Multivariate product adapted grading of Scots pine sawn timber for an industrial customer,part 1: Method development

ABSTRACT

Rule-based automatic grading (RBAG) of sawn timber is a common type of sorting system used in sawmills, which is intricate to customise for specific customers. This study further develops an automatic grading method to grade sawn timber according to a customer's resulting product quality. A sawmill's automatic sorting system used cameras to scan the 308 planks included in the study. Each plank was split at a planing mill into three boards, each planed, milled, and manually graded as desirable or not. The plank grade was correlated by multivariate partial least squares regression to aggregated variables, created from the sorting system's measurements at the sawmill. Grading models were trained and tested independently using 5-fold cross-validation to evaluate the grading accuracy of the holistic-subjective automatic grading (HSAG), and compared with a re-substitution test. Results showed that using the HSAG method at the sawmill graded on average 74% of planks correctly, while 83% of desirable planks were correctly identified. Results implied that a sawmill sorting station could grade planks according to a customer's product quality grade with similar accuracy to HSAG conforming with manual grading of standardised sorting classes, even when the customer is processing the planks further.     

Customer adapted grading of Scots pine sawn timber using a multivariate method

To define new grading rules, or to customize the ones in use in a rule-based automatic grading (RBAG) system of boards, is a time-consuming job for a sawmill engineer. This has the effect that changes are rarely made. The objective of this study was to continue the development of a method that replaces the calibration of grading rule settings by a holistic-subjective automatic grading, using multivariate models. The objective was also to investigate if this approach can improve sawmill profitability and at the same time have a satisfied customer. For the study, 323 Scots pine (Pinus sylvestris L.) boards were manually graded according to the preferences of an important customer. That is, a customer that regularly purchases significant volumes of sawn timber. This manual grading was seen as reference grading in this work. The same boards were also scanned and graded by a RBAG system, calibrated for the same customer. Multivariate models for prediction of board grade based on aggregated knot variables, obtained from the scanning, were calibrated using partial least squares regression. The results show that prediction of board grades by the multivariate models were more correct, with respect to the manual grading, than the grading by the RBAG system. The prediction of board grades based on multivariate models resulted in 76–87% of the boards graded correctly, according to the manual grading, while the corresponding number was 63% for the RBAG system.     

Multivariate product adapted grading of Scots Pine sawn timber for an industrial customer,part 2: Robustness to disturbances

ABSTRACT

Holistic-subjective automatic grading (HSAG) of sawn timber by an industrial customer's product outcome is possible through the use of multivariate partial least squares discriminant analysis (PLS-DA), shown by part one of this two-part study. This second part of the study aimed at testing the robustness to disturbances of such an HSAG system when grading Scots Pine sawn timber partially covered in dust. The set of 308 clean planks from part one of this study, and a set of 310 dusty planks, that by being stored inside a sawmill accumulated a layer of dust, were used. Cameras scanned each plank in a sawmill's automatic sorting system that detected selected feature variables. The planks were then split and processed at a planing mill, and the product grade was correlated to the measured feature variables by partial least squares regression. Prediction models were tested using 5-fold cross-validation in four tests and compared to the reference result of part one of this study. The tests showed that the product adapted HSAG could grade dusty planks with similar or lower grading accuracy compared to grading clean planks. In tests grading dusty planks, the disturbing effect of the dust was difficult to capture through training.     

Optimization potential for perception-oriented appearance classification by simulated sawing of computed tomography-scanned logs of Norway spruce

Abstract

Wood, as a natural material, has favourable properties in both technical and aesthetic aspects. Due to its inherent variability, production of high-quality sawn timber demands adequate control of log conversion, which is feasible with computed tomography (CT) log scanning. Existing appearance grading rules for sawn timber might not fully reflect people's visual perception of wood surfaces, and therefore, an alternative, more perception-oriented appearance classification could be beneficial. An appearance classification of sawn timber based on partial least squares discriminant analysis (PLS-DA) of knot-pattern variables was developed and tested. Knot-pattern variables derived from images of board faces were used in training PLS-DA models against an initial classification of the board faces previously established by aid of cluster analysis. Virtual board faces obtained from simulated breakdown of 57 CT-scanned Norway spruce logs were graded according to the developed classification. Visual assessment of the grading results indicated that the classification was largely consistent with human perception of board appearance. An initial estimation of the potential to optimize log rotation, based on CT data, for the established appearance grades was derived from the simulations. Considerable potential to increase the yield of a desired appearance grade, compared to conventional log positioning, was observed.     

