Strength grading of structural timber is a process by which value is added to sawn products. It is to the greater part carried out using machine grading based on statistical relationships between so called indicating properties and bending strength. The most frequently applied indicating property (IP) on the European market is the stiffness in terms of average modulus of elasticity (MOE) of a timber piece, although MOE is a material property that varies within timber. A major limitation of today’s grading methods is that the described relationships are relatively poor, which means that there is a potential for more accurate techniques. The main purpose of this research has been to initiate development of more accurate and efficient machine grading methods. Strength of timber is dependent on the occurrence of knots. At the same time, knot measures applied as indicating properties until today have shown to be poor predictors of strength. However, results from this research, and from previous research, has shown that not only size and position of knots but also fibre deviations in surrounding clear wood are of great importance for local stiffness and development of fracture under loading. Thus, development of new indicating properties which take account of knots as well as properties of surrounding fibres, determined on a very local scale, was considered as a possible path towards better strength grading. In the research, results from contact-free deformation measurements were utilized for analysis of structural behaviour of timber on both local and global level. Laser scanning was used for detection of local fibre directions projected on surfaces of pieces. Scanned information, combined with measures of density and average axial dynamic MOE, was applied for calculation of the variation of local MOE in the longitudinal board direction. By integration over cross-sections along a piece, a stiffness profile in edgewise bending was determined and a new IP was defined as the lowest bending MOE along the piece. For a sample of Norway spruce planks, a coefficient of determination of 0.68 was achieved between the new IP and bending strength. For narrow side boards to be used as laminations in wet-glued glulam beams, the relationship between IP and tensile strength was as high as 0.77. Since the intended use of the narrow boards was as laminations in wet-glued beams, the possibility of grading them in a wet state was also investigated. Grading based on axial dynamic excitation and weighing gave just as good results in a wet state as when the same grading procedure was applied after drying. It was also found that the relationship between the new IP and strength was dependent on what scale the IP was determined. Optimum was reached for moving average MOE calculated over lengths corresponding with approximately half the width of investigated pieces. Implementation of the new IP will result in grading that is more accurate than what is achieved by the great majority of today’s grading machines. The new method will probably also be particularly favourable for development of engineered wood products made of narrow laminations. / Hållfasthetssortering av konstruktionsvirke innebär att värdet på sågade produkter ökar. Sorteringen genomförs oftast med maskinella metoder baserade på statistiska samband mellan s.k. indikerande egenskaper och böjhållfasthet. Den indikerande egenskap (indicating property, IP) som är vanligast på den Europeiska marknaden är styvhet uttryckt som ett medelvärde för elasticitetsmodulen (modulus of elasticity, MOE) i ett virkesstycke, trots att MOE är en materialegenskap som varierar i virket. En betydande begränsning med dagens sorteringsmetoder är att de beskrivna sambanden är förhållandevis svaga, vilket innebär att det finns en potential för metoder med högre noggrannhet. Det huvudsakliga syftet med detta doktorandprojekt har varit att initiera en utveckling mot sådana metoder. Hållfasthet hos virke är beroende av förekomst av kvistar. Samtidigt har de kvistmått som fram till idag kommit till användning visat sig vara dåliga prediktorer av hållfasthet. Resultat från såväl denna som tidigare forskning har dock visat att inte bara kvistars storlek och läge, utan också variationen i fiberriktning i omgivande träfibrer, är av stor betydelse för lokal styvhet och brottförlopp under inverkan av last. Utveckling av nya IP som tar hänsyn till såväl kvistar som omgivande träfibrers egenskaper fastställda på mycket lokal nivå bedömdes vara en möjlig väg för att uppnå bättre hållfasthetssortering. I detta doktorandprojekt användes beröringsfri deformationsmätning för analys av det strukturella beteendet hos virkesstycken på såväl lokal som global nivå. Laserskanning utnyttjades för detektering av lokala fiberriktningar projicerade på virkesstyckenas ytor. Med utgångspunkt från skannad information, virkesdensitet och medelvärde för axiell dynamisk elasticitetsmodul kunde variationen i lokal elasticitetsmodul i virkesstyckenas längdriktning bestämmas. Genom integration över tvärsektioner längs ett virkesstycke kunde en profil över hur böjstyvheten i styva riktningen varierade i virkesstyckets längdriktning beräknas. En ny IP definierades som den lägsta elasticitetsmodulen i böjning utmed virkesstyckets längd. För ett urval av granplankor erhölls en förklaringsgrad på 0.68 mellan den nya indikerande egenskapen och böjhållfasthet. För smala sidobrädor avsedda att användas som lameller i våtlimmade limträbalkar var motsvarande förklaringsgrad mellan samma IP och draghållfasthet så hög som 0.77. Eftersom sidobrädorna var avsedda att användas som lameller i våtlimmade balkar genomfördes en studie avseende möjligheten att hållfasthetssortera i vått tillstånd med hjälp av axiell dynamisk excitering och vägning. Det visade sig att sådan sortering gav lika bra resultat som då samma metod användes efter torkning. Sambandet mellan den nya indikerande egenskapen och hållfasthet visade sig också vara beroende av på vilken lokal nivå som egenskapen beräknades. Optimum uppnåddes då den bestämdes som ett glidande medelvärde beräknat över en längd motsvarande ungefär halva virkesstyckets höjd. Implementering av den nya sorteringsmetoden kommer att resultera i sortering som är noggrannare än vad som kan erhållas med det stora flertalet av de sorteringsmetoder som finns idag. Den nya indikerande egenskapen kommer sannolikt att bli särskilt gynnsam att använda för utveckling av ingenjörsmässiga träprodukter bestående av smala lameller.

For manufacturers in the wood industry, an important way to make the production more effective is to automate the process of detecting defects and different attributes on boards. One important attribute on most boards is heartwood and sapwood. This thesis project was conducted at the company MiCROTEC and aims to investigate methods to classify heartwood and sapwood on boards. The dataset used in this project consisted of oak boards. In order to increase the amount of information retrieved from the boards, hyperspectral imaging was used instead of conventional RGB cameras. Based on this data, deep learning models in the form of U-Net and U-within-U-Net architecture as well as different spectral dimensionality reduction methods were developed to segment boards in heartwood and sapwood. The performance of these deep learning models was compared to PLS-DA and SVM. PLS-DA has already been used at MiCROTEC and has been used in this work for comparison as a baseline model.   The result of the thesis work showed that a deep learning approach could increase the F1-Score from 0.730 for the baseline classifier PLS-DA to an F1-Score of 0.918, and that the different spectral reduction methods only had a small impact on the result. The increase in F1-score was mainly due to an increase in precision, since the PLS-DA had a similar recall as the deep learning models.

Through the selection of six representative typical Swedish chairs from each era, I present manufacturing techniques and how the craftsmanship has affected the style of the chairs over time. My main source material is the objects themselves. I chose to analyze the constructions and what can distinguish them apart in techniques, material choices, aesthetics and design. The main part of my work has been to find out what are the most occurring damages and its causes. Most damages from renaissance to biedermeier can be linked to the mortise joint. Through interviews with experienced conservators and with my own experiences as a cabinet maker and conservator, I discuss the concept of the ultimate chair design in terms of stability. I have come to the conclusion that the most solid chair is the eight to twelve framed chairs from the Renaissance, but the chair must be adapted to the intended use.