incorporated, 2012. ProQuest Ebook Central,

Cutting Energy on Wood and Wood Products Machining

61

subjected to a process that requires a particular operational condition, which in turn requires a specific cutting tool (dimensional data as diameter, length… and geometrical data as cutting angles, shape…), a cutting material (HSS, HW, PCD, etc.) adapted to the type of wood material (solid wood, panels, composites) being machined and operating conditions (cutting speed, feed speed, cutting direction, depth and width of cut). All these variables, including the particular characteristics of the wood material will influence the process performance in terms of quality and cutting precision, tool wear and finally the cutting forces and cutting power, useful parameters for the online control of the performance limits of the cutting tool under a given operating condition. The following chapter will address general issues around the main cutting kinematics aspects, which will facilitate the understanding of the subject of energy and cutting power, including research results in both solid wood and panels.

Copyright © 2012. Nova Science Publishers, Incorporated. All rights reserved.

2. CUTTING KINEMATICS ELEMENTS When machining kiln dried wood (8 to 18% moisture content MC), the required cutting power and cutting forces depend on several factors such as: the chip formation mechanism through the cutting kinematics, the cutting geometry, cutting material and raw material in terms of moisture content and density. These are determining factors that must be considered along with surface roughness and tool wear. Part of the research in the wood machining process is aimed at studying the cutting processes that allow the optimal cutting conditions selection, decrease the loss of raw materials, improve the tool performance and increase the process safety with a good resulting surface quality. The cutting process is complex due to its unique characteristics, depending on the type of cutting tool, i.e. linear parameters, geometry, speed, and the materials the tool is made of, also it must be considered the wood with all its peculiarities, density, moisture content, anisotropy, and defects. The result of the tool – wood interaction, allows to identify the technological processes involved in the mechanical wood processing to reach a finished product, showing dimensional changes and modifying the shape and surface quality without altering its qualitative properties.Under these terms, the cutting theory will aim to meet the physical process and the factors involved on the cutting process, to allow the understanding of the factors that affect the surface quality and how to control and improve the machining conditions of the tool, with an increased productivity.

Wood and Wood Products, Nova Science Publishers, Incorporated, 2012. ProQuest Ebook Central,

Copyright © 2012. Nova Science Publishers, Incorporated. All rights reserved.

62

Alfredo Aguilera and Pierre-Jean Méausoone

The penetration of the tool in the workpiece is not possible when it is driven to move with a given energy. Often, the cutting movement is performed by the tool as is the case of sawing, planing or drilling, but sometimes it is the piece that moves like in turning and peeling. The characteristic of wood cutting is the machine work on the fly, because in most cases we have a feed for wood pieces without locking. When a tooth cuts wood, the necessary work is done, which means a power consumption because of chip formation. The cutting edge direction about the workpiece feed is generally perpendicular because it is the easiest solution; however, the tool is subjected to violent forces which can produce vibrations and breaks, a problem solved by tilting the edge allowing better distribution of the cutting forces and then allowing a constant and progressive machining. When the engaged teeth moves counter the workpiece movement, this action is called conventional cutting mode, but if both the engaged teeth and the workpiece move in the same direction, the process is called climb cutting mode. The conventional cutting mode demands a lower cutting force compared to climb cutting mode, when the forces increase from zero to their maximum value, however, in woods that have an unfavorable fiber direction (or around the knots), it results in poor surface quality. As indicated by Aguilera (2011), the chips in conventional cutting are built as the tooth advances, but it does not dull the tooth as quickly as the climb cutting mode, requiring less cutting power but a rougher surface, compared to the climb cutting mode.

2.1. Cutting Geometry Wood machining tools must ensure a good working quality from the standpoint of appearance and accuracy. They should have an adequate tool life, easy sharpening and easy to assembly and disassembly, they should maintain their cutting edge during as many sharpenings as possible; they should not be dangerous and should not cause abnormal heating of the wood. At the same time, they must have a good size gullet in order to collect the chips to allow maximum feed rates and ensure adequate energy consumption. The cutting tools are defined mainly by their cutting circle diameter, number of teeth, teeth shape, and cutting angles. It must be remembered that the cutting geometry considers three fundamental angles, each of great importance and influence: the feed rate, energy consumption, and the surface quality to the wood cutting process.

Wood and Wood Products, Nova Science Publishers, Incorporated, 2012. ProQuest Ebook Central,

Cutting Energy on Wood and Wood Products Machining

63

The cutting tool is related to the cutting surface by the following angles (Figure 3): 



Clearance angle α: angle formed between the back of the knife and the tangent to the cutting circle (on planing or moulding process) or to the new cutting plane (on band sawing process). This angle is related to rake and sharpness angle, depending also on the type of wood and feed rate. Denser wood requires a slower feed, then a small clearance angle (increasing the cutting friction), but a great sharpness angle. On softwood, the increase of the clearance angle implies a faster feed process; however, caution should be taken against the risk of rupture – breakage, because of a weaker knife. Sharpness angle β: is the angle formed between the face and the back of the knife or tooth. This angle gives strength to the knife, where the magnitude of this angle depends on the type of wood and cutting material (Table 1), increasing the sharpness angle when the rake angle is reduced and vice versa.

Copyright © 2012. Nova Science Publishers, Incorporated. All rights reserved.

Table 1. Sharpness angles according to type of wood and cutting material for planing knives Type of wood Softwood Hardwood Softwood Hardwood Softwood Hardwood

Cutting material HSS HSS Chrome-vanadium Chrome-vanadium HM HM

Sharpness angle 30-45º 50-55º 35-45º 45-48º 50º 55º

Sharpness angle β

Clearance angle α Vc

Vf

Rake angle γ

 



Figure 3. Cutting geometry on planing or molding process.

Wood and Wood Products, Nova Science Publishers, Incorporated, 2012. ProQuest Ebook Central,

Alfredo Aguilera and Pierre-Jean Méausoone

64

The sharpness angle is a function of cutting material, and as a general rule, at greater hardness there is an increase of the fragility of the cutting material. If this angle changes, the surface quality will be affected as well the tool life. When jointing, this angle is affected, causing problems such as fibers compression, more friction, temperature rise and edge wear.

Copyright © 2012. Nova Science Publishers, Incorporated. All rights reserved.



Rake angle γ is very important. It is mainly responsible for consumed cutting power, surface quality and tool life. Basically it is the angle of the tooth (or knife) in relation to the cut, being the angle between the face of the tooth and the normal to the cutting surface (in the case of band saw), or the angle between the face of the knife (tooth) and the projection line from the cutting edge through the center of the cutterhead or circular saw. This angle allows the release of the chip and determines the penetration of the edge, mainly affecting the produced finish, the consumed cutting power and it determines the feed resistance. A small angle causes a great effort and the tooth scrapes the wood rather than cut it, a big angle reduces the effort, but the tearing of fibers is accentuated, allowing faster feeds. On planing or molding, the choice depends on the type of wood, in particular its hardness, moisture content and expected surface finish.

If the angle is badly selected, it can cause defects such as torn fibers, due to insufficient cutting force, because of a very large angle between the main component of the cutting force which is projected beyond the edge and tends to pull the fibers. Fuzzy fibers are generated by compression of the fibers, which tend to lift because of an increase in wood moisture content and a very small rake angle. On planing and molding, as a general rule, it is adviced not to intervene in the head holder to modify the rake angle as it can cause the following problems:   

Unbalancing in the body of the cutterhead, Loss of holding power, Decrease in mechanical strength of the cutterhead.

The decrease of the rake angle can be done by projecting the knife, only up to a maximum of ½ of its total width outside of the cutterheads.

Wood and Wood Products, Nova Science Publishers, Incorporated, 2012. ProQuest Ebook Central,

Cutting Energy on Wood and Wood Products Machining

65

2.2. Cutting Process There are a number of factors involved in the cutting process which affect energy consumption and surface roughness, these factors are:

Copyright © 2012. Nova Science Publishers, Incorporated. All rights reserved.

        

Cutting speed (Vc) Feed speed (Vf) Cutterhead diameter (D) Number of knives (Z) Depth of cut (ap) Width of cut (b) Cutting geometry (mainly rake angle γ) Cutting direction (A 90-90, B 90-0 or C 0-90) Cutting mode (conventional or climb cutting)

As indicated by Aguilera (2011b) “two main movements must be applied to perform the cutting process: one on the cutting tool, called cutting speed (Vc), and the other on the work piece, called the feed rate (Vf). The outcome of these two movements produces a wood surface layer that will deform, shedding and generating a new surface with a certain surface quality. This wood layer is called chip, which in rotational cutting tools as the cutterheads and circular saw blades, has a comma appearance”.

Chip Formation In the case of cutting parallel to the fibre (as in planing or moulding process) the excess wood is removed in the form of individual chips. These individual chips are formed by intermittent and successive penetration of the cutting edge into the wood, where one or more knives are mounted on the perimeter of a rotating cutterhead (Figure 4). Figure 4 shows a typical S4S (Surfaced Four Sides) planing process where it is possible to observe both lateral cutter heads (Z10) generating a bad chip quality, where large fragments evidence problems like excessive cutting depth, over dried lumber and/or a feed speed that is not right for the rotational speed and number of knives being used (advance per tooth problem, see equation 3). A relevant feature of the peripheral cut means that the knife is changing its relative position angle to the predominant fibre direction. Although it happens at the beginning of the cut, this is mostly carried out parallel to the fibres, and

Wood and Wood Products, Nova Science Publishers, Incorporated, 2012. ProQuest Ebook Central,

66

Alfredo Aguilera and Pierre-Jean Méausoone

Copyright © 2012. Nova Science Publishers, Incorporated. All rights reserved.

when the knife starts to emerge, its produces a cut at an angle to the grain of the wood (Figure 5) (Koch, 1972). As the knives have a rotational movement (circular saws or planing heads), the produced chip varies constantly. Therefore, in these cases it is more appropiate to speak of mean chip thickness (Kivimaa 1950).

Figure 4. Chip formation in planing process.

Vc

Vf Figure 5. Knife‟s angle to the grain of the wood.

Wood and Wood Products, Nova Science Publishers, Incorporated, 2012. ProQuest Ebook Central,

Cutting Energy on Wood and Wood Products Machining

67

emax fz

Vc  em

ap

fz Vf

Copyright © 2012. Nova Science Publishers, Incorporated. All rights reserved.

Figure 6. Chip formation process in conventional cutting mode. Comma shape of the chip.

The mean chip thickness (em) is related directly to the feed per tooth or bite (fz) and cutting height (depth of cut) (ap), but inversely to the cutter head diameter, as shown in the following expression (1) for planning heads. The shape of the chip is like a comma (Figure 6).

e m  fz

ap D

(1)

For circular saws, the mean chip thickness (em) shape‟s is a truncated comma, directly related to the feed per tooth or bite (fz), the cutting height (or depth of cut) (ap) and the saw blade projection or over ride (f), but inversely to the cutter head diameter (2R), as shown in the following expression.

em 

fz  2 f 2 f  ap        2  R R 

(2)

As shown in equation (1) and (2), the mean chip thickness depends directly on the bite, which is determined as follows: Wood and Wood Products, Nova Science Publishers, Incorporated, 2012. ProQuest Ebook Central,

Alfredo Aguilera and Pierre-Jean Méausoone

68

fz 

1000*Vf N *Z

(3)

Where the mean chip thickness (em) and the advance per tooth (fz) explain the changes in the surface quality and the cutting forces (Juan, 1992 and Kivimaa, 1950). The advance per tooth is directly related with the feed speed (Vf) and indirectly with the number of knives (Z) and rotational speed (N) (3). Being the rotational speed directly related with the cutting speed (Vc), as it is shown in (4).

Vc    D  N

(4)

Copyright © 2012. Nova Science Publishers, Incorporated. All rights reserved.

The cutting depth is related to cutting energy and surface roughness. Davis (1964) concluded that surface roughness increases gradually when decreasing the depth of cut. Regarding energy requirements, Koch (1964) says they increase as the depth of cut increases, in addition to its influence on the aforementioned mean chip thickness (em).