Automatic grading of sawlogs: A comparison between X-ray scanning,optical three-dimensional scanning and combinations of both methods

As sawmills become increasingly efficient, the importance of focusing on value recovery becomes obvious. To maximize value recovery, sawmills require the ability to sort logs according to quality. This study compares four different combinations of three-dimensional (3D) and X-ray scanning that can be used to grade logs automatically. The study was based on 135 Scots pine (Pinus sylvestris L.) logs that had been scanned with both a 3D scanner and an X-ray scanner with two X-ray sources. The percentage of boards with correct grade sawn from automatically graded logs varied from 57% when using only 3D scanning to 66% when using a combination of 3D scanning and X-ray scanning in two directions. The highest possible result, with ideal log grading, was 81%. The result also shows that the combination of a 3D scanner and one X-ray direction results in higher accuracy than a scanner based on two X-ray directions.     

Simulated grading of logs with an x‐ray log scanner‐grading accuracy compared with manual grading

Today sawmills have started to use automatic methods for log grading. The methods used are either optical or gamma‐ray scanners. However, the signals from these scanners are too coarse for accurate log grading and for good control of the sawing process at the single log level. The objective of the present study was to determine the grading accuracy of a log‐scanner with two industrial X‐ray sources. The grading accuracy was compared with the accuracy of manual grading. The results showed that the manual grading of logs and boards is difficult. The accuracy of manual grading was low and the automatic grading systems were more reliable than manual ones. Possibilities for improving the automatic grading systems are discussed.     

Comparing Grade Classification Criteria for Automatic Sorting of Norway Spruce Saw Logs

Grades of centre and side boards from 277 Norway spruce logs were combined to form binary response variables, here denoted as sorting criteria. Four different grading systems were tested. The log geometry variables unevenness, butt taper and top taper were used in logistic regression models. The classification accuracy ranged from 58 to 83%. The accuracy was higher for visual stress grade criteria than for more complex criteria such as the Nordic timber grading rules. The number of tested criteria and thus possible comparisons limited the ability to establish significant differences. The low associations between board grades within logs and between graders, highlight key issues when developing and improving automatic log sorting systems.     

Structural development of Pinus sylvestrís stands with varying initial density: A preliminary model for quality of sawn timber as affected by silvicultural measures

The effects of stand density, thinning and pruning on the quality of sawn timber of Scots pine (Pinus sylvestris L.) were studied based on a computer model. Procedures for the tree growth and crown structure produce a tree, whose properties in terms of location, dimensions and quality (living, dead) of branches are known for any moment of the selected rotation. Consequently, the size, quality (living, dead) and location of knots in sawn timber are recognizable allowing the grading of sawn pieces. The model computations showed that the natural dynamics of the crown system in narrowly‐spaced stands could yield sawn timber nearly of the same quality as resulted from pruning in widely‐spaced stands. Thinning increased the branch growth and the branchiness of the wood yielding lower grade for sawn pieces, respectively.     

Effects of initial spacing on the stem and branch properties and graded quality of Picea abies (L.) karst

The effects of spacing on the stem and branch properties of Norway spruce growing on a good soil in southwestern Sweden were evaluated. Butt logs from sample trees were test sawn, and the boards were machine‐stress graded and graded according to Swedish export rules. Branch diameter and bole diameter both responded positively to an increase in spacing. Quality, as reflected by stress and export grades, increased in response to a decrease in spacing.     

