Influence of growth stresses and material properties on distortion of sawn timber — numerical investigation

? The board distortion that occurs during the sawing and the drying process causes major problems in the utilisation of sawn timber. The distortion is highly influenced by parameters such as spiral grain angle, modulus of elasticity, shrinkage, growth stresses and sawing pattern.

? In this study a finite element simulation of log sawing and timber drying was performed to study how these parameters interact to affect board distortion. A total of 81 logs with different material combinations were simulated. From each simulated log four boards with different annual ring orientation were studied.

? The results showed that the elastic modulus, shrinkage coefficient and growth stresses had a large influence on the final bow and spring deformation. After sawing of the log into boards, the release of growth stresses was the main contributor to the bow and spring deformation. For boards with low modulus of elasticity, the bending distortion became larger than for the boards with high modulus of elasticity. The twist deformation was very small after sawing but increased significantly during drying of the boards. The results showed that spiral grain angle and the board location within the log were the main contributors to the twist deformation.

     

A theory of cross-sectional shrinkage distortion and its experimental verification

Summary A theory has been developed for calculating the cup and edge distortion that will occur when green boards are dried, or the moisture content of dry boards changes in service. The parameters on which the calculations are based are the annual ring orientation coordinates R and of the boards, and the transverse shrinkage factors. For boards of square-cross-section the theory agrees very closely with the shrinkage predicted by the older theory of Greenhill, MacLean and Keylwerth. For Beilschmiedia tawa specimens 50×50 mm in cross-section the experimental and calculated width and thickness shrinkage were in excellent agreement. For radiata pine boards 200×50 mm in cross-section the experimental cup, edge distortion and shrinkage in width and thickness agreed very closely with the values predicted by the theory.     

Comparing predictability of board strength between computed tomography,discrete X-ray,and 3D scanning of Norway spruce logs

Strength graded boards of Norway spruce (Picea abies (L.) Karst.) are important products for many Scandinavian sawmills. If the bending strength of the produced boards can be predicted before sawing the logs, the raw material can be used more efficiently. In previous studies it is shown that the bending strength can be predicted to some extent using discrete X-ray scanning of logs. In this study, we have evaluated if it is possible to predict bending strength of Norway spruce boards with higher accuracy using computed tomography (CT) scanning of logs compared to a combination of discrete X-ray and 3D scanning. The method was to construct multivariate models of bending strength for three different board dimensions. Our results showed that CT scanning of logs produces better models of bending strength compared to a combination of discrete X-ray and 3D scanning. The main reason for this difference was the benefit of knowing the position of where the boards were cut from the logs and therefore detailed knot information could be used in the prediction models. Due to the small number of observations in this study, care should be taken when comparing the resulting prediction models to results from other studies.     

Predicting the Stiffness of Sawn Products by X-ray Scanning of Norway Spruce Saw Logs

The aim of the study was to investigate the possibility of strength grading Norway spruce [Picea abies (L.) Karst.] saw logs on the basis of simulated X-ray LogScanner measurements and to evaluate the potential accuracy of X-ray LogScanner measurements of green heartwood density and percentage of heartwood. The study was based on 272 logs for strength grading and 29 logs for measurements of green heartwood density and percentage of heartwood. The logs were scanned using computed tomography (CT). After sawing, the modulus of elasticity (MOE) of the centre boards was measured using a strength-grading machine. The CT images were used for simulations of an X-ray LogScanner, resulting in simulated measurements of different variables such as diameter, taper, percentage of heartwood, density and density variations. Multivariate models for prediction of MOE were then calibrated using partial least squares (PLS) regression. The MOE of a log was defined as the mean value of the MOE of the two centre boards. The study showed that the simulated X-ray LogScanner measured the percentage of heartwood and green heartwood density with relatively high accuracy (R ² = 0.94 and R ² = 0.73, respectively, after removing two outliers) and that these and other variables measured by the simulated X-ray LogScanner could be used to predict the stiffness of the centre boards. These predictions were used to sort the logs according to the predicted MOE. When sorting out 50% of the logs (''high-strength'' logs), the percentage of C30 boards increased from 73% (all logs in the study) to 100% (only ''high-strength'' logs). The rest of the logs could then be divided into two groups, one of them with 100% C24 and C30 boards.     