3. CUTTING ENERGY AND CUTTING POWER According to Koch (1972), the factors affecting the cutting power are multiple and they are interrelated with surface roughness and tool wear, but it is still possible to differentiate three groups: first, factors related to the material, second, those relating to the cutting tool and finally, those related to the cutting process. When wood or wood based products are machined, the inherent characteristics of this material will be critical on surface roughness, tool wear and cutting forces – cutting power. Anatomical wood constituents (tracheids, vessels, rays), as described by Aguilera (2011b), have mainly an effect on surface roughness, depending on whether it is a hardwood or softwood, where more cell cavities will require a machining process that minimizes surface exposure of these elements, i.e., cutting kinematic conditions, and type of cutting tool adapted to the particular type of material to be processed. The performance of the cutting tool in terms of tool wear will be affected by solid wood, its density and wood based panels; its abrasiveness (due mainly by adhesives and different surface coatings); the degree of moisture content

Wood and Wood Products, Nova Science Publishers, Incorporated, 2012. ProQuest Ebook Central,

Copyright © 2012. Nova Science Publishers, Incorporated. All rights reserved.

Cutting Energy on Wood and Wood Products Machining

69

and chocks during process (chip thickness magnitude, knots presence). Tool wear is an ineviTable mechanism that combines the interaction of the cutting tool, the cutting material and machining conditions, and it is derived from a combination of several factors acting on the cutting edge that contribute to deform the cutting geometry. The main factors affecting the cutting tool are mechanical, thermal, chemical and abrasive. It should also be considered the material of the cutting edge, which must withstand the various phenomena of wear, such as abrasion, diffusion, oxidation, fatigue and adhesion, but in practice, the cutting speed has the most significant influence on tool life followed by feed and depth of cut (Lajis et al. 2008). Orthogonal wood cutting mechanics was studied by McKenzie (1961), Kivimaa (1950) and Franz (1958); they considered the relation between the cutting forces at a corresponding rake angle (Figure 8) and the resultant chip thickness. McKenzie (1961) arrives to illustrate the three basic types of orthogonal cutting that are today well known. The author uses two numbers to represent first the angle between the cutting edge of the tool and the grain direction, and second the angle between the cutting direction and the grain direction. 90 – 90° or “A” direction or cross cutting, perpendicular to the pith, corresponding to trimming operation (index 1 or 6); 90 – 0° or “B” direction, parallel to the pith or longitudinal cut, corresponding to sawing – re-sawing, rip sawing, planing and molding operation (index 2 or 5); 0 – 90° or “C” direction, tangential to growth rings, corresponding to the slicing or peeling process to obtain veneers (index 3 or 4) (Figure 7).

6

1

2

5 3

4

Figure 7. Cutting orientation.

Wood and Wood Products, Nova Science Publishers, Incorporated, 2012. ProQuest Ebook Central,

Alfredo Aguilera and Pierre-Jean Méausoone

Cutting force (N)

70

90-90° = A

90-0° = B 0-90° = C

Rake angle (γ)

Copyright © 2012. Nova Science Publishers, Incorporated. All rights reserved.

Figure 8. Cutting forces behavior according to rake angle and cutting direction.

The experimental studies developed by the authors allowed to establish relations between the cutting plans and the behavior of cutting forces, the effect of the degree of tool wear, and the rake angle. The rake angle has a strong effect on cutting forces (Figure 8) where their decrease determines a necessity of more cutting force (and hence more cutting power), but if cutting conditions are not well adjusted, or the wood or wood based material is not considered in terms of its moisture content, density or abrasiveness, the cutting process will result in problems of surface quality and accelerated tool wear, demanding more power from the process. Each type of wood material and process (cutting direction) will require a specific rake angle that is at same time associate to a specific cutting material (type of steel or alloy). In this sense, for the case of 90-0° or “B” direction (planing, molding or sawing process), the reduction of the rake angle must be accompanied with a harder cutting material and vice versa, and of course adapting the cutting kinematics to this specific processing conditions.

3.1. Cutting Forces and Tool Wear Wyeth et al. (2009) indicate that the study of the chip formation at different grain angles is fundamental to understand what occurs during the cutting process, adding that the role of fracture toughness, strength and friction

Wood and Wood Products, Nova Science Publishers, Incorporated, 2012. ProQuest Ebook Central,

Copyright © 2012. Nova Science Publishers, Incorporated. All rights reserved.

Cutting Energy on Wood and Wood Products Machining

71

between the chip and the tool are important on the cutting forces. In this sense, Orlowski and Wierzbowski (2010) specify that accelerated drying causes a decrease of wood mechanical properties such as fracture toughness and shear yield strength, additionally these phenomena also caused a reduction in the specific cutting resistance. Aknouche et al. (2009) studied the tool wear effect on cutting forces. The findings indicate that the shape of the wear evolution according to cutting length curve is similar to the typical wear curve on metal machining. They found a progressive increase of the cutting forces following the same tendencies of the tool wear. However, only one knife was evaluated and they did not perform any surface quality measure. Gauvent et al. (2006) deals with the wear phenomenon decomposition between corrosion and abrasion analysis in order to allow a study and a global understanding of this complex phenomenon. They validate corrosion and abrasion interaction in order to generate the cutting edge global wear. With the sights of their work, the chemical composition of the cutting material is as important as the material mechanic properties against the wear. McKenzie et al. (2001) concluded, on a study designed to predict cutting forces on routing layered boards, that the “rake angle and dullness must be included as variables, and wider sampling of the range of material is required. A suiTable lathe adapted for force component measurement has been found to have advantages over the milling machine in allowing faster testing and inclusion of dullness as a variable”. It is important to highlight that in this study the models of cutting forces reached high levels of correlation with tool wear for both parallel and normal force component. Others studies regarding the relationship between cutting conditions and cutting forces have been developed by Boucher et al. (2007), Eyma et al. (2004, 2005), Palmqvist et al. (2003, 2005), and Aguilera and Martin (2001), who found, in general, well correlated results between chip thickness and the resulting cutting forces. Also, Porankiewicz et al. (2005, 2006) analyzed the dull of high speed steel cutting tools after milling five wood species of different densities and mineral content, where a variable behavior in tool wear was found partly due to inherent wood characteristics.

3.2. The Cutting Power The kinematics of the cutting process will determine the power requirements in association with the characteristics of the raw material and the

Wood and Wood Products, Nova Science Publishers, Incorporated, 2012. ProQuest Ebook Central,

72

Alfredo Aguilera and Pierre-Jean Méausoone

cutting tool. It is worthwhile to remember that two movements are required to perform the wood cutting process, the primary movement on the cutting tool which at the same time is the one that needs more power, and the secondary or feed movement of the workpiece. It is necessary then, by opposition of both movements, to detach a wood layer with a given form and dimensions. This action is facilitated through the cutting edge of the tool, which is subjected to a force that opposes the wood to change its shape. There is a movement – the primary one– which can be expressed as a velocity (Vc) measured in m/s, and a cutting force (Fc) to produce the chip detachment measured in Newton (N). Therefore, the product of the force and speed generate a needed cutting power (Pc) or net power expressed in kilowatts (kW). Then:

Copyright © 2012. Nova Science Publishers, Incorporated. All rights reserved.

Pc  Fc *Vc

(5)

The cutting power in relation to the feed speed on wood machining depends on the section being processed (Koch, 1964) and the cutterhead involved in the process, in planing process this means that the power requirements will be higher for the upper head rather than the lateral heads who work the edges of the pieces. For the feed rate, power requirements will increase when increasing the feed rate (increase of the mean chip thickness and the advance per tooth), and for the upper head (which is doing more work), beyond a certain speed, the cutting power takes an exponential trend. The general behavior of the cutting power related to processing time or cutting distance is presented in Figures 9 and 10. Figure 9 shows the power behavior for the three typical zones during the motor operation on cutting process. The cutting power will result from the difference between the total power and the idle power, in this case measured in kilowatts or amperes, where at the beginning when the process starts and the tool engages the wood material, the power gradually increases to reach the stabilization zone through the entire process, with constant cutting conditions. Figure 10 shows that in long cutting distance or when machining abrasive material, or using inadequate tool material, the expected cutting power tends to increase progressively with time or cutting distance progress, i.e. for all cutting condition constants this increase can be explained by tool wear, where the lower cutting capacity of the tool will require more power and consequently more cutting force to be able to detach the chips. The consequence is a poor surface quality.

Wood and Wood Products, Nova Science Publishers, Incorporated, 2012. ProQuest Ebook Central,

Power (kW)

Cutting Energy on Wood and Wood Products Machining

73

Total power Stabilization zone during cutting process Idle power

a

b

c Time (s)

Copyright © 2012. Nova Science Publishers, Incorporated. All rights reserved.

Power (kW)

Figure 9. Total and idle power versus time at different process stages: a) starting process, b) stabilization zone when “normal” operating conditions, c) ending process.

Time or Cutting distance

Figure 10. Cutting power trend according to time or cutting distance processed.

In order to machine for long time or longer cutting distance without expending more power, obtaining a better quality, it is necessary to have a better performing tool and appropriate cutting conditions. The cutting forces or the specific energy are difficult to measure at the production line, but in practice, monitoring the electrical output of the engines is a reality, through amperes or kilowatts measures or recording devices. The main factors affecting the cutting power are the following:   

Number of knives Rake angle Wood density

Wood and Wood Products, Nova Science Publishers, Incorporated, 2012. ProQuest Ebook Central,

Alfredo Aguilera and Pierre-Jean Méausoone

74     

Moisture content Cutting edge Depth of cut Cutting width Chip thickness

As previously mentioned, the cutting power depends on several factors, which could be summarized in the following formula (6), cutting power formula used in 90-0 machining (planing) which can provide a good approximation:

Copyright © 2012. Nova Science Publishers, Incorporated. All rights reserved.

Pc  Z * N * Fc * ap * D

(6)

where: Pc = cutting power (W) Z = number of teeth N = rotation speed (Rd/s) Fc = cutting forces (N) ap = depth of cut (m) D = cutting circle (m) In this regard the work specific (or specific cutting energy) is the ratio between the cutting work and the chip volume generated in the cut. It is known that the work "T" is the product of a force (cutting force Fc) and distance (m)(7), and the cutting power is obtained between the ratio work "T" and a time "t" (s)(8), i.e. T = Fc * distance (Nm)

(7)

and the cutting power is:

Pc  T / t (Nm/s)

(8)

Thereby work specific Wsp = T / O, where O is the chip volume, or even according to the following equations:

Pc  Wsp * b * ap *Vf

Wood and Wood Products, Nova Science Publishers, Incorporated, 2012. ProQuest Ebook Central,

(9)

Cutting Energy on Wood and Wood Products Machining

75

or:

Wsp  Pc /(b * ap *Vf )

(10)

or:

Wsp 

Z * N * Fc * ap * D b * ap * Vf

(11)

Copyright © 2012. Nova Science Publishers, Incorporated. All rights reserved.

where: Wsp = work specific Pc = cutting power b = width of cut ap = depth of cut Vf = feed speed It can be said that the cutting energy is equal to the work of all the forces exerted on the tool. The power corresponds to the volume of material turned into chip per second, multiplied by the work specific. This is a function of the parameters associated with the material and working conditions. The calculated result should never be considered as an exact or even approximate value, but only as a range where they would be around values. The cutting force increases when increasing the chip thickness, but the absorbed energy decreases. This means that for the same working conditions, it is cheaper to process thick chips, although it will be needed to install a more powerful engine. This apparent contradiction is explained by the fact that the product between the cutting power and machining time decreases as the chip thickness increases. It be remembered that the cutting power calculated this way is required to detach the chips, but it is necessary to add all the powers that are necessary and useful to overcome the different rub. This rub can be explained as a loss of power due to a decrease in the electrical and mechanical efficiency transmission of the motors (13). When studying for different cutting modes (A, B, C and their combinations) the main components of the cutting force and the work specific (Wsp) as a function of chip thickness, Kivimaa (1950) found that the parallel force (Fy) increases gradually with the chip thickness, but the work specific presents an exponential growth when the chip thickness reaches very small values (Figure 11).

Wood and Wood Products, Nova Science Publishers, Incorporated, 2012. ProQuest Ebook Central,

Alfredo Aguilera and Pierre-Jean Méausoone

Work specific

76

Target zone Mean chip thickness

Copyright © 2012. Nova Science Publishers, Incorporated. All rights reserved.

Figure 11. Work specific according to mean chip thickness variation. Target zone where there are a well compromise between energy and quality.

The work specific is affected by the moisture content and wood density, but specially by fiber direction. The crosscut (A) demands 2.5 to 3 times more force than the longitudinal cut (B). Longitudinal cut is most common in the planing / molding process, but with an increase of the depth of cut the required force is close to the crosscut process. Table 2. Work specific for mean chip thickness, planing process of conifer in fiber direction Mean chip thickness (mm) 0.01 0.02 0.03 0.04 0.05 0.07 0.10 0.15 0.20 0.25 0.30 0.40 0.50

Work specific (hp/mm2) 15.0 12.0 9.0 7.0 6.0 4.7 3.6 2.6 2.0 1.8 1.5 1.2 1.0

Wood and Wood Products, Nova Science Publishers, Incorporated, 2012. ProQuest Ebook Central,

Cutting Energy on Wood and Wood Products Machining

77

Since Wsp depends on the mean chip thickness (em), we have:

Vf Z *N

em 

ap D

(12)

If the value of em is known Wsp can be obtained from Table 2, but these are approximate values according to the factors listed above. A net power is obtained when applying the formula, comparable to the power required by a motor (free power plus friction losses etc.). As an example: Vf = 50 m/min N = 5000 1/min Z=6

ap = 2 mm b = 10 cm D = 160 mm

From equation (12), the mean chip thickness is: 50* 1000 2  0.19mm 5000* 6 160

Copyright © 2012. Nova Science Publishers, Incorporated. All rights reserved.

em 

with em = 0.19 mm, the corresponding work specific from Table 2 is: Wsp ≈ 2.0 And from equation (9), the cutting power is:

Pc 

2.0 * 10 * 2 * 50  3.3kW 6 *102

The motor power (Pt) for those cutting conditions is:

Pt 

Pc e *m

(13)

where εe is the electrical efficiency (0.9) and εm is the mechanical efficiency of the motor (0.8), considering normal and periodical motor maintenance. Pt 

3.3  4.6kW *1.36  6.3hp 0.9 * 0.8

Wood and Wood Products, Nova Science Publishers, Incorporated, 2012. ProQuest Ebook Central,

Alfredo Aguilera and Pierre-Jean Méausoone

78

In these cutting conditions, the resulting surface quality in terms of advance per tooth is: fz 

50 *1000  1.67mm 5000 * 6

or 25.4/1.67 = 15 marks per inch, so this is a very good finish, for appearance grade products. And, if the objective is to improve the surface quality by reducing fz without affecting the production level, there are two options: change the cutter head incorporating more knives, or by making the cutter head to rotate at higher speeds. For the first option, i.e. Z = 6, to Z = 10, we calculate the advance per tooth (fz) to re-evaluate the surface quality:

fz 

50*1000  1.0mm 5000*10

Copyright © 2012. Nova Science Publishers, Incorporated. All rights reserved.

Then, 25.4/1.0 = 25 marks per inch, and the mean chip thickness results in:

em 

50 * 1000 2  0.11mm 5000*10 160

with em = 0.11 mm, the corresponding work specific from Table 2 is: Wsp ≈ 3.6, and the cutting power:

Pc 

3.6 *10 * 2 * 50  5.9kW 6*102

The total power at the motor is:

5.9  8.2kW *1.36  11.2hp 0.9 * 0.8 Now keeping the original Z and increasing the rotational speed to reach fz = 1.0 mm: Pt 

Wood and Wood Products, Nova Science Publishers, Incorporated, 2012. ProQuest Ebook Central,

Cutting Energy on Wood and Wood Products Machining

fz 

79

50*1000  1.0mm N *6

The new N will be: 8333 1/min, and then, we recalculate the mean chip thickness:

Copyright © 2012. Nova Science Publishers, Incorporated. All rights reserved.

em 

50* 1000 2  0.11mm 8333* 6 160

So Wsp ≈ 3.6 and the cutting power 5.9 kW, total motor power = 11.2 hp. In both cases, remember the restriction on the example: do not reduce production rate, then to improve surface quality, more required power is needed. If the option of having more knives is possible, another cost is an accelerated tool wear because of the reduction on the mean chip thickness. So more knives for better quality means more power but more tool wear. At production site, it is necessary to evaluate electricity costs versus tools costs. It must be indicated that the regulation of tools is more difficult. Indeed, obtaining a regular cutting circle is more difficult when the number of knives is more important. For the second option, fewer knives are rotating faster; they are subjected to more stress, and then a faster tool wear may appear. If this option is available, as in most modern machines, with more knives at the cutter head, there are more possibilities to increase significantly the production rate with better surface quality as well. If planing / molding requirements are not as strict, fz can be increased through higher energy savings. In general, the following considerations are required: An advance per tooth up to 3 – 4 mm provides in most cases a good surface finish. b. The tool life is improved if the rake angle is as large as possible; a thicker chip as possible, avoiding compression and friction with the wood. a.

Table 3 shows a summary of the relationship between different cutting conditions.

Wood and Wood Products, Nova Science Publishers, Incorporated, 2012. ProQuest Ebook Central,

Alfredo Aguilera and Pierre-Jean Méausoone

80

Table 3. Combinations between cutting parameters and expected responses Parameter to evaluate em, fz em Vf ae ap Z Z N N

Situation increase increase increase increase increase increase increase increase increase

Response parameter Fc, Pc Wsp em, fz, Fc, Pc Fc, Pc Fc, Pc, em em, fz, Fc Pc em, fz, Fc Pc

Situation

Note

increase decrease increase decrease increase decrease increase decrease increase

em, fz constants fz, Z, N constants ap, N constants ap, N constants ap, Z constants ap, Z constants

Copyright © 2012. Nova Science Publishers, Incorporated. All rights reserved.

Figures 12 and 13 show the cutting power behavior for Blackwood (Acacia melanoxylon) 90-0° planing, under following conditions: 14% moisture content, 0.63 g/cm3 mean wood density, Z4 conventional cutter head, HSS, 25° rake angle. The effect of the chip thickness increase on the cutting power for different levels of cutting depth, ranging from 1 to 3 mm is shown in Figure 12.

Figure 12. Cutting power of Blackwood planing as change in chip thickness and cutting depth.

In Figure 12, it is possible to observe that when the mean chip thickness increases its magnitude, then the total cutting power increases, and for a Wood and Wood Products, Nova Science Publishers, Incorporated, 2012. ProQuest Ebook Central,

Copyright © 2012. Nova Science Publishers, Incorporated. All rights reserved.

Cutting Energy on Wood and Wood Products Machining

81

constant level of chip thickness, the deeper cuts requires more power. On wood cutting process, a similar trend was found by Aguilera and Martin (2001) when evaluating the planing of beech and spruce lumber, studying the behavior of the cutting power and depth of cut with the change of the wood density, to finally analyze the surface quality. In their findings, the increase in wood density and depth of cut presented a continuous rise of the cutting power requirements, with no significant differences at different depths of cut, but a still clear deterioration of the surfaces with more feed rate. On solid wood, the wood density can also result in significant variations of cutting power, with Aguilera and Martin (2001) and Aguilera and Muñoz (2011) concluding that for a constant level of advance per tooth, the cutting power behaves directly to rotational speed, where an increase in cutting power requirements was detected for high density wood, and finally, the surface quality was strongly influenced by the mean chip thickness. The work specific trend (from equation 10) is shown in Figure 13 with the corresponding cutting power measures for different levels of the mean chip thickness. Here the increase of the chip thickness produces consequently a raise of the required cutting power, but demands less cutting energy as work specific, i.e. a thicker chip will be cheaper to process (processing at higher feed rate) and, however, a more powerful engine will be needed.

Figure 13. Work specific and cutting power according to mean chip thickness for Blackwood planing.

Wood and Wood Products, Nova Science Publishers, Incorporated, 2012. ProQuest Ebook Central,

Copyright © 2012. Nova Science Publishers, Incorporated. All rights reserved.

82

Alfredo Aguilera and Pierre-Jean Méausoone

The cutting energy on wood based panels was evaluated by Aguilera et al. (2000) and Aguilera (2011c), the first one measured the cutting forces, cutting power and surface roughness for MDF‟s routing process, where the main findings indicate a direct relationship between panel density increase and cutting power (cutting forces) for different levels of the chip thickness, and a deteriorated surface quality at the core of the panel (lower density area). The second author evaluated the effect of the chip thickness on the resulting surface roughness for the rip sawing process of MDF‟s panels, monitoring and correlating the cutting energy with acoustic emission probe. The main findings are presented in Figure 14: the increase of the cutting energy with the decrease of the chip thickness, then the chip thickness is a very important parameter that influences the quality of the whole machining process, affecting significantly the cutting energy and at the same time the surface quality. The monitoring of cutting energy can be also possible with the sound pressure technique, with satisfactory correlation with the surface roughness. Figure 15 shows the results of rip sawing process on MDF panels processing where changes on the cutting mode (conventional and climb cutting) and cutting speed are analyzed. The increasing of the cutting speed caused an improvement in the surface roughness and therefore an increase in the sound pressure. This improvement was related to the cutting conditions through the mean chip thickness.

Figure 14. Cutting energy and surface roughness for MDF rip sawing.

Wood and Wood Products, Nova Science Publishers, Incorporated, 2012. ProQuest Ebook Central,

Cutting Energy on Wood and Wood Products Machining

83

Copyright © 2012. Nova Science Publishers, Incorporated. All rights reserved.

Figure 15. Sound pressure monitoring cutting energy on MDF‟s rip sawing process according to cutting mode and cutting speed.

Moreover, a good relationship can be observed between mean chip thickness and surface roughness, and also sound pressure, so the mean chip thickness is a good factor to accurately assess both response variables, and can be very useful to control the feed rates in the machining process.

CONCLUSION The wide varieties of wood-based products available on the market require specific manufacturing methods and machinery linked to unique operating conditions. For each machining operation, the correct choice of cutting tool, cutting material and process kinematic significantly influence the quality of finished products, process productivity and costs. The appropriate selection of these parameters allow to achieve production goals in terms of quantity and quality, with a proper performance of cutting tools in terms of tool life and energy consumption. In this sense, cutting energy expressed through cutting forces, cutting power and specific cutting work, or through the use of acoustic emission sensors (voltage) or sound (sound pressure) or simply by controlling the power consumption of each motor (amperage), allow the monitoring of the cutting processes (either on solid wood or panels), a relatively well estimate of the

Wood and Wood Products, Nova Science Publishers, Incorporated, 2012. ProQuest Ebook Central,

84

Alfredo Aguilera and Pierre-Jean Méausoone

cutting tool change before a severe deterioration of the workpiece surface happens. However, it must be remembered that wood is a heterogeneous material whose response to cutting processes will never be evident, since it is subjected to a variability that may condition the precise sensors data acquisition, therefore, necessary to establish broad ranges for response parameters to an adequate safety margin estimating the limits of the cutting process, tools and consequently the resulting quality.

ACKNOWLEDGMENTS

Copyright © 2012. Nova Science Publishers, Incorporated. All rights reserved.

The authors would like to acknowledge the support of the Direction of Research and Development (DID) and the Forest Science and Natural Resources Faculty of the Universidad Austral de Chile at Valdivia, Chile, and the School of Wood Engineers (École Nationale Supérieure des Technologies et Industries du Bois) of the Lorraine University (formerly Université Henri Poincaré, Nancy 1) at Épinal, France.

ABOUT THE AUTHORS Alfredo AGUILERA: PhD in Wood Science, University Nancy 1, France. Current Associate Professor at Forest Science and Natural Resources Faculty, Universidad Austral de Chile, Valdivia, Chile. [email protected]. Pierre-Jean MEAUSOONE: PhD in Wood Science, University Nancy 1, France. Current Associate Professor at École Nationale Supérieure des Technologies et Industries du Bois, Lorraine University, Épinal, France. [email protected].

REFERENCES Aguilera, A. 2011. Surface roughness evaluation in medium density fibreboard rip sawing. Eur. J. Wood Prod., 69(3):489-493. DOI: 10.1007/s00107010-0481-3 Aguilera, A. 2011b. Monitoring surface quality on molding and sawing processes for solid wood and wood panels. In: Wood machining. Wiley ISTE Ltd. London, England, p. 159-216. ISBN: 978-1-84821-315-9.

Wood and Wood Products, Nova Science Publishers, Incorporated, 2012. ProQuest Ebook Central,

Copyright © 2012. Nova Science Publishers, Incorporated. All rights reserved.

Cutting Energy on Wood and Wood Products Machining

85

Aguilera, A. 2011c. Cutting energy and surface roughness in medium density fiberboard rip sawing. Eur. J. Wood Prod., 69(1):11-18. DOI: 10.1007/s00107-009-0396-z. Aguilera, A. Martin, P. 2001. Machining qualification of solid wood of Fagus silvatica L. and Picea excelsa L.: cutting forces, power requirements and surface roughness. Holz als Roh- und Werkstoff 59 (6), 483-488. DOI: 10.1007/s001070100243. Aguilera, A., Méausoone, P.J., Martin, P. 2003. A new methodology for wood cutting optimization in the secondary manufacturing processes. Holz als Roh- und Werkstoff, 61(5):358 - 362. DOI: 10.1007/s00107-003-0403-8. Aguilera, A., Méausoone, P.J., Martin, P. 2000. Wood material influence in routing operations: the MDF case. Holz als Roh und Werkstoff, 58(4):278283. DOI: 10.1007/s001070050425. Aguilera, A., Muñoz, H. 2011. Surface roughness and cutting power on Blackwood and Redwood planing. Maderas, Cienc. Tecnol., 13(1):19-28. DOI: 10.4067/S0718-221X2011000100002. Aknouche, H., Outahyon, A., Nouveau, C., Marchal, R., Zerizer, A., Butaud, J.C. 2009. Tool wear effect on cutting forces: In routing process of Aleppo pine wood. Journal of materials processing technology 209: 2918–2922. DOI:10.1016/j.jmatprotec.2008.06.062. Boucher, J., Méausoone, P.J., Martin, P., Auchet, S., Perrin, L. 2007. “Influence of helix angle and density variation on the cutting force in wood-based products machining”. Journal of Materials Processing Technology, vol. 189, p. 211–218. Davis, E.M. 1964. “Machining and related characteristics of United States hardwoods”. U.S.A. U.S. Department of Agriculture, Forest Service. 68 p. (Technical Bulletin – 1267). Eyma, F., Méausoone, P.J., Larricq, P., Marchal, R. 2005. Utilization of a dynamometric pendulum to estimate cutting forces involved during routing. Comparison with actual calculated values. Ann. For. Sci. 62: 441– 447. Eyma, F., Méausoone, P.J., Martin, P. 2004. Study of the properties of thirteen tropical wood species to improve the prediction of cutting forces in mode B. Ann. For. Sci. 61: 55–64. Franz, N. 1958. “An analysis of the wood-cutting process”. Ph.D. Thesis. Univ. Mich., Ann Arbor. 152 pp. Gauvent, M., Rocca, E., Méausoone P-J., Brenot, P. 2006. Corrosion of materials used as cutting tools of wood. WEAR, Volume 261, Issue 9, pp 1051-1055.

Wood and Wood Products, Nova Science Publishers, Incorporated, 2012. ProQuest Ebook Central,

Copyright © 2012. Nova Science Publishers, Incorporated. All rights reserved.

86

Alfredo Aguilera and Pierre-Jean Méausoone

Juan, J. 1992. “Comment bien usiner le bois”. CTBA, Paris, p. 140. Kivimaa, E. 1950. “Cutting force in wood-working”. The State Institute for Technical Research, Finland, Julkaisu 18 Publication, p. 102. Koch, P.1972. “Utilization of the Southern Pines. Vol. I: the Raw Material”. Agriculture Handbook No. 420, USDA-Forest Service, p. 734. Koch, P. 1964. “Wood machining processes”. A volume in the wood processing series. Edited by Frederick F. Wangaard, Yale University. Lajis, M.A., Mustafizul Karim, A.N., Nurul Amin, A.K.M., Hafiz, A.M.K., Turnad, L.G. 2008. Prediction of Tool Life in End Milling of Hardened Steel AISI D2. European Journal of Scientific Research, Vol.21 (4):592602. McKenzie, W.M. 1961. Fundamental analysis of the wood cutting process. PhD thesis, University of Michigan, Department of Wood Technology. McKenzie, W.M., Ko, P., Cvitkovic, R., Ringler, M. 2001. Towards a model predicting cutting forces and surface quality in routing layered boards. Wood Sci. Technol. 35 (6): 563-569. DOI: 10.1007/s002260100115. Orlowski, K.A., Wierzbowski, M.A. 2010. Fracture toughness and shear yield strength determination of steam kiln–dried wood. The Future of Quality Control for Wood and Wood Products, 4-7th May, Edinburgh, The Final Conference of COST Action E53. 8 pp. Palmqvist, J., Lenner, M., Gustafsson, S.I. 2005. Cutting-forces when upmilling in beech. Wood Sci. Technol. 39 (8): 674–684. DOI: 10.1007/ s00226-005-0010-4. Palmqvist, J., Lenner, M., Gustafsson, S.I. 2003. Cutter head forces and load cell scanning. Wood Sci. Technol. 37 (3-4): 199–211. DOI: 10.1007/ s00226-003-0174-8. Porankiewicz, B., Iskra, P., Sandak, J., Tanaka, C., Józwiak, K. 2006. Highspeed steel tool wear during wood cutting in the presence of hightemperature corrosion and mineral contamination. Wood Sci. Technol. 40 (8):673-682. DOI: 10.1007/s00226-006-0084-7. Porankiewicz, B., Sandak, J., Tanaka, C. 2005. Factors influencing steel tool wear when milling wood. Wood Sci. Technol. 39 (3): 225–234. DOI: 10.1007/s00226-004-0282-0. Wyeth, D.J., Goli, G., Atkins, A.G. 2009. Fracture toughness, chip types and the mechanics of cutting wood. A review. COST Action E35 2004–2008: Wood machining – micromechanics and fracture. Holzforschung, Volume 63 (2), D.O.I. 10.1515/HF.2009.017.

Wood and Wood Products, Nova Science Publishers, Incorporated, 2012. ProQuest Ebook Central,

In: Wood and Wood Products Editor: J. Paulo Davim

ISBN: 978-1-62081-973-9 © 2012 Nova Science Publishers, Inc.

Chapter 4

INVESTIGATION OF OPTIMUM PARAMETERS FOR MULTIPLE PERFORMANCE CHARACTERISTICS IN DRILLING WOOD COMPOSITES (MDF) USING GREY –TAGUCHI METHOD Copyright © 2012. Nova Science Publishers, Incorporated. All rights reserved.

K. Palanikumar1,, S. Prakash and J. Paulo Davim 1

Department of Mechanical Emgineering, Sri Sai Ram Institute of Technology, Chennai, India 2 Department of Mechanical and Production Emgineering, Sathyabama University, Chennai, India 3 Department of Mechanical Engineering, University of Aveiro, Campus Santiago, Aveiro, Portugal

ABSTRACT This paper presents a novel effective method for optimizing machining parameters on drilling wood composites that have multiple characteristics identified using GRA. Drilling parameters such as spindle speed, feed rate and drill diameter were optimized based on multiple performance characteristics. The characteristics of interest are thrust force and average surface roughness. Taguchi L 27 orthogonal array is used for 

Corresponding author, E-mail: [email protected].

Wood and Wood Products, Nova Science Publishers, Incorporated, 2012. ProQuest Ebook Central,

88

K. Palanikumar, S. Prakash and J. Paulo Davim experimentation. Experiments are conducted on MDF panel as per IS 12406 standard. Drilling tests are carried out using Physical Vapor Deposition (PVD) TiN coated carbide step drill bits. Grey Relational analysis is used for the optimization of machining parameters on drilling MDF panel. Analysis of the grey relational grade indicates that parameter significance and the optimal parameter combination for the drilling process are identified. Experimental results have shown that the machining performance in the wood composite drilling process can be improved effectively using this approach.

Copyright © 2012. Nova Science Publishers, Incorporated. All rights reserved.

1. INTRODUCTION Wood-based composites encompass a range of products, from fiberboard to laminated beams. Wood-based composites are used for a number of nonstructural and structural applications in product lines ranging from panels for interior covering purposes to panels for exterior uses and even in furniture, support structures in buildings. Medium density fiberboard (MDF) is one of the most rapidly growing composite panel products in forest products. MDF is used in many applications in wood industries because of its favorable properties such as surface characteristics, dimensional stability and excellent machinability. MDF is a wood substitute form which is made from fine wood fibers in a resin bonded with application of heat and pressure. It is manufactured by a dry processing technology at a lower temperature [1]. However basic concepts of wood machining have received little attention because lack of fundamental knowledge precludes the possibility of obtaining either empirical or theoretical data that can be utilized. Drilling is a popular and widely used machining process in industries. The main considerations during drilling are hole quality, and surface finish. Drilling operation is fundamental in the manufacturing industry to drill holes especially in furniture industries. Drilling operations are evaluated based on the performance characteristics such as surface roughness, material, and cutting force. These performance characteristics are strongly correlated with cutting parameters such as spindle speed, feed rate, and drill diameter. It is an important task to select cutting parameters for achieving high cutting performance [2-3]. Achieving desired surface quality is of great importance for the functional behavior of the assembly parts. Most of the researchers on the machining of composites have focused on turning and facing while drilling of MDF has not received much attention. In drilling of MDF composites, thrust force is the important response which affects the hole quality. The effect

Wood and Wood Products, Nova Science Publishers, Incorporated, 2012. ProQuest Ebook Central,

Copyright © 2012. Nova Science Publishers, Incorporated. All rights reserved.

Investigation of Optimum Parameters …

89

of thrust force on drilling of MDF composites is studied by many researchers. The influence of drill geometry on thrust force has been extensively studied due to the fact that the higher the machining forces, the larger the damaged area on the work material. Therefore, considerable efforts have been focused on developing drill geometries capable of reducing forces and consequently, minimizing the machined part damage. Most of the defects in drilling of composites are due to thrust force experienced by the work piece [4]. Engin et al. [5] presented some observations to enhance the understanding of the machining characteristics of MDF in which the focus had mainly been on the forces (machining and friction). They studied the orthogonal cutting mechanisms of MDF by developing the models of mechanics of orthogonal cutting. Langella et al. [6] presented a model for use in predicting thrust during composite materials drilling. They stated that thrust increases due to the feed rate. The most effective way of achieving good quality holes while drilling fiber reinforced plastics is by reducing the thrust. Hocheng [7] presented the prediction and evaluation of thrust force in drilling of composite material using candle stick drill. They indicated that the feed rate and the drill diameter are the most significant factors affecting the thrust force. Surface roughness, which is used to evaluate the quality of a product, is one of the quality attributes in drilling of composites product. Surface roughness was investigated by several authors. Paulo Davim and Pedro Reis [8] have studied the cutting parameters (cutting velocity and feed rate) under machining force (Fm), specific cutting pressure (Ks) and surface roughness (Ra). They conducted experiments based on the methodology of Taguchi and indicated that the surface roughness (Ra) increases with the feed rate. Palanikumar et al [9] developed mathematical model for predicting for minimum surface roughness in turning of GFRP composite using design of experiment. The models are expressed in terms of input parameters (speed, depth of cut, feed rate and fiber orientation). Paulo Davim and Pedro Reis [10] presented a study that evaluates the cutting parameters (cutting velocity and feed rate) under the surface roughness, and damage in milling laminate plates of carbon fiberreinforced plastics (CFRPs). Recently Nemli et al [11] presented the influencing factors affecting the surface roughness of the particle board by contact stylus method. This study showed that the raw material type, pressure, panel density and shelling ratio were some manufacturing parameters affecting the surface quality of the particle board. Akbulut and Koc [12] investigated the effects of fiber-board density, panel temperature, and cutter sharpness on the roughness of surface. This study showed that increasing the panel density and the use of a sharp cutter decreased the roughness of the profiled areas.

Wood and Wood Products, Nova Science Publishers, Incorporated, 2012. ProQuest Ebook Central,

Copyright © 2012. Nova Science Publishers, Incorporated. All rights reserved.