Evaluation of the visual stress grading standard on French Spruce (Picea excelsa) and Douglas-fir (Pseudotsuga menziesii) sawn timber

? In this paper an evaluation of the visual grading standard for softwood sawn timber was made.

? In order to do so, visual grading according to EN 518 and theoretical grading according to EN 338 and EN 384 (measurements of MOE, MOR and density) were applied to lumber. Two batches of 111 and 102 French boards were graded, respectively, of Spruce (Picea excelsa) and Douglas-fir (Pseudotsuga menziesii). For the visual grading the most discriminant criterion was noted: knots, cracks, wane, etc.

? Finally, the results of the two grading methods were compared, and it was shown that the visual stress grading gave quite low results for our two French species.

     

Improving strength grading of timber by grain angle measurement and mechanical modeling

Abstract

Timber strength grading has become a major issue in the European Union during the last years, due to the introduction of the Eurocode 5 and all its related standards. Currently, the most performing strength grading machines are able to locally detect the boards’ knots sizes and positions and interpret this information through adapted grading models. The best lead to improve their accuracy seems to be the introduction of new information about the boards and adapt the mechanical model to take them in account. Small grain angle causes high reduction of clear wood's mechanical properties; local value of slope of grain appears to be of high interest. The aim of this study is to quantify the additional accuracy that grain angle information can bring to an optical scanner used as a strength grading machine. A specific grading model has been developed accordingly, and the results obtained for different machine/model/loading combinations are presented. These results show that slope of grain measurement can significantly improve the accuracy of the optical scanner, for both modulus of elasticity and modulus of rupture estimations.     

北美规格材目测分等概述

目测分等方法是北美一种很重要的锯材分等体系, 即依据目测分等规则通过肉眼观测方式测定木材的缺陷状况进而确定规格材的材质等级。因此, 对木材缺陷与材质等级之间关系的正确把握以及对分等规则的正确应用, 是对规格材进行科学、正确分等的基本要求。合理地依据目测分等规则划分规格材等级, 不仅可以使规格材生产厂家的经济效益最大化, 而且有利于规格材产品使用性能的充分、合理发挥。经过目测分等的规格材具有客观的力学性质以及外观质量等级, 可以为其作为结构用材使用时提供科学的依据。直到今天, 目测分等方法依然是北美进行规格材分等的主要分等方法和基础。文中介绍了北美目测分等方法的发展、原理、意义以及世界上其他一些国家和地区的目测分等标准, 并对中国目测分等标准与北美目测分等规则进行了比较, 希望对完善我国的目测分等规则有所启示。     

Differences of tensile strength distributions between mechanically high-grade and low-grade Japanese larch lumber III: effect of knot restriction on the strength of lumber

It is well known that the presence of knots in structural lumber is one of the most important strengthreducing factors. For practical purposes, visual grading including knot restriction is an effective method for nondestructive evaluation of strength. Edge knot restriction for not only visually graded lumbers but also mechanically graded lumbers is specified in the Japanese agricultural standards for glued laminated lumber. We conducted experimental studies on differences of tensile strength distributions between mechanically high-grade and low-grade Japanese larch (Larix kaempferi, carriere) lumbers daily used for manufacturing glued laminated timbers in Nagano, Japan. We then examined the additional visual grading of mechanically graded lumbers for nondestructive evaluation. We visually graded the prepared mechanically graded lumber by focusing on the knots' area ratio of grouped knots. We confirmed that the higher visual grade related to the stronger tensile strength, similar to our present knowledge; but the effects of knot restriction were reduced when the length of the lumber increased in view of nonparametric 5th percentiles of tensile strength. The differences in the strength/elasticity ratio between mechanically high-grade and low-grade lumber were negligible. It was clear that the length effect on the ratio in visually graded high-grade lumber was smaller than that of visually graded low-grade lumber. It was thus concluded that knot restriction should have little effect on the tensile strength of mechanically graded lumber.     

