| 1. |
|
|
| 2. |
- Couceiro, José, 1983-
(författare)
-
X-ray computed tomography to study moisture distribution in wood
- 2019
-
Doktorsavhandling (övrigt vetenskapligt/konstnärligt)abstract
- X-ray computed tomography (CT) has been used as an analysing tool for different features in wood research since the beginning of the1980s, but it can also be used to study wood-water interactions in different ways, such as by determining wood moisture content (MC). The determination of wood MC with CT requires two CT images: one at the unknown moisture distribution and a second one at a known reference MC level, usually at oven-dry condition. The two scans are then compared, and the MC is calculated based on the differences between the images. If the goal is to determine the MC in local regions within the wood volume, e.g. when studying moisture gradients in wood drying, wood shrinkage must be taken into account during the data processing of the images. The anisotropy of wood shrinkage creates an obstacle, however, since the shrinkage is not uniform throughout the wood specimen. The technique is thus limited in two ways: it cannot measure MC in local regions and it cannot do it in real time.The objective of this thesis was to study methods to overcome these two limitations. The work explores up to three different methods to estimate local MC from CT images in real time. The first method determines shrinkage for each pixel using digital image correlation (DIC) and is embedded in a broader method to estimate the MC, which verified against a reference. It involves several steps in different pieces of software, making it time-consuming and creating many sources of possible experimental errors. The determination of shrinkage within this method is further explored to enable the implementation of all steps in a unique piece of software. It is shown that it is possible to calculate MC through this method with a root mean square error of prediction of 1.4 percentage points for MC between 6 and 25%.The second method studied succeeds in determining the MC distribution in research applied to wood drying, but the calculation of shrinkage differs from the previous method: instead of calculating shrinkage in the radial and tangential directions, it does so by using the displacement information generated from the spatial alignment of the CT images. Results show that the algorithm can provide consistent data of internal MC distribