• Journal of the Korean Wood Science and Technology
• /
• 제43권6호
• /
• pp.701-713
• /
• 2015

소나무와 금강송의 수종 분류를 위해 근적외선(NIR) 분광법과 주성분분석(PCA) 및 부분최소자승법 판별분석(PLS-DA)을 결합하여 수종 분류 모델을 설계하였다. 측정된 모든 NIR 스펙트럼을 이용하여 PCA를 실시한 결과 소나무와 금강송의 수종 분류는 불가능하였다. 그러나 2차 미분된 스펙트럼을 이용하여 시험편의 단면과 심 변재 구분에 따른 수종 분류에서는 변재부에서 수종 분류가 가능하였으며, 특히 방사단면의 변재에서는 명확하게 수종이 분류되었다. 그리고 개발된 PLS-DA 예측 모델을 통해 명확한 수종 분류가 가능하였다. 2차 미분으로 전처리된 스펙트럼을 이용하였을 때 가장 좋은 분류 결과 얻을 수 있었다. 2차 미분 스펙트럼을 이용한 예측 모델은 100%의 분류 정확도를 나타내었으며, 예측 모델의 $R_p{^2}$ 값은 0.86, RMSEP는 0.38로 나타났다. 전처리하지 않은 스펙트럼과 2차 미분 스펙트럼을 이용한 예측 모델의 신뢰도는 유사하였다. 근적외선 분광법과 부분최소자승법 판별분석을 결합한 수종 분류 모델은 소나무와 금강송의 분류에 적합하였다.

• Composites Research
• /
• 제28권2호
• /
• pp.58-64
• /
• 2015

본 논문에서는 샌드위치 복합재 패널로 제작되는 사용후 핵연료 수송용기 충격완충체의 유효등가 유한 요소모델을 제시하는데 목적을 둔다. 샌드위치 복합재 패널은 금속재 면재와 각각 우레탄 폼, 발사목 그리고 레드우드 심재로 구성되었다. 충격완충체의 유효등가 유한요소 모델은 샌드위치 복합재 패널의 저속충격 시험과 해석결과와의 비교를 통해 제시되었으며, LS-DYNA 3D를 사용한 동적 외연 유한요소해석에 의해 수행되었다. 시험과 해석 결과, 충격완충체 샌드위치 패널의 유한요소 모델은 적층쉘 요소의 면재와 솔리드요소의 심재를 사용한 기존의 혼합모델링 기법에 비해 면재와 심재 모두 솔리드 요소를 적용하는 방법이 더 정확한 결과를 나타냄을 확인하였다. 이때 발사목과 레드우드 심재는 요소제거 기능을 갖는 솔리드 요소로 모델링 되는 것이 추천되어진다.

• Journal of the Korean Wood Science and Technology
• /
• 제23권3호
• /
• pp.58-65
• /
• 1995

The possibilities of the estimation of the moisture content(MC) for sitka-spruce (Picea sitchensis Carr.) by measuring density have been investigated. The method is based on the relationships between the wood density and moisture content of wood expressed by Equations (8)~(9). The purpose of this study is examining the estimation of the moisture content of wood by density and the variation of moisture content with annual ring width of wood. The following conclusions were obtained; 1. This method is very convenience because of the average moisture content of wood can be obtained by a simple estimation. This estimation can be made from the easy measurement of the weight and volume of wood. 2. Coefficient of determination between the experimental MCs and theoretical MCs which is calculated by the oven-dry densities of each specimens and Equations (8), (9) is 0.98. This Correlation is very remarkable. Therefore the model Equations on the estimation of moisture content by wood density was available. 3. Relationship between experimental MCs and theoretical MCs which is estimated by average oven-dry density of total specimens showed positive correlation(Fig.2). But from the Fig.4. we can concluded that the number of specimens is two groups. This phenomenon is considered that the variation of MC by the annual ring width from the specimens' observations. Consequently, the MCs of wood by density, is likely to be successful method. can be estimate using by the average oven-dry densities divided with the annual ring widths of wood.