The flexural properties of young Pinus elliottii x Pinus caribaea var. hondurensis timber from the Southern Cape and their prediction from acoustic measurements

The objectives of this study were to determine the bending strength and stiffness properties of young Pinus elliottii x P. caribaea var. hondurensis timber from the Southern Cape, South Africa, and to evaluate the predictability of these properties from acoustic measurements on standing trees, logs and their sawn boards. The timber has good bending strength (MOR) properties but the mean stiffness (MOE_stat) was significantly lower than that required for structural-grade timber. The mean MOE_stat was also about 30% lower than a P. radiata control sample. The MOE_stat of sawn boards could be predicted from acoustic stress wave measurements on standing trees with a correlation (r) of 0.48 and from acoustic resonance frequency measurements on sawn boards with a correlation of 0.85. The MOR of sawn boards was not significantly correlated with acoustic stress wave measurements on standing trees and only moderately correlated with acoustic resonance frequency measurements on sawn boards (r = 0.45).     

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.     

Influence of initial plant density on sawn timber properties for Douglas-fir (Pseudotsuga menziesii (Mirb.) Franco)

? Context

Mechanical wood properties are increasingly relevant for structural applications and are influenced by growing space availability. Douglas-fir (Pseudotsuga menziesii (Mirb.) Franco) has an increasing market share in Europe and is mainly processed to sawn timber.

? Aim

A sample of 164 thinning trees was taken from two Douglas-fir long-term forestry research plots in Germany. The end-use quality of about 2,000 side and center boards was analyzed as a function of initial plant density (1,000, 2,000, and 4,000 trees per hectare) and log position within the stem.

? Methods

Sawn timber quality was described by knottiness, density, modulus of elasticity, and strength. Explanatory parameters were radial position, longitudinal position, and initial plant density. All boards were strength graded visually and by the grading machine GoldenEye-706 using both X-rays for detecting densities and size as well as position of knots and laser interferometry for detecting eigenfrequency (DIN 4074, DIN 2012; EN 14081-2, CEN 2010).

? Results

High plant density led to better mechanical sawn timber quality. Significant differences were especially observed between 1,000 and 2,000 trees per hectare. The yield of machine strength-graded center boards of strength class C24 increased from 50 to 89 % at low and high initial plant density, respectively.

? Conclusion

Foresters are able to improve end-product quality by controlling planting density in particular. The roundwood price that foresters get should be based on the proportion of higher strength classes within logs to give incentives for a more quality-oriented forest management.     

Simulated yield and quality distribution of sawn timber from final felling in a Norway spruce [Picea abies (L.) Karst.] stand with varying thinning regimes: A case study

Abstract

An integrated simulation tool, formed by integrating the InnoSIM sawing simulation system with the RetroSTEM simulator, was used to convert available wood raw materials from final felling into sawn timber, allowing for calculation of the three-dimensional wood properties of individual stems (stem geometry, heartwood formation, knottiness) as well as the volume, quality and value of sawn timber in a Norway spruce stand with different thinning regimes (unthinned, normal and intensively thinned). Based on the input data of sawing patterns, the simulations indicate that there are relatively small differences (<8%) in the volume yield (m³ha^?1) of sawn goods resulting from sawlogs available from final felling with different thinning practices. However, intensive thinning yielded the largest stem diameters and the greatest volumes (m³ha^?1) of large-sized centre goods (thickness: 50, 63, 75 mm) of rather poor quality. Normal thinning yielded the largest volume of A-grade side boards and centre goods (m³ha^?1), as well as the best total value ([euro]ha^?1) of sawn timber. Differences observed in sawn timber quality distribution can contribute to even more significant variation in value yields, if pricing mechanisms of timber products change to favour higher grade timber products.     

DIE INVLOED VAN RADIALE HARSGEïNFILTREERDE KERNSKEURE OP DIE GESAAGDE OPBRENGS UIT PINUS ELLIOTTII ENGELM.-SAAGBLOKKE / THE EFFECT OF RADIAL RESIN-INFILTRATED SPLITS ON THE YIELD OF SAWN TIMBER FROM PINUS ELLIOTTII ENGELM. SAWLOGS

A number of complaints, regarding the presence of resin-infiltrated splits in seasoned boards produced from Pinus elliottii sawlogs supplied from certain stands in State Forests, led to an investigation to determine the incidence of the defect and to find possible causes.

This paper deals mainly with the study of the first aspect. In the investigation two samples of stems were selected by a semi-random method from one mature stand, being clearfelled, in Entabeni State Forest (Northern Transvaal) and from three stands receiving, or due to receive, their fifth thinning. The clearfelling and thinning had produced logs containing heart shakes and this had given rise to the complaints.

During log preparation it was attempted to prepare as much as possible of the stem length into the longest sawlogs, working as closely as possible to the maximum limits for sweep and knots allowed by the departmental log specifications. This attempt resulted in the preparation of some logs containing slightly more than the permissible defects. Observations were made on stump surfaces of the orientation of shakes, and on log ends of their occurrence higher up in the stem, in an attempt to find a link between shake orientation and the known direction of strong winds, and to determine the height in the stem to which shakes extended from the stump. The orientation study produced no definite proof that wind was responsible, although there were strong indications.

This was followed by sowings up the logs mainly into structural timber, kiln seasoning and grading in two ways, viz A ignoring the resin-infiltrated checks and splits but taking all other defects pertinent to the relevant grades into consideration, and B taking all defects, including the checks and splits into consideration (Table 4). During the grading further measurements were taken on boards containing splits, to determine to what heights in the stem the shakes extended.

From the grading results, total and graded yields of sawn timber were computed, the differences in yields between the two grading procedures being a measure of the incidence and seriousness of the shake defect. Within log classes, (see footnote to Table 2) the yields obtained separately for each 2 cm top U.B. diameter class into which the logs had been segregated prior to sawing, were weighted according to the percentage size distributions of the logs yielded by the clearfelling and thinning operations in the relevant compartments, to obtain a weighted average for the log class.

The heights to which shakes were present in the stems of the two samples are shown in Table 3(a) while their incidence in logs of the various size classes are shown in Table 3(b). It is seen that shakes were more frequent and serious in the larger logs.

The presence of shake in the log samples, taken as a whole, caused only minimal yield losses (Table 4) but had the effect of somewhat reducing average board length (Table 5). However, sawn and graded yields are still excellent and the species should not be discriminated against in afforestation, because of the sporadic occurrence of the defect.

The yield figures should be of some value to sawmillers processing both normal and shake-containing logs of this species.     

落叶松锯材外观分等与机械应力分等的试验研究

根据中国、日本、美国三国的外观分等标准,对落叶松试材进行了以强度为基础的外观分等;采用机械应力分等机对落叶松试材进行了应力分等试验,并分析比较了三国不同的外观分等结果与机械应力分等检测的相关关系。     

Flexural and visual characteristics of fibre-managed plantation Eucalyptus globulus timber

ABSTRACT

The main goal of this study was to investigate the visual characteristics, recovery rate, and flexural properties of sawn boards from a fibre-managed plantation Eucalyptus globulus resource as a potential raw material for structural building applications. The impacts of the visual characteristics, strength-reducing features, and variation in basic density and moisture content on the bending modulus of elasticity (MOE) and modulus of rupture (MOR) of the boards were investigated. The reliabilities of different non-destructive methods in predicting MOE and MOR of the boards were evaluated, including log acoustic wave velocity measurement and numerical modellings. The MOE and MOR of the boards were significantly affected by the slope of grain, percentage of clear wood, and total number of knots in the loading zone of the boards. The normal variation in basic density significantly influenced the MOE of the boards while its effect on the MOR was insignificant. The numerical models developed using the artificial neural network (ANN) showed better accuracies in predicting the MOE and MOR of the boards than traditional multi-regression modelling and log acoustic wave velocity measurement. The ANN models developed in this study showed more than 78.5% and 79.9% success in predicting the adjusted MOE and MOR of the boards, respectively.