90

K. Palanikumar, S. Prakash and J. Paulo Davim

For the practical machining of MDF, it is necessary to determine the optimal machining parameters to achieve less thrust force, and good surface finish. Optimization of process parameters is the important criterion in the machining process to achieve high quality. To optimize the data based on the experimental results, the traditional statistical regression requires large amount of data that causes the difficulty in treating the typical normal distribution of data and the lack of variant factors. Handling the more demanding multiple performance characteristics are still an interesting research problem [13–16]. The optimization of the multiple machining responses the is much more complicated than optimization of a single machining response. Grey relational theory can solve both unclear problems and those involving incomplete data, can compensate for the short comings of statistical regression, and effectively analyze relationships between sequences in situations involving limited data. The grey relational analysis theory initialized by Deng [17, 18] makes use of grey relational generating and calculates the grey relational coefficient to handle the uncertain systematic problem under the status of only partial known information The grey relational coefficient can express the relationship between the desired and actual experimental results and the grey relational grade is simultaneously computed correspondent to each machining response. The single grey relational grade can provide an optimal constitute of process parameters in which manufacturing simultaneously requests multiple machining responses. The optimal level of process parameters were confirmed through the level with the highest grey relational grade [19, 20]. The grey relational analysis is successfully applied already for machining process. Chang et al [21] used to optimize the injection molding process of short glass fiber reinforced polycarbonates composites using grey relational analysis. Chang and Lu [22] studied the grey relational analysis to obtain the optimal cutting parameter in milling for SUS304 stainless steel. Recently, Palanikumar [23] used Taguchi method with grey relational analysis for the optimization of drilling glass fiber reinforced composites with multiple performance characteristics. In the present work, experiments are designed using Taguchi L27 orthogonal array matrix by considering feed rate, spindle speed and drill diameter as the decision (control) variables and the performance measures namely, surface roughness and thrust force as output responses. Grey relational analysis has been considered for optimization of multiple performance characteristics. Analysis of variance and confirmation test have been conducted to validate the test result.

Wood and Wood Products, Nova Science Publishers, Incorporated, 2012. ProQuest Ebook Central,

Investigation of Optimum Parameters …

91

2. EXPERIMENTAL WORKS 2.1. Description of Experimental Setup and Measurements

Copyright © 2012. Nova Science Publishers, Incorporated. All rights reserved.

In this work, plain MDF panel of IS 12406 standard manufactured through dry processing technology is used for experimentation. The mechanical and physical properties of the medium density fiberboard as per IS 12406 are described in Table 1. The drilling tests are carried out on ARIX VMC 100 machining centre. In order to conduct experiments, the panel materials are cut into 150 x 50 x 9 mm specification. Then the panel is fastened to the drilling fixture attached to the Kistler dynamometer which is mounted on the table. Equal spacing is maintained between successive drilled holes in the board. The drill bit used in the investigation is „Step drill‟ TiN coated carbide drill bits having drill diameter of 4, 8, and 12 mm. The experimental setup, thrust force plot and the drill bits used are presented in Figure 1. Surface roughness is an important factor which is to be controlled. The average surface roughness (Ra) which is mostly used in industries is taken for this study. Table 1. Mechanical and Physical Properties of plain MDF panel tested Tensile strength Modulus of rupture Elasticity modulus Moisture content N/mm2 N/mm2 N/mm2 % 0.8 28 2800 5-10

Figure 1. Experimental setup with thrust force plot. Wood and Wood Products, Nova Science Publishers, Incorporated, 2012. ProQuest Ebook Central,

Density kg/mm3 600-900

K. Palanikumar, S. Prakash and J. Paulo Davim

92

Figure 2. Equipment used for measuring surface roughness and the corresponding plot.

It is the average of the absolute deviation Y(x) of the surface profile from the centerline over a sampling length L and is given by Equation 1. The surface roughness of each hole is taken as the mean of three circumferential readings. Figure 2. shows the equipment used for measuring typical surface roughness profile of average roughness (Ra) and their plot. L

Copyright © 2012. Nova Science Publishers, Incorporated. All rights reserved.

Ra 

1 y ( x ) dx L 0

(1)

2.2. Experimental Design In this study, three factors were studied and their low–middle–high levels are given in Table 2. A L 27 Taguchi standard orthogonal array is adopted as the experimental design. The most suitable array is L 27, which needs 27 runs and has 26 degrees of freedoms (DOF). To check the DOF in the experimental design, for the three levels test, the three main factors take 6 DOFs (3 X 2) and the remaining DOFs are taken by interactions. The 3 level L 27 orthogonal array is shown in Table 3, where the numbers 1, 2 and 3 stand for the values of the factors. The columns chosen for the main factors are 1, 2, and 5. Table 2. Control factors and their levels S.No 1 2 3

Control factor Feed rate Spindle speed Drill diameter

Symbol for coded value f N d

Levels 1(low) 100 1000 4

Unit 2 (middle) 300 3000 8

Wood and Wood Products, Nova Science Publishers, Incorporated, 2012. ProQuest Ebook Central,

3(high) 500 5000 12

mm/min rpm mm

Investigation of Optimum Parameters …

93

Copyright © 2012. Nova Science Publishers, Incorporated. All rights reserved.

Table 3. Taguchi L 27 orthogonal array L27 (313) Run 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27

1 1 1 1 1 1 1 1 1 1 2 2 2 2 2 2 2 2 2 3 3 3 3 3 3 3 3 3

2 1 1 1 2 2 2 3 3 3 1 1 1 2 2 2 3 3 3 1 1 1 2 2 2 3 3 3

3 1 1 1 2 2 2 3 3 3 2 2 2 3 3 3 1 1 1 3 3 3 1 1 1 2 2 2

4 1 1 1 2 2 2 3 3 3 3 3 3 1 1 1 2 2 2 2 2 2 3 3 3 1 1 1

5 1 2 3 1 2 3 1 2 3 1 2 3 1 2 3 1 2 3 1 2 3 1 2 3 1 2 3

6 1 2 3 1 2 3 1 2 3 2 3 1 2 3 1 2 3 1 3 1 2 3 1 2 3 1 2

7 1 2 3 1 2 3 1 2 3 3 1 2 3 1 2 3 1 2 2 3 1 2 3 1 2 3 1

8 1 2 3 2 3 1 3 1 2 1 2 3 2 3 1 3 1 2 1 2 3 2 3 1 3 1 2

9 1 2 3 2 3 1 3 1 2 2 3 1 3 1 2 1 2 3 3 1 2 1 2 3 2 3 1

10 1 2 3 2 3 1 3 1 2 3 1 2 1 2 3 2 3 1 2 3 1 3 1 2 1 2 3

11 1 2 3 3 1 2 2 3 1 1 2 3 3 1 2 2 3 1 1 2 3 3 1 2 2 3 1

12 1 2 3 2 3 1 3 1 2 3 1 2 1 2 3 2 3 1 2 3 1 3 1 2 1 2 3

13 1 2 3 3 1 2 2 3 1 1 2 3 3 1 2 2 3 1 1 2 3 3 1 2 2 3 1

The experimental parameters used and the corresponding responses are given in Table 4. The first column of the table is assigned to the feed rate (f), the second to the spindle speed (N), and the third to the drill diameter (d). The measured responses are thrust force and surface roughness.

Wood and Wood Products, Nova Science Publishers, Incorporated, 2012. ProQuest Ebook Central,

K. Palanikumar, S. Prakash and J. Paulo Davim

94

Table 4. Experimental design, designation and their measured responses

Copyright © 2012. Nova Science Publishers, Incorporated. All rights reserved.

Exp No 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27

Factor f

N

d

100 100 100 100 100 100 100 100 100 300 300 300 300 300 300 300 300 300 500 500 500 500 500 500 500 500 500

1000 1000 1000 3000 3000 3000 5000 5000 5000 1000 1000 1000 3000 3000 3000 5000 5000 5000 1000 1000 1000 3000 3000 3000 5000 5000 5000

4 8 12 4 8 12 4 8 12 4 8 12 4 8 12 4 8 12 4 8 12 4 8 12 4 8 12

Measured responses Designation Thrust force Surface roughness, (N) Ra(micron) f1 N1 d1 120.25 5.48 f1 N1 d2 126.23 6.25 f1 N1 d3 132.53 7.54 f1 N2 d1 89.1 5.45 f1 N2 d2 133.2 6.81 f1 N2 d3 155.93 7.98 f1 N3 d1 98.77 4.41 f1 N3 d2 110.03 5.48 f1 N3 d3 149.18 6.63 f2 N1 d1 138.15 7.85 f2 N1 d2 182.25 8.53 f2 N1 d3 204.98 9.29 f2 N2 d1 161.55 6.62 f2 N2 d2 172.8 7.64 f2 N2 d3 211.95 8.34 f2 N3 d1 182.5 5.65 f2 N3 d2 185.48 6.35 f2 N3 d3 188.78 7.49 f3 N1 d1 211.5 9.47 f3 N1 d2 222.75 10.29 f3 N1 d3 291.6 11.53 f3 N2 d1 231.18 8.57 f3 N2 d2 246.15 9.13 f3N2 d3 252.45 9.93 f3 N3 d1 195.3 7.74 f3 N3 d2 269.1 8.45 f3 N3 d3 262.13 8.96

3. GREY-TAGUCHI OPTIMIZATION Notably, GRA utilizes the mathematical method when analyzing correlations between series comprising a grey relational system, and thereby determines the difference in contribution between a reference series and each compared series.

Wood and Wood Products, Nova Science Publishers, Incorporated, 2012. ProQuest Ebook Central,

Investigation of Optimum Parameters …

95

Copyright © 2012. Nova Science Publishers, Incorporated. All rights reserved.

Figure 3. Block Diagram of Grey Based Taguchi Method.

The compared series are alternative vectors created from sets based on attribute characteristics, which are the larger-the-better and the smaller-thebetter, or optimization of specific values between the maximum and minimum values of an attribute. Applying a GRA algorithm can rank different alternatives by determining their grey relational grades. The grey relational grades of different series can be used to rank various alternatives, where higher values indicate superior alternatives [24]. The optimization of process parameters are carried out using Taguchi method with grey relational analysis. The block diagram for Grey taguchi method is presented in Figure 3. For optimization of process parameters, the following steps are followed: Step 1. Calculate S/N ratio for the corresponding response using the following formula. (1) Larger- the -better quality

S/N ratio (ε) =

1 n 1 10log 10    2  n  i1 yji

(2)

where n= number of replications yij= observed response value where i= 1.2,3..n: j=1,2..k. Wood and Wood Products, Nova Science Publishers, Incorporated, 2012. ProQuest Ebook Central,

K. Palanikumar, S. Prakash and J. Paulo Davim

96

This is applied where maximization of the quality characteristics of interest is required. (2) Smaller –the-better quality

S/N ratio (ε) =

1 n  10log 10   yij2  n  i 1 

(3)

This is applied where minimization of the quality characteristics of interest is intended. (3) Nominal-the best quality 

S/N ratio (ε) =

where μ

=



10log 10   2  

2



(4)

y1  y 2  y3  ....  yn n

Copyright © 2012. Nova Science Publishers, Incorporated. All rights reserved.

 2

σ2 =

( yj  y )  n 1

This is called nominal-the-best type of problem where one tries to minimize the mean squared error around the specified target. High signal-tonoise ratios are always preferred in a Taguchi experiment. The responses considered in the experiment are thrust force and surface roughness which are having smaller-the-better characteristics. Based on equation 3 the S/N ratio of the responses considered are calculated and presented in Table 3. Step 2. Normalize the experimental results of thrust force and surface roughness (Data pre-processing). Normalization is a transformation performed on a single data input to distribute the data evenly and scale it into an acceptable range for further analysis. Data preprocessing involves transforming an original sequence into a comparable sequence. In the grey relational analysis, data preprocessing is first performed in order to normalize the raw data. In this study the experimental responses (thrust force and surface roughness) is normalized in the range of zero to unity, which is called as data preprocessing or grey relational generating.

Wood and Wood Products, Nova Science Publishers, Incorporated, 2012. ProQuest Ebook Central,

Investigation of Optimum Parameters …

97

Table 3. The S/N ratio for the responses thrust force and surface roughness

Copyright © 2012. Nova Science Publishers, Incorporated. All rights reserved.