Variation of the dynamic elastic modulus and wave velocity in the fibre direction with other properties during the drying of Eucalyptus regnans F. Muell

Dynamic elastic modulus (E_L) and wave velocity (V) were determined using resonance vibrations from initially green, 100 × 50 mm sample boards of Eucalyptus regnans F. Muell., and after several stages of drying to oven dry. E_L and V were determined from impact induced vibrations and spectral analysis. E_L and V from green wood were positively related to basic density and normal shrinkage, only V was negatively related to green density, and both E_L and V were negatively related to green moisture content and the number of internal checks after drying. The latter relationship has the potential to provide a simple method of segregating highly check prone material. No significant relationships were obtained with collapse. Outside the hygroscopic range, in low shrinkage material, E_L increased little or gradually, while in high shrinkage, collapse prone material, it increased more rapidly, but no clear breakpoint was evident. In the hygroscopic range, E_L increased rapidly in all samples. V increased curvilinearly throughout the entire moisture range, but no difference between collapse prone and non-collapse prone material was observed. Received 16 February 1998     

Predicting lumber grade and by-product yields for Scots pine trees

The purpose of this study was to develop models for estimating yields of lumber grades and by-products of individual Scots pine (Pinus sylvestris L.) trees using stem and crown dimensions as explanatory variables. Two separate data sets were used: (1) one simulated by the process-based growth model, PipeQual, which provides information about stem form and branch properties. The model was used to predict the 3D structure of Scots pine stems from thinning regimes of varying intensity and rotation periods and (2) an empirical data set with detailed 3D measurements of stem structure. The stems were sawn using the WoodCim sawing simulator and the yields and grades of the individual sawn pieces, as well as by-products, were recorded. The sawn timber was classified on A, B, C and D-grades for side and centre boards separately (Nordic Timber grading). By-products were pulpwood, sawmill chips, sawdust and bark.     

Simulation of the yield of knot-free components from Scots pine (Pinus sylvestris L.): A comparative study of star-sawing and square-sawing

The production of knot-free wood is important since the market demands wood without knots for reasons of both appearance and material properties. This work describes a simulation of the removal of knots from star-sawn and square-sawn timber. The efficiency of the two methods is compared in terms of the length of the knot-free components obtained and the volume yield. The simulation was based on data for trees and logs taken from the Swedish Stem Bank. These data were then used to simulate the sawmill process in a computer program called the Virtual Sawmill. Data related to the boards obtained were used in a MATLAB model simulating the cross-cutting of knots. Simulated star-sawing of logs with a top diameter exceeding 230 mm gave a mean knot-free component length of 417±321 mm, while the mean length of knot-free components for simulated square-sawing of the same logs was 298±122 mm. The volume yield of knot-free components from the two sawing patterns was 91% for star-sawing and 87% for square-sawing. For timber with cross-section dimensions of 38×75 mm², the mean length and yield of knot-free components from simulated star-sawing were 451±349 mm and 90%, respectively. In simulated square-sawing, the corresponding values were 263±197 mm and 82%, respectively. This shows that star-sawing has potential for the production of knot-free timber.     

