No abstract

Teve-se por objetivo avaliar a qualidade da madeira de mogno brasileiro proveniente de plantação comercial para a produção de serrados, aos 16 anos. Como índices de qualidade, avaliou-se as rachaduras de topo de toras (IRTT), os empenamentos e rachaduras de tábuas e o máximo desvio angular (MAD). Ao longo de cinco dias de estocagem das toras houve a evolução do IRTT, porém a abertura das rachaduras foi praticamente constante e inferior a 1 mm. Para tábuas úmidas e secas ao ar, os índices de encurvamento e arqueamento foram menores que o limite máximo de 0,50% preconizado em norma técnica vigente, o que demonstrou a alta estabilidade dimensional da espécie. Uma parcela de 34% das tábuas secas ao ar não apresentou rachaduras de extremidades. O MAD aumentou da medula para a casca e ao longo da altura do fuste, e essa avaliação permitiu classificar a grã da madeira como tendo tendência ao intercruzamento. Houve a diminuição do IRTT com o aumento do MAD e do diâmetro das toras. A madeira jovem de mogno brasileiro possui qualidade satisfatória à produção de serrados.

Microscopic wood identification plays a critical role in many economically important areas in wood science. Historically, producing and curating relevant and representative microscopic cross-section images of wood species is limited to highly experienced and trained anatomists. This manuscript demonstrates the feasibility of generating synthetic microscopic cross-sections of hardwood species. We leveraged a publicly available dataset of 119 hardwood species to train a style-based generative adversarial network (GAN). The proposed GAN generated anatomically accurate cross-section images with remarkable fidelity to actual data. Quantitative metrics corroborated the capacity of the generative model in capturing complex wood structure by resulting in a Fréchet inception distance score of 17.38. Image diversity was calculated using the Structural Similarity Index Measure (SSIM). The SSIM results confirmed that the GAN approach can successfully synthesize diverse images. To confirm the usefulness and realism of the GAN generated images, eight professional wood anatomists in two experience levels participated in a visual Turing test and correctly identified fake and actual images at rates of 48.3 and 43.7%, respectively, with no statistical difference when compared to random guess. The generative model can synthesize realistic, diverse, and meaningful high-resolution microscope cross-section images that are virtually indistinguishable from real images. Furthermore, the framework presented may be suitable for improving current deep learning models, helping understand potential breeding between species, and may be used as an educational tool.

AbstractThe relationship between grain angle and wood properties has not been focus of researches in wood industry. The aim of this study was to establish grain angle variations in commercial Eucalyptus logs and their effects on physical-mechanical wood properties. Wood maximum angular deviation (MAD) was correlated with density, volumetric shrinkage, compressive strength parallel to grain, flexural strength and stiffness as determined by bending and acoustic methods in wood of seven Eucalyptus grandis × Eucalyptus urophylla clones at 13 years old. The relationship between MAD at pith-bark and base-top positions and its effect on the physical and mechanical properties were evaluated. Amplitude of MAD values was small for the seven clones, and the mean was 6.2°. The grain deviation decreased by only 8% in base-top direction, and the correlations among MAD and three logs heights were small and negative (r = −0.13). MAD values presented an increasing trend of 33% in pith-bark direction, with a small positive correlation (r = 0.42). Basic density (BD) presented a significant correlation with the MAD (r = 26). There was no significant correlation between the MAD and volumetric shrinkage, mechanical properties and modulus of elasticity dynamic (determined by stress wave timer, ultrasound or transverse vibration).

Eucalyptus-based glued laminated timber (glulam) was produced to determine the feasibility of a non-destructive method (drilling resistance) to predict the properties of structural elements and add value to low-value hardwood species. Glulam was manufactured with both formaldehyde and bio-based adhesives in two assembly schemes. The shear strength of the glue line (fv0), compression perpendicular to grain (Ec90 and fc90), delamination (DL), main and extended glue lines thicknesses were evaluated. Our findings indicated that the assembly schemes did not statistically influence properties and linear regression models adjusted for drilling resistance satisfactorily estimated the Ec90 and fc90.