Exp No 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27

S/N ratio(dB) Thrust force (N) -41.60 -42.02 -42.45 -39.00 -42.49 -43.86 -39.89 -40.83 -43.47 -42.81 -45.21 -46.23 -44.17 -44.75 -46.52 -45.23 -45.37 -45.52 -46.51 -46.96 -49.30 -47.28 -47.82 -48.04 -45.81 -48.60 -48.37

Surface roughness, Ra (micron) -14.78 -15.92 -17.55 -14.73 -16.66 -18.04 -12.89 -14.78 -16.43 -17.90 -18.62 -19.36 -16.42 -17.66 -18.42 -15.04 -16.06 -17.49 -19.53 -20.25 -21.24 -18.66 -19.21 -19.94 -17.77 -18.54 -19.05

Normally, there are three categories of performance characteristics in the analysis of normalized values, ie. the larger-the better, the smaller-the better, and nominal the best. Depending upon the type of performance characteristics, the output responses re normalized using the following formula, i.e. yij is normalized as Zij (0≤ Zij≤ 1) to avoid the effect of adopting different units and to reduce the variability.

Wood and Wood Products, Nova Science Publishers, Incorporated, 2012. ProQuest Ebook Central,

K. Palanikumar, S. Prakash and J. Paulo Davim

98

(4) Larger- the -better quality

yij  min( yij ,i 1,2,... n ) Zij =

max( yij ,i 1,2,... n )min( yij ,i 1,2,... n )

(5)

(To be used for S/N ratio with larger the better) (5) Smaller-better-quality

Zij =

max( yij ,i 1,2,... n )  yij max( yij ,i 1,2,... n )min( yij ,i 1,2,... n )

(6)

(To be used for S/N ratio with smaller the better) (6) Nominal-the-best quality ( yij  Target)- min( yij Target, i 1,2,... n )

Copyright © 2012. Nova Science Publishers, Incorporated. All rights reserved.

Zij =

max( yij Target, i 1,2,... n )min( yij Target, i 1,2,... n )

(7)

(To be used for S/N ratio with nominal the better) In this study the smaller the better is considered for the output responses. Using the Equation 6, output responses are normalized and presented in Table 4. Referring the result from Table 3. thrust force („the smaller, the better‟ characteristic) for carbide TiN step drill bit, the minimum thrust force occur at the 4th trial and is taken as reference sequence equivalent to 1(89.10 N). The maximum cutting forces occur at the 21st trial (291.60 N) and the normalized value for the 1st trial (120.5 N) can be calculated as follows:

max( yij ,i 1,2,... n )  yij max( yij ,i 1,2,... n )min( yij ,i 1,2,... n ) / (291.60 – 89.10)

=

(291.60 – 120.25)

= 0.846

Step 3. Calculate the deviation sequence and grey relational co efficient for the normalized S/N ratio values. Wood and Wood Products, Nova Science Publishers, Incorporated, 2012. ProQuest Ebook Central,

Investigation of Optimum Parameters …

99

Table 4. The result of normalization of two response variables and their deviation sequence Normalized values of responses * Deviation sequence

Copyright © 2012. Nova Science Publishers, Incorporated. All rights reserved.

Exp No

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27

Thrust force (N)

Surface roughness, Thrust force Surface roughness, Ra(micron) (N) Ra(micron)

0.846 0.817 0.786 1.000 0.782 0.670 0.952 0.897 0.703 0.758 0.540 0.428 0.642 0.587 0.393 0.539 0.524 0.508 0.396 0.340 0.000 0.298 0.224 0.193 0.476 0.111 0.146

0.850 0.742 0.560 0.854 0.663 0.499 1.000 0.850 0.688 0.517 0.421 0.315 0.690 0.546 0.448 0.826 0.728 0.567 0.289 0.174 0.000 0.416 0.337 0.225 0.532 0.433 0.361

0.154 0.183 0.214 0.000 0.218 0.330 0.048 0.103 0.297 0.242 0.460 0.572 0.358 0.413 0.607 0.461 0.476 0.492 0.604 0.660 1.000 0.702 0.776 0.807 0.524 0.889 0.854

0.150 0.258 0.440 0.146 0.337 0.501 0.000 0.150 0.312 0.483 0.579 0.685 0.310 0.454 0.552 0.174 0.272 0.433 0.711 0.826 1.000 0.584 0.663 0.775 0.468 0.567 0.639

* Calculated by using Eq (6).

From the data (normalized value) available in Table 4, deviation sequence is computed and presented in Table 4. In grey relational analysis after Wood and Wood Products, Nova Science Publishers, Incorporated, 2012. ProQuest Ebook Central,

K. Palanikumar, S. Prakash and J. Paulo Davim

100

calculating deviation sequence, the grey relational coefficient is calculated and presented in Table 5. For calculating the grey relational coefficient the following formula is used.

 ( y (k ), yi (k ))  ij

 min   max ij (k )   max

(8)

Copyright © 2012. Nova Science Publishers, Incorporated. All rights reserved.

where (a) j=1,2,3..n: k=1,2,..m, n is the number of experimental data and m is the number of responses. (b) yo (k) is the reference sequence (yo (k)=1, k=1,2..m: yj (k) is the specific comparison sequence. (c) Δij=|| yo (k)- yj (k)||= The absolute value of the difference between yo (k) and yj (k) (d) Δ min = min min || yo (k)- yj (k)|| is the smallest value of yj (k) νj€i νk (e) Δ max = max max || yo (k)- yj (k)|| is the largest value of yj (k) νj€i νk (f) δ is the distinguish coefficient. A value of the δ is the smaller and the distinguished ability is the larger δ = 0.5 is generally used [18-20] The deviation sequence Δ0i, Δ max (k) and Δ min (k) for i = 1–27, k = 1– 2 can be calculated as follows: Δ01 (1) = | yo*(1) – y1*(1)| = |1.00- 0.846| = 0.154

(9)

Δ01 (2) = | yo*(2) – y1*(2)| = |1.00- 0.817| = 0.183

(10)

The deviation sequence obtained after the data preprocessing is presented in Table 4. Using Table 4, Δmax and Δmin can be found. The values for Δmax = 1 and Δmin =0. The calculated grey coefficient using Eq.(8) for different drilling condition is presented in Table 5. Step 4 Generate the grey relational grade. After calculating the grey relational coefficient the grey relational grade can be computed by the following Equation 11. It is usual to take the average value of the grey relational coefficient as the grey relational grade [25–27].

Wood and Wood Products, Nova Science Publishers, Incorporated, 2012. ProQuest Ebook Central,

Investigation of Optimum Parameters …

101

Table 5. The calculated grey relational coefficient and grey relational grade Factor

Copyright © 2012. Nova Science Publishers, Incorporated. All rights reserved.

Exp No 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27

f

N

d

100 100 100 100 100 100 100 100 100 300 300 300 300 300 300 300 300 300 500 500 500 500 500 500 500 500 500

1000 1000 1000 3000 3000 3000 5000 5000 5000 1000 1000 1000 3000 3000 3000 5000 5000 5000 1000 1000 1000 3000 3000 3000 5000 5000 5000

4 8 12 4 8 12 4 8 12 4 8 12 4 8 12 4 8 12 4 8 12 4 8 12 4 8 12

Grey relational coefficient Thrust Surface roughness, force (N) Ra(micron) 0.765 0.769 0.732 0.659 0.700 0.532 1.000 0.774 0.697 0.597 0.602 0.499 0.913 1.000 0.829 0.769 0.628 0.616 0.674 0.509 0.521 0.464 0.466 0.422 0.583 0.617 0.547 0.524 0.452 0.475 0.520 0.742 0.512 0.647 0.504 0.536 0.453 0.413 0.431 0.377 0.333 0.333 0.416 0.461 0.392 0.430 0.383 0.392 0.488 0.517 0.360 0.468 0.369 0.439

Grey relational grade

Rank

0.767 0.695 0.616 0.887 0.647 0.551 0.956 0.799 0.622 0.591 0.492 0.444 0.600 0.536 0.464 0.631 0.580 0.520 0.433 0.404 0.333 0.439 0.411 0.387 0.502 0.414 0.404

4 5 9 2 6 13 1 3 8 11 17 19 10 14 18 7 12 15 21 24 27 20 23 26 16 22 25

The grey relational grade is defined as follows:

 j

1 m  yi j k i1

(11)

where γj is the grey relational grade for the jth experiment and k is the number of performance characteristics. Wood and Wood Products, Nova Science Publishers, Incorporated, 2012. ProQuest Ebook Central,

102

K. Palanikumar, S. Prakash and J. Paulo Davim

Copyright © 2012. Nova Science Publishers, Incorporated. All rights reserved.

Figure 4. Grey relational grade graph.

The grey relational grade shows the important relationships among the sequences and indicates their degree of influence (Table 5). The grey relational graph with respect to experiment number is presented in Figure 4. Step 5 From the value of grey relational grade in Table 5, by using Eq. (11), response table is presented in Table 6 and the factor effects and the plotted in Figure 5. Step 6 Considering maximization of grade values (Table 5 and Figure 4) we can obtain the optimal parameter conditions f1N3d1. Step 7 Using the grade value, ANOVA is formulated for identifying the significant factors. The results of ANOVA are given in the Table 7. From ANOVA it is clear that feed rate (34.27%) influences more on drilling MDF panels followed drill diameter (9.2%) and spindle speed (2%).

4. EXPERIMENTAL RESULTS AND DISCUSSION To determine the performance of the machining parameters applied to drilling MDF panel, GRA is applied to determine the optimal factor level conditions. Table 4 lists the multiple performance characteristics of average surface roughness and thrust force with different drilling parameters. The smaller-the-better case is desirable in terms of multiple quality characteristics. In total, twenty seven experiments were conducted to identify the three important input control parameters and normalize the data of two quality characteristics. Notably, GRA mathematical conversion enables limited experiments to obtain comparable coefficients and grades. The grey relational

Wood and Wood Products, Nova Science Publishers, Incorporated, 2012. ProQuest Ebook Central,

Investigation of Optimum Parameters …

103

grade can acquire the optimal combination of control factors and effect contribution of each experimental factor. According to GRA results (Table 5 and Figure 4), the drilling parameter setting from the seventh experiment has the highest grey relational grade, indicating that the best multiple performance characteristics were obtained with the combination of f1N3d1. In terms of grey relational grades, the seventh experiment was followed by the fourth experiment, eighth experiment, first experiment, and twenty first experiment. The grey relational grade for each factor level was derived from the factor levels (Table 4), which are related to the grey relational grades (Table 5). The grey relational grade of each factor level is calculated as the average for the same level in each column; the grey relational grades are shown in intersecting rows. For example, the grey relational grades for factor f at level 1, f at level 2, f at level 3, calculated as follows: γf1 = 1/9(0.767+0.695+0.616+0.887+0.647+0.551+0.956+0.799+0.622) = 0.726 (12)

Copyright © 2012. Nova Science Publishers, Incorporated. All rights reserved.

γf2 = 1/9(0.592+0.493+0.444+0.599+0.534+0.464+0.630+0.579+0.520) = 0.539 (13) Table 6 lists the average grey relational grade for each factor level and its grey relational grade. Since grey relational grades represent the correlation between the reference and comparable sequences, a large grey relational grade means the comparability sequence is closely correlated with the reference sequence. This investigation selects the level that provides the largest average response. Furthermore, these drilling parameters with their levels f1 N3 and d1. list the highest grey relational grade for factors feed rate, spindle speed, and drill diameter, respectively. Table 6. The response table for grey relational analysis Control factors Feed Speed Drill diameter

Average grey relational grade for each factor level Level 1 Level 2 Level 3 Delta Rank 0.7266 * 0.5397 0.4142 0.3214 1 0.5306 0.5467 0.6031 * 0.0724 3 0.6451 * 0.5531 0.4823 0.1627 2

* Optimal level foe each factor. Wood and Wood Products, Nova Science Publishers, Incorporated, 2012. ProQuest Ebook Central,

K. Palanikumar, S. Prakash and J. Paulo Davim

Thus, f1N3d1 comprises the optimal parameter combination for the drilling process. Average feed rate is 100 mm/min, spindle speed is 5000 rpm, and drill diameter is 4 mm. Figure 5 shows the influence of factors with their levels for each drilling parameter. The grey relational grade graph for each factor illustrates the variable response for the three different levels. Table 6 lists the difference between the maximum and minimum value of the grey relational grade for each factor, which indicates that as the difference increases, the effect of drilling factor on the multiple performance characteristics also increases. This difference can be defined as the effect contribution of control factors. Figure 5 presents the graph for effect contribution of control factors, which follow the order of f, d, and N. This order demonstrates that the first effect value of 0.3124 is the f-feed rate, which has a stronger effect than drilling factors; the second effect value of 0.1627 is the drill diameter, the third effect value of 0.0724 is the spindle speed. From Table 5, the optimal parameters achieved are spindle speed at level 3 (5000 rpm), feed rate at level 1 (100 mm/min) and drill diameter at level 1 for achieving better thrust force and surface roughness. It has been found that experiment No. 7 (GRG = 0.956) machining parameter setting has the highest grey relational grade. Therefore experiment No. 7 (spindle speed = 5000 rpm; feed rate = 100 mm/min and drill diameter 4 mm) machining parameters setting is optimal parameter setting for attaining multiple performances simultaneously among 27 experiments. Similarly the response table for thrust force and surface roughness is calculated and presented in Table 7. For analyzing the results, statistical analysis of variance is used. The purpose of the analysis of variance (ANOVA) is to analyze which machining parameters significantly affect the performance characteristic.

grey relational grade

Copyright © 2012. Nova Science Publishers, Incorporated. All rights reserved.