Potential for increasing volume yield by reducing planing allowance

About half the volume of sawlogs ends up as sawn and planed timber. The rest is lost due to drying shrinkage or is turned into by products. As the raw material is a major expense for a sawmill, it is important to reduce waste. To investigate how much the volume yield in the production of sawn and planed timber could be increased by reducing the target dimensions in the sawing stage in a sawmill, two groups of sawn timber were planed under similar conditions. One group consisted of sawn Scots pine timber with a large variation in twist. The other group consisted of sawn Norway spruce timber planed under different pressure settings. Using X-ray images, the minimum dimension for avoiding planer misses was calculated for each board, to find the smallest green target dimension. This was compared to actual measured dimensions. It was found that most sawn timber had unnecessarily large dimensions, and it was also found that a reduction in the target dimensions could increase the volume yield for sawn and planed timber by more than 3 percentage points. Boards with large twist would however need a higher planing allowance. The effect of the planer pressure setting was negligible.     

The mechanical and physical properties of strand boards treated with preservatives at different stages of manufacture

The physical and mechanical properties of boards treated with a preservative at different points during the manufacture process were evaluated to determine the best stage for the application of preservative. A copper boron tebuconazole amine water-based preservative was used in 3% PF-bonded strand boards to achieve five different retentions. Preservative addition was examined at different stages of the manufacture cycle, namely, green strand diffusion, dry strand vacuum treatment, glue-line spray addition, heat and cold quench of manufactured board, and by post-manufacture vacuum treatment. The treatment methods had marked effects on the mechanical properties of some of the boards when the boards with the highest preservative retention were compared with their respective untreated controls. The best results were achieved where the preservative was applied by vacuum treatment of dry strands or by diffusion of green strands before board manufacture. Increasing preservative retention had minimal effects on board properties with these two methods but significant deterioration was noted when the preservative was applied by spraying dry strands or by post-board-manufacture heat and cold quench. An increase of pressing temperature resulted in significant improvements to the mechanical properties of the spray-treated boards. Post-manufacture vacuum treatment of boards caused excessively high losses in internal bond strength.     

Simulation of the yield of knot-free components from Scots pine (Pinus sylvestris L.): A comparative study of star-sawing and square-sawing

Abstract

The production of knot-free wood is important since the market demands wood without knots for reasons of both appearance and material properties. This work describes a simulation of the removal of knots from star-sawn and square-sawn timber. The efficiency of the two methods is compared in terms of the length of the knot-free components obtained and the volume yield. The simulation was based on data for trees and logs taken from the Swedish Stem Bank. These data were then used to simulate the sawmill process in a computer program called the Virtual Sawmill. Data related to the boards obtained were used in a MATLAB model simulating the cross-cutting of knots. Simulated star-sawing of logs with a top diameter exceeding 230 mm gave a mean knot-free component length of 417±321 mm, while the mean length of knot-free components for simulated square-sawing of the same logs was 298±122 mm. The volume yield of knot-free components from the two sawing patterns was 91% for star-sawing and 87% for square-sawing. For timber with cross-section dimensions of 38×75 mm², the mean length and yield of knot-free components from simulated star-sawing were 451±349 mm and 90%, respectively. In simulated square-sawing, the corresponding values were 263±197 mm and 82%, respectively. This shows that star-sawing has potential for the production of knot-free timber.     

Analysis of sawn lumber production from logging residues of branchwood of <Emphasis Type="Italic">Aningeria robusta</Emphasis> and <Emphasis Type="Italic">Terminalia ivorensis</Emphasis>

One way of meeting the supply demands of the wood industry is through the reduction of waste in timber processing. It has been estimated in Ghana that for every tree felled, nearly 50% of the tree volume is left in the forest in the form of branches, crownwood, and stumps. In this study, the potential of utilizing branchwood as raw material from logging residues to help meet the timber demands of the downstream processing sector was analyzed experimentally. The sawing characteristics of the branchwood of Aningeria robusta and Terminalia ivorensis with diameters in the range 10–25 cm using live sawing and cant sawing methods were determined using a randomized complete block design, 2 × 4 factorial with three replications. Experimental results indicate that, the relative lumber value yields for first and second grade boards of A. robusta and T. ivorensis branches were 25 and 20%, respectively, while the green lumber volume yields were 40 and 32%, respectively. A high proportion of sapwood and the occurrence of washboarding in T. ivorensis contributed to that species’ relatively low lumber value and volume yield. The experimental results further indicate that the sawing technique used had no significant influence on lumber yields. However, at 5% level of significance, significant differences in lumber yield existed between some of the diameter classes.     

Modelling of adequate pretwist for obtaining straight timber

Abstract

Wood in general and wooden studs in particular are often distorted owing to uneven shrinkage during the drying process in the sawmill. Twist is often the most detrimental of all types of distortion, and it is caused by spiral grain in combination with variations in moisture content. For sawmills, the objective is to produce dried, straight boards, and one method of dealing with boards with excessive spiral grain is to sort them out and then dry them in a pretwisted position to obtain straight boards after drying. A model using the finite element (FE) method for the simulation of drying twist distortions was first calibrated against laboratory experiments in which boards were dried with and without restraints and pretwists. After the calibration, the FE results were compared with industrial test results for boards that were dried without restraints or with restraints with zero pretwist, i.e. straight restraints. The FE model used an elastic–ideally plastic material model to obtain permanent deformations. The calibration was to set the yield stresses so that there was a good match between FE results and results from the laboratory experiments. The comparison between the industrial test results and the FE results showed that the FE model is capable of realistic simulations of drying boards with and without restraints and presumably also pretwists.     

毛竹竹材物理力学性能研究

为了解不同竹龄毛竹生材含水率、线性干缩率、气干密度、抗弯强度、抗弯弹性模量和顺纹抗压强度等物理性能,对其加工应用的影响,笔者以2-7年生毛竹为材料进行研究,结果表明：竹材的生材含水率、气干干缩率（弦向、径向、纵向）和全干缩率（弦向、径向、纵向）随着竹龄的增加呈减小的趋势;从基部到梢部竹材的生材含水率、线性干缩率均减小;竹材线性干缩率弦向＞径向＞纵向.竹材气干密度、抗弯强度、抗弯弹性模量和顺纹抗压强度均随着竹龄的增加呈增大的趋势,尤其是3年生竹材的这些物理力学性能与2年生差异显著,但3年后生竹材差异不大;从基部到梢部竹材的气干密度、抗弯强度、抗弯弹性模量和顺纹抗压强度逐渐增加.综合考虑毛竹的物理力学性能和竹林的经济效益,适合采伐的是3年后生竹材,锯截之后的竹材也应根据部位不同进行区分,以便于加工应用过程中合理利用,提高产品的理化性能和质量的稳定性.     

Finite element study of growth stress formation in wood and related distortion of sawn timber

Lack of straightness in timber is the most frequent complaint regarding solid (and laminated) timber products worldwide. Nowadays, customers demand higher quality in the shape stability of wood products than they did earlier. The final distortion of timber boards is mostly caused by moisture-related stresses in wood (drying distortions) and growth-related stresses (distortions appearing when logs are split up to timber boards by sawing). To get more knowledge on how these distortions can be reduced in wooden products, there is a need for improved understanding of this material behaviour through good numerical tools developed from empirical data. A three-dimensional finite element board distortion model developed by Ormarsson (Doctoral thesis, Publ. 99:7, 1999) has been extended to include the influence of growth stresses by incorporating a one-dimensional finite element growth stress model developed here. The growth stress model is formulated as an axisymmetric general plane strain model where material for all new annual rings is progressively added to the tree during the analysis. The simulation results presented include how stresses are progressively generated during the tree growth, distortions related to the redistribution of growth stresses during log sawing, and distortions and stresses in drying reflecting the effects of growth stresses. The results show that growth stresses clearly vary during tree growth and also form a large stress gradient from pith to bark. This in itself can result in significant bow and crook deformations when logs are sawn into timber boards. The distortion results from the simulations match well with the results observed in reality. The parametric study also showed that the radial growth stress distribution is highly influenced by parameters such as modulus of elasticity, micro fibril angle and maturation strain.     