• 설비공학논문집
• /
• 제23권9호
• /
• pp.620-626
• /
• 2011

The purpose of this study is to evaluate a simulation model of a stacking guidance system (SGS) with a roller guide applicable to the mobile harbor. The study used a small-scale model (1/20) made of wood with rollers in order to compare the dynamic analysis with experiment results. The law of similarity was applied for the validation of the scaled model. In order to construct a more realistic simulation model, the damping coefficient of the dynamic model was adjusted to 0.5 Ns/mm for the wood-to-wood contact condition based on the experimental results. Using this validated model, dynamic simulations were also carried out for containers of 20, 30, and 40 tons. The results showed that the reaction force of the roller guide was increased from 74.7 kN to 91.2 kN as the weight of container increased. For the design of a roller guide for SGS, the results obtained in this study can be used to reduce the reaction force by employing a rubber roller or a highly damped rotational joint.

• Journal of the Korean Wood Science and Technology
• /
• 제45권6호
• /
• pp.797-808
• /
• 2017

Automatic wood species identification systems have enabled fast and accurate identification of wood species outside of specialized laboratories with well-trained experts on wood species identification. Conventional automatic wood species identification systems consist of two major parts: a feature extractor and a classifier. Feature extractors require hand-engineering to obtain optimal features to quantify the content of an image. A Convolutional Neural Network (CNN), which is one of the Deep Learning methods, trained for wood species can extract intrinsic feature representations and classify them correctly. It usually outperforms classifiers built on top of extracted features with a hand-tuning process. We developed an automatic wood species identification system utilizing CNN models such as LeNet, MiniVGGNet, and their variants. A smartphone camera was used for obtaining macroscopic images of rough sawn surfaces from cross sections of woods. Five Korean softwood species (cedar, cypress, Korean pine, Korean red pine, and larch) were under classification by the CNN models. The highest and most stable CNN model was LeNet3 that is two additional layers added to the original LeNet architecture. The accuracy of species identification by LeNet3 architecture for the five Korean softwood species was 99.3%. The result showed the automatic wood species identification system is sufficiently fast and accurate as well as small to be deployed to a mobile device such as a smartphone.

• Journal of Microbiology and Biotechnology
• /
• 제11권4호
• /
• pp.598-606
• /
• 2001

A mathematical model is proposed that can depict the kinetics of simultaneous saccharification and fermentation (SSF) using steam-exploded wood(SEW) with a glucose- and cellobiose-fermenting yeast strain. Brettanomyces custersii. An expression to describe the reduction of the relative digestibility during the hydrolysis of the SEW is introduced in the hydrolysis model. The fermentation model also takes two new factors into account, that is, the effects of the inhibitory compounds present in the SEW hydrolysates on the microorganism and the fermenting ability of Brettanomyces custersii, which can use both glucose and cellobiose as carbon sources. The model equations were used to simulate the hydrolysis of the SEW, the fermentation of the SEW hydrolysates, and a batch SSF, and the results were compared with the experimental data. The model was found to be capable of representing ethanol production over a range of substrate concentrations. Accordingly, the limiting factors in ethanol production by SSF under the high concentration of the SEW were identified as the effect of inhibitory compounds present in the SEW, the enzyme deactivation, and a limitation in the digestibility based on the physical condition of the substrate.