104

0.75

0.75

0.75

0.7

0.7

0.7

0.65

0.65

0.65

0.6

0.6

0.6

0.55

0.55

0.55

0.5

0.5

0.5

0.45

0.45

0.45

0.4

0.4 0

100

200

300

400

feed(m m /m in)

500

600

0

1000

2000

3000

4000

5000

6000

0.4

spindle speed(rpm )

Figure 5. Grey relational grade for drilling parameters levels.

Wood and Wood Products, Nova Science Publishers, Incorporated, 2012. ProQuest Ebook Central,

0

4

8 drill dai(mm)

12

16

Investigation of Optimum Parameters …

105

Table 7. Response graph for thrust force and surface roughness

1

Thrust force Feed Spindle Drill dia rate(mm/min) speed(rpm) (mm) 123.9133 183.8122 158.7

Surface roughness Feed Spindle Drill dia rate(mm/min) speed(rpm) (mm) 6.225556 8.47 6.804444

2

180.9378

182.3633

183.11

7.528889

3

242.4622

181.1378

Delta

118.5489

2.6744

Rank

1

3

Level

7.83

7.658889

205.5033 9.341111

6.795556

8.632222

46.8

3.115555

1.674444

1.827778

2

1

3

2

Table 8. ANOVA table for grade

Copyright © 2012. Nova Science Publishers, Incorporated. All rights reserved.

Source f N d Error Total

DF 2 2 2 20 26

Seq SS

Adj SS

Adj MS

F

0.4452 0.0260 0.1199 0.0506 0.6418

0.4452 0.0260 0.1199 0.0506

0.2226 0.0130 0.0599 0.0025

87.98 5.15 23.69

Based on the results of analysis of variance (Table 8), it has been understood that feed rate is the main parameter which influences thrust force and surface roughness for MDF composites. In this study, the application of GRA was successful and improved the multiple quality characteristics of drilling parameters laser for drilling MDF panel. Hence the grey relational analysis based on Taguchi method for the optimization of the multi response problems is a very useful tool for predicting the thrust force and surface roughness in the drilling of MDF composites. The combination obtained from the analysis of the grey relational grade, f1N3d1.

CONCLUSION Drilling experiments were conducted on a CNC drilling machine with TiN coated carbide step drill and MDF panels as work material. The thrust force and surface roughness values are obtained as out put response under different

Wood and Wood Products, Nova Science Publishers, Incorporated, 2012. ProQuest Ebook Central,

106

K. Palanikumar, S. Prakash and J. Paulo Davim

Copyright © 2012. Nova Science Publishers, Incorporated. All rights reserved.

cutting conditions for various combinations of drilling parameters. The following conclusions were drawn. 1. Grey relational analysis in the Taguchi method for the optimization of the multi response problems is a very useful tool for predicting thrust force and surface roughness in the drilling of MDF panels. 2. From the analysis, it is revealed that feed rate, drill diameter and spindle speed are prominent factors which affect the drilling MDF panels. 3. The best performance characteristics was obtained with TiN coated carbide step drill with low feed rate (100 mm/min) , high spindle speed (5000 rpm) and low drill diameter. 4. From the response table of the average Grey relational grade, it is found that the largest value of the Grey relational grade for the spindle speed of 5000 rpm, drill diameter 4 mm and feed rate of 100 mm/min. It is the recommended levels of the controllable parameters of the drilling operations as the minimisation of the thrust force and surface roughness are simultaneously considered. 5. For the analytical data generated by the twenty seven experiments using three drilling control factors, GRA obtained the grade distribution and verified the drilling parameters that improve multiple performance characteristic. 6. The accuracy can be improved by including more number of parameters and levels.

REFERENCES [1]

[2]

[3] [4]

Tomimura, Y., Khoo, K.C., Ong, C.L. and Lee, T.W. (1988). Rubber wood for Medium Density Fiberboard. Journal of Tropical Forest Science 2(3): 175-179. Dippon, J., Ren, H., Amara, F.B. and Altintas, Y. (2000). Orthogonal cutting mechanics of medium density fiberboards. Forest Products Journal vol 50:no.7/8. Engin, S., Altintas, Y. and Amara, F.B. (2000). Mechanics of routing medium density fiberboard. Forest Products Journal vol 50:no.9. Arul S., Vijayaraghavan L., Malhotra S.K. and Krishnamurthy R. (2006) The effect of vibratory drilling on hole quality in polymeric composites,

Wood and Wood Products, Nova Science Publishers, Incorporated, 2012. ProQuest Ebook Central,

Investigation of Optimum Parameters …

[5]

[6]

[7]

[8]

[9]

Copyright © 2012. Nova Science Publishers, Incorporated. All rights reserved.

[10]

[11]

[12]

[13]

[14]

[15]

107

International Journal of Machine Tools and Manufacture. Vol. 46, pp. 252-259. Engin S., Altintas Y. and Amara F.B. (2000) Mechanics of routing medium density fiberboard, Forest Production Journal. Vol. 590, pp. 65-69. Langella A., Nele L.and Maio A.(2005) A torque and thrust prediction models for drilling of composite materials‟, Composites Part A: Applied Science and Manufacturing. Vol. 36, pp. 83-93. Hocheng H. and Tsao C.C. (2003), „Comprehensive analysis of delamination in drilling of composite materials with various drill bits‟, Journal of Materials Processing Technology. Vol. 140, pp. 335339. Paulo Davim J. and Pedro Reis (2003), Drilling carbon fiber reinforced plastics manufactured by autoclave experimental and statistical study, Materials and Design. Vol. 24, pp. 315-324. Palanikumar K. (2004), „Studies on machining characteristics of glass fiber reinforced polymer composites‟, Ph.D. thesis, Anna University, Chennai, India. Paulo Davim J. and Pedro Reis (2005), „Damage and dimensional precision on milling carbon fiber-reinforced plastics using design experiments‟, Journal of Materials Processing Technology. Vol. 160, pp. 160-167. Nemli.G, Ibrahim Ozturk, Ismail Aydin (2005) Some of the parameters influencing surface roughness of particle board‟, International journal of Building and Environment. Vol 40, pp.1337-1340. Akbulut T, Koc¸ E. (2004) Effects of panel density, panel temperature, and cutter sharpness during edge machining on the roughness of the surface and profiled areas of medium density fiberboard. Forest Prod. J. Vol 54 (12), pp.67–70. E.A. Elsayed, A. Chen, (1993) Optimal levels of process parameters for products with multiple characteristics, Int. J. Prod. Res. Vol 31 (5), pp 1117–1132. Y.S. Tarng, W.H. Yang, (1998) Application of the Taguchi method to the optimization of the submerged arc welding process, Mater. Manuf. Process. Vol.13 (3), pp. 455–467. J.L. Lin, K.S. Wang, B.H. Yan, Y.S. Tarng, (2000) Optimization of the electrical discharge machining process based on the Taguchi method with fuzzy logics, J. Mater. Process. Technol. Vol, 102 pp.48–55.

Wood and Wood Products, Nova Science Publishers, Incorporated, 2012. ProQuest Ebook Central,

Copyright © 2012. Nova Science Publishers, Incorporated. All rights reserved.

108

K. Palanikumar, S. Prakash and J. Paulo Davim

[16] J. Antony, (2001) Simultaneous optimization of the multiple quality characteristics in manufacturing processes using Taguchi‟s quality loss function, Int. J. Adv. Manuf.Technol. Vol, 17 pp.134–138. [17] Deng JL. (1989)Introduction to grey system. J. Grey Syst.; Vol 1(1), pp 1–24. [18] Deng JL. Control problems of grey systems. (1982) Syst. Contr. Lett. Vol, 52, pp.88–294. [19] Lin JL, Lin CL. (2002) The use of orthogonal array with Grey relational analysis to optimize the electrical discharge machining process with multiple performance characteristics. Int. J. Mach. Tools Manufact. Vol, 42, pp.237–44. [20] Narender Singh P, Raghukandan K, Pai BC. (2004) Optimization by grey relational analysis of EDM parameters on machining Al-10%SiCp composites. J. Mater Process Technol. Vol. 155–156, pp.1658–61. [21] Chang. SH, Hawang JR, Doong JL (2000) Optimization of the injection molding process of short glass fiber reinforced polycarbonate composites using grey relational analysis. J. Master Process technology Vol, 97, pp.186-193. [22] Chang CK, Lu HS (2007) The optimal cutting process parameter selection of heavy cutting process in side milling for SUS304 stainless steel. Int. J. Adv. Manufacturing Technology Vol, 34, pp.440-447. [23] Palanikumar. K (2011).Experimental investigation and optimization in drilling of GFRP composites. Measurement.Vol (44), pp 2138-2148. [24] Lu M, Wevers K. (2007) Grey system theory and applications: a way forward. J Grey Syst Vol 10(1, pp.47–54. [25] Tosun, N. (2006) Determination of optimum parameters for multiperformance characteristics in drilling by using grey relational analysis, The International Journal of Advanced Manufacturing. Vol, 28 pp, 450–455. [26] Huang.J.T,. Lin, J.L (2008) Optimization of machining parameters setting of die-sinking EDM process based on the Grey relational analysis with L18 orthogonal array, Journal of Technology Vol. 17. pp, 659–664. [27] Fung, C.P. (2003) Manufacturing process optimization for wear property of fiber-reinforced polbutylene terephthalate composites with grey relation analysis, Wear Vol, 254 pp. 298–306.

Wood and Wood Products, Nova Science Publishers, Incorporated, 2012. ProQuest Ebook Central,

In: Wood and Wood Products Editor: J. Paulo Davim

ISBN: 978-1-62081-973-9 © 2012 Nova Science Publishers, Inc.

Chapter 5

ELECTRICAL RESISTANCE VS MOISTURE CONTENT IN FOUR DIFFERENT TYPES OF CORK PRODUCTS

Copyright © 2012. Nova Science Publishers, Incorporated. All rights reserved.

José R. González-Adrados1,, Florentino González-Hernández1, Juan I. Fernández-Golfin1, José L. García De Ceca1, María Conde García1 and Francisco García Fernández2 1

INIA. Forest Research Centre. Forest Products Dept.. Madrid, Spain 2 Univ Politecn Madrid, Escuela Tecn Super Ingenieros Montes, Dept Ingn Forestal, Madrid, Spain

ABSTRACT Cork is the main non-wood forest product in the western Mediterranean; development of its potential uses requires a detailed knowledge of its properties. One of the variables affecting cork behaviour is moisture content, currently determined by means of electric resistance moisture meters. We have studied the hygroscopic behaviour of different types of cork products (natural and agglomerated), and obtained electrical resistance-moisture curve for each. Specimens were conditioned in climate chambers (20ºC, four different relative humidity –RH,%

Corresponding author e-mail: [email protected].

Wood and Wood Products, Nova Science Publishers, Incorporated, 2012. ProQuest Ebook Central,

110

J. R. González-Adrados, F. González-Hernández et al. conditions). Equilibrium moisture (EMC, %) and electrical resistance (R,) were determined for each specimen. Resistance-moisture curves were adjusted using the equation: log (log (R) + 1) = a + b·EMC. Results showed significant differences for EMC and R-EMC curves among the different types. Individual models adjusted better (R2 between 0.909 and 0.992) than the aggregated model, and showed smaller typical errors (0.01 %EMC - 0.04 %EMC vs 0.05 %EMC). The aggregated model showed systematic errors up to 1 %EMC (for RH=85%). We have concluded that precision of resistance hygrometers increases significantly by using resistance-moisture curves adjusted for each type.