Rotational position of curved saw logs and warp of the sawn timber

With the development of scanning technology in sawmills, it is possible to optimise log rotational position when sawing. However, choosing a different rotational position than horns down might be detrimental for the board shape after drying, especially for curved logs. Thus, there is a need to investigate at what level of log curve it is possible to freely rotate logs without causing board warp. This study was carried out through a test sawing that was conducted at a sawmill situated in the middle of Sweden. The tests were made on 177 Norway spruce logs, with varying amount of curve. Half of the logs were sawn in the horns-down position, half were sawn rotated perpendicular to horns down. Log shape and warp of the dried boards were measured. The results indicated a relationship between board spring, log curve and choice of rotational position. Furthermore, board bow was related to log curve but not rotational position. It can be concluded that for straight logs, with a bow height of less than 15 mm, an unconventional rotational position does not cause excess spring in the boards. Bow and twist are not affected by the rotational position at all.     

不同下锯法锯解的分析评价

本文就制材生产中通常使用的三种下锯法,采用计算机进行模拟锯解,然后进行分析评价,找出三种下锯法对主产锯材宽度、原木出材率和经济效果的影响。为合理利用原木、选择适当的下锯法,提供了理论依据。     

Finite element analysis of bending stiffness for cross-laminated timber with varying board width

ABSTRACT

Cross laminated timber (CLT) is a wood panelling building system that is used in construction, e.g. for floors, walls and beams. Because of the increased use of CLT, it is important to have accurate simulation models. CLT systems are simulated with one-dimensional and two-dimensional (2D) methods because they are fast and deliver practical results. However, because non-edge-glued panels cannot be modelled under 2D, these results may differ from more accurate calculations in three dimensions (3D). In this investigation, CLT panels with different width-to-thickness ratios for the boards have been simulated using the finite element method. The size of the CLT-panels was 3.0 m × 3.9 m and they had three and five laminate layers oriented 0°–90°–0° and 0°–90°–0°–90°–0°. The thicknesses of the boards were 33.33, 40.0, and 46.5?mm. The CLT panel deformation was compared by using a distributed out-of-plane load. Results showed that panels with narrow boards were less stiff than wide boards for the four-sided support setup. The results also showed that 2D models underestimate the displacement when compared to 3D models. By adjusting the stiffness factor k88, the 2D model displacement became more comparable to the 3D model.     

Effect of mean moisture content on the steam reconditioning of collapsed <Emphasis Type="Italic">Eucalyptus regnans</Emphasis>

Sixteen quarter-sawn boards (100 × 40 mm²) of regrowth Eucalyptus regnans (Mountain Ash) were conditioned to various moisture contents to investigate the effect of mean moisture content on collapse recovery. The results support the recommendation that boards should be reconditioned at a mean moisture content of between 15 and 20%. It is likely that the actual amount of collapse recovery was nearly as good for moisture contents up to about 25%. The main disadvantage with reconditioning boards with a moisture content of between 20 and 25% was the additional normal shrinkage that occurs because of the early reduction or removal of drying stresses. The samples in this study were dried under mild conditions for long periods of time to minimise the presence of moisture gradients.     

Estimation of shrinkage coefficients in radial and tangential directions from CT images

The aim of the present work was to use the displacement information generated from the spatial alignment in order to compute wood shrinkage in the radial and tangential directions in computed tomography (CT) images, and to compare the results with those obtained with computer-aided design software on the same images. To estimate the shrinkage coefficients from tomography images, wood specimens in the green state, equilibrium moisture content 15% and 8% state and oven dry condition were scanned. Specimens were taken from Norway spruce and Scots pine logs. The root-mean-square-error calculations showed acceptable small differences between the two measuring methods, which means that the algorithm is a useful tool for estimating the shrinkage coefficients in radial and tangential direction from CT images. This provides an image processing tool to monitor the dimensional changes during the drying and heat treatment process.