• 대한환경공학회지
• /
• 제39권10호
• /
• pp.568-574
• /
• 2017

본 연구는 폐목재로부터 organosolv 공정을 이용해서 리그닌을 분리할 때 영향을 미치는 주요 3개의 반응조건(반응시간($X_1$), 산 촉매의 농도($X_2$) 및 반응온도($X_3$))을 리그닌 회수율(y) 기준으로 최적화하였다. 중심합성계획법(central composite design, CCD)에 따라 반응온도 $136.4-203.6^{\circ}C$, 산촉매 농도 0-2.5%, 반응시간 26.36-93.64 분의 범위를 가진 실험계획을 수행해서 2차 모델식 및 최적조건을 수립하였다. 2차 모델식은 $y=-79.89+0.91X_1+9.8X_2-2.54{\times}10^{-3}X_1{^2}-2.11X_2{^2}$와 같이 얻었으며, 결정계수(coefficient of determination, $R^2$) 값은 0.8531으로 10% 이내의 유의수준에서 유의성을 나타냈다. 2차 모델식에 따라 예측되는 최고 리그닌 회수율은 12.46 g/100 g of dry wood이며 이때 최적 반응 조건은 반응온도 $178.2^{\circ}C$, 산 촉매 농도 2.32%으로 나타났다. 폐목재 대상 organosolv 공정에서의 리그닌 수율은 반응온도보다는 산 촉매 농도의 영향이 더 크게 나타났으며 반응시간에 의한 영향은 없는 것으로 나타났다. 모델의 변동성 분석(analysis of variance, ANOVA)에 따르면 리그닌 수율(y)에 대한 모델식의 유의확률은 p<0.001로 높은 유의성을 보였다. 최적조건에서 모델의 재현성을 검증한 결과 모델식이 실제공정을 적절하게 모사한 것으로 나타났다.

• Journal of the Korean Wood Science and Technology
• /
• 제40권5호
• /
• pp.352-362
• /
• 2012

Mass transfer behavior in wood was estimated through its microscopic structure. The diffusion coefficients which were decided by theoretical equations are influenced by different anatomical properties of wood. From the experiment, the moisture flux was linear to the square root of time. The diffusion coefficients had a regular tendency during the time elapse. During the modeling, it is necessary to understand the limitation of parameters and consider the particular situation to be simulated. In hardwood, because the apertures were not considered, tangential mass transfer simulation was totally different from experiment. As a result, a hardwood model design should consider the apertures which are even on the fiber walls.

• Journal of the Korean Society for Industrial and Applied Mathematics
• /
• 제11권1호
• /
• pp.95-105
• /
• 2007

A mathematical modeling for the drying process of hygroscopic porous media, such as wood, has been developed in the past decades. The governing equations for wood drying consist of three conservation equations with respect to the three state variables, moisture content, temperature and air density. They are involving simultaneous, highly coupled heat and mass transfer phenomena. In recent, the equations were extended to account for material heterogeneity through the density of the wood and via the density variation of the material process, capillary pressure, absolute permeability, bound water diffusivity and effective thermal conductivity. In this paper, we investigate the drying behavior for the three primary variables of the drying process in terms of control volume finite element method to the heterogeneous transport model on one-dimensional grid.

• Journal of the Korean Wood Science and Technology
• /
• 제49권5호
• /
• pp.491-503
• /
• 2021

Wood identification is regularly performed by observing the wood anatomy, such as colour, texture, fibre direction, and other characteristics. The manual process, however, could be time consuming, especially when identification work is required at high quantity. Considering this condition, a convolutional neural networks (CNN)-based program is applied to improve the image classification results. The research focuses on the algorithm accuracy and efficiency in dealing with the dataset limitations. For this, it is proposed to do the sample selection process or only take a small portion of the existing image. Still, it can be expected to represent the overall picture to maintain and improve the generalisation capabilities of the CNN method in the classification stages. The experiments yielded an incredible F1 score average up to 93.4% for medium sample area sizes (200 × 200 pixels) on each CNN architecture (VGG16, ResNet50, MobileNet, DenseNet121, and Xception based). Whereas DenseNet121-based architecture was found to be the best architecture in maintaining the generalisation of its model for each sample area size (100, 200, and 300 pixels). The experimental results showed that the proposed algorithm can be an accurate and reliable solution.