Copyright © 2012. Nova Science Publishers, Incorporated. All rights reserved.

INTRODUCTION Natural cork is a hygroscopic material that loses or gains moisture when air relative humidity varies. This behaviour is due to the presence of hydrophilic points (from the hydroxyl and methoxyl groups) in its cell wall, to which water molecules easily attach through hydrogen bridges (Lequin et al. 2010; de Palacios et al. 2011). Most hydrophilic compounds are polysaccharides (celluloses and hemicelluloses), minoritary structural components in cork cell structure, which is characterized by a thick secondary wall, formed mostly by suberin, a highly hydrophobic polymer, with accounts for approximately 50% of its weight. The other two structural components are lignin and cellulose, located mainly in the primary and tertiary walls. Polysaccharide content varies between 13% and 26% (Pereira 2007). Moisture content is a parameter that greatly influences the physical, mechanical and surface properties of cork (Fortes et al. 2004); that is why it is checked at different times of the manufacturing process of most cork products. In the case of cork stoppers, the main cork product, moisture control is also important from a sanitary point of view, as a way of avoiding microbiological contamination (Chatonnet and Labadie 2003; Mazzoleni and Dallagiovanna 2003). Cork hygroscopic behaviour has been studied applying BET, Henderson and GAB models to cork planks, granulates and small samples, verifying the existence of hysteresis and cork‟s stability through time (de Palacios et al. 2011; González-Adrados and Calvo 1994; Lequin et al. 2010). Two factors affecting hygroscopic behaviour are shape and size of the material employed, and significant differences have been found for cork granulates depending on particle size (Abdulla et al. 2009). In spite of the high volume of agglomerate cork products manufactured, we have not found any literature focused on their hygroscopic behaviour. Nowadays, different types of

Wood and Wood Products, Nova Science Publishers, Incorporated, 2012. ProQuest Ebook Central,

Copyright © 2012. Nova Science Publishers, Incorporated. All rights reserved.

Electrical Resistance vs Moisture Content …

111

agglomerate cork products are manufactured, differing widely in size of granules, type and proportion of glue used, use of synthetic polymers or other additives, and also manufacture by individual molding or extrusion. In the case of stoppers, behaviour differences (resulting from physical properties, oxygen ingress, and sorption capacity) between the different types used in the bottling process are notable (Godden et al. 2001; Kontoudakis et al. 2008; Blake et al. 2009). Hygroscopicity differences may also appear, as has been found for a similar product, wood (Esteban et al. 2008). Bibliography related to determination of cork moisture content by electrical resistance-moisture curves (ER – MC) is scarce. Available studies show that intrinsic electrical conductivity is low (2.9 x 10(-14) S m(-1) at 25ºC), and is considerably affected by the material‟s moisture content and the temperature (MaratMendes and Neagu 2003; Marat-Mendes and Neagu 2004). Like other biological materials (Jensen et al. 2006), dielectric properties differ greatly with product type, and are affected by the process of manufacture: natural or agglomerate, and the presence of additives such as paraffin (Lanca et al. 2007; Lanca et al. 2010). To our knowledge, there are no studies related to adjustment of ER-MC curves, apart from a recent one by our team (FernandezGolfin et al. 2010), which applied Samuelson‟s model (Forsén and Tarvainen 2000), used normally for wood, to two types of stoppers, natural and agglomerate. Its results showed that errors when estimating moisture in stoppers are small (±0.5% for natural stoppers, ±0.3% for agglomerate ones), and that there can be differences between models depending on type of stopper and direction of measurement. In contrast, other factors such as presence of surface treatment or geographic origin of the samples have no influence over adjustments. The aim of this chapter is to analyze the existing differences in hygroscopic behaviour in the four main types of cork stoppers currently manufactured. Additionally, we want to obtain the electrical resistancemoisture curves (ER – MC) for each type of stopper and find out if there are significant differences among them, in the understanding that the results can be applied to other cork products with the same characteristics.

MATERIALS AND METHODS Stoppers Four types of commonly manufactured cork stoppers were studied (Celiège 2006):

Wood and Wood Products, Nova Science Publishers, Incorporated, 2012. ProQuest Ebook Central,

112

J. R. González-Adrados, F. González-Hernández et al. 1. Natural (NAT): completely made of corkwood, top quality. 2. Agglomerated (AGLO): obtained by extrusion of cork granulates (grain size between 1 mm and 8 mm), with addition of a polyurethane glue (10% - 12% w/w). 3. Microgranulated: obtained by molding cork granulates (grain size between 0.25 mm and 2.5 mm) with expanded synthetic materials (up to 51% w/w). Two kinds were selected: a. For still wines (STW; manufacturer declares synthetic microspheres in its composition), and b. For champagne wines (CAVA).

Twenty four stoppers of each type were studied. Microgranulated stoppers came from different manufacturers and had different sizes. Table 1 summarises the characteristics of the material examined.

Copyright © 2012. Nova Science Publishers, Incorporated. All rights reserved.

Conditioning Six cork stoppers of each type were placed in four climate chambers providing the following environment conditions: 20ºC/RH 40±5%, 20ºC/RH 65±5%, 20ºC/RH 80±5% and 25ºC/RH 85±5%. These chambers provided the conditions required to cover the typical cork stopper moisture content range. The chambers were monitored weekly, recording gain or loss of mass experienced by each stopper. Equilibrium was deemed to have been reached when weekly mass variation was ≤0.1%. Stoppers were kept in the chambers until their masses had stabilised (four months). The stopper masses were determined using a METTLER TOLEDO balance (Delta Range PB 303) with 1 mg. Both chambers and balance were subjected to periodic calibration and maintenance in accordance to ISO 17025 Manual of Laboratory Quality. Table 1. Characteristics of the material examined Type

Code

Diameter (mm)

Length (mm)

Density (kg·m-3)

Natural cork for still wines

NAT

24

44

157.64 ± 8.49

Agglomerated for still wines

AGLO

23

44

291.88 ± 1.24

Microgranulated for still wines

STW

24

44

264.22 ± 2.07

Microgranulated for champagne wines

CAVA

31

48

265.48 ± 0.74

Wood and Wood Products, Nova Science Publishers, Incorporated, 2012. ProQuest Ebook Central,

Electrical Resistance vs Moisture Content …

113

To prevent any moisture content loss or gain during the measurement process, all stoppers were placed in small sealed plastic transport containers previously conditioned in the same climate chambers. After introduction into their containers, stoppers from the 25ºC/HR 85±5% chamber became thermally conditioned at 20ºC over 2 h in the 20ºC/HR 80%±5% chamber. When all measurements were done, stoppers were dried in an oven at 103°C ± 3°C to obtain their anhydrous weight. Equilibrium Mositure Content (EMC) was calculated as:

EMC (%) 

Ww  W0  100 W0

Ww: Wet weight (g) W0: Anhydrous weight (g)

Copyright © 2012. Nova Science Publishers, Incorporated. All rights reserved.

Measurement of Electrical Resistance The electrical resistance of each stopper was measured using an AGILENT 4339B high resistance meter (range 103-1015 , accuracy 0.5%, display resolution 5 digits). The measuring specifications were:       

Measuring voltage: 10 V Measuring temperature (material and laboratory): 20ºC Measuring delay: 5 s The electrodes were non-insulated steel needles: Diameter: 1.8 mm Length: 25 mm Separation between electrodes: 8.5 mm

Distance of separation between electrodes was the standard one in cork manufacturing, given the small diameter of stoppers (as small as 23 mm), in contrast with the one commonly employed in the wood industry (separation between electrodes 30 mm). To take readings, the electrodes were introduced into the stoppers in the following manner (section nomenclature for NAT stoppers according Pereira (2007)):

Wood and Wood Products, Nova Science Publishers, Incorporated, 2012. ProQuest Ebook Central,

J. R. González-Adrados, F. González-Hernández et al.

114 



For measurements perpendicular to the grain (transverse section, radial direction,T): introduced 20 mm into the base, perpendicular to the growth rings (NAT type stoppers), avoiding visible pores and anomalies. For measurements parallel to the grain (tangential section, axial direction, L): introduced 11 mm on the lateral surface, parallel to the pores (NAT type stoppers), avoiding visible pores and anomalies.

All measurements were taken after verifying the high resistance meter using a TINSLEY 4721 decade box (in possession of an ISO 17025 calibration certificate). Afterwards, moisture content of each stopper was taken by drying in an oven at 103±2ºC until a constant mass was reached.

Copyright © 2012. Nova Science Publishers, Incorporated. All rights reserved.

Statistical Methods A two-factor (type of stopper and air relative humidity) analysis of variance was used to study the differences in moisture content among types of stoppers. Additionally, t tests for multiple comparison among average moisture content of stoppers were used to define homogeneous groups in each type of environmental conditions considered. SPSS Statistics 17.0.0 (IBM Corp., NY) package was used. Electrical resistance-moisture curves (ER – MC) were obtained through the use of linear regression for each type of stopper. Covariates used were electrical resistance (R, Ω) measurements (both perpendicular (T) and parallel to the grain (L)), and the equilibrium moisture content (EMC, %) following Samuelsson‟s model (Forsén and Tarvainen 2000): log (log (R) + 1) = a + b·EMC Curve fitting was carried out using Statgraphics Centurion XV software, which provided the model coefficients obtained by standard linear regression techniques.

Wood and Wood Products, Nova Science Publishers, Incorporated, 2012. ProQuest Ebook Central,

Electrical Resistance vs Moisture Content …

115

RESULTS Equilibrium Moisture for Different Types of Stoppers

Copyright © 2012. Nova Science Publishers, Incorporated. All rights reserved.

Table 2 and Figure 1 shows equilibrium moisture content (EMC) reached at different conditions for all the types. No cases have been excluded, so there are some slight differences with previously published results (FernandezGolfin et al. 2010). Differences among the types let us sort them out a priori into two groups: that of NAT and AGLO stoppers, with very similar EMC, always higher than those present for STW and CAVA stoppers, which in turn also show some differences between them. Data variability is noteworthy, maximum in the case of natural stoppers (coefficient of variation, CV, between 2.8% and 9.2%), and minimum in the case of STW (CV between 0.6% and 1.3%). Results for multiple comparison tests (Table 2) confirm the existence of statistically significant differences among the different types of stoppers for all conditions, which allows to differentiate three groups: NAT+AGLO, STW and CAVA. Microgranulated stoppers (STW y CAVA) are different from the first group (NAT+AGLO), and between them not only in equilibrium moisture, but in the way it evolves when relative humidity increases. Table 2. Equilibrium moisture content (EMC, %) for each type at 20°C. Characters on the right of each column show homogeneous groups (multiple comparison test) Type

RH = 40%

RH = 65%

RH = 80%

RH = 85%

NAT

4.34 ± 0.32

a 5.58 ±0.25 a 6.83 ± 0.19

a

8.29 ± 0.76

a

AGLO

4.11 ± 0.08

a 5.79 ±0.10 a 6.82 ± 0.14

a

8.30 ± 0.15

a

STW

2.82 ± 0.03

b 4.29 ±0.04 b 5.33 ± 0.07

b

6.74 ± 0.04

b

CAVA 3.11 ± 0.05

c 4.56 ±0.07 c 5.27 ± 0.18

b

6.10 ± 0.09

c

Figure 1 shows average values besides those of one of the theoretical models proposed by the literature for humidity at 35°C (de Palacios et al. 2011). STW and CAVA stopper moistures (at 20°C) were considerably below those of the theoretical model, in spite the difference in temperature.

Wood and Wood Products, Nova Science Publishers, Incorporated, 2012. ProQuest Ebook Central,

116

J. R. González-Adrados, F. González-Hernández et al.

Copyright © 2012. Nova Science Publishers, Incorporated. All rights reserved.

Figure 1. Equilibrium moisture content (EMC) at 20°C for different cork types and comparison with isotherms at 35°C for natural cork slices (de Palacios, 2011).

Analysis of variance results (table 3) reflect that almost all variability in moisture content is absorbed by the model (97.8%). The two factors considered reach a high level of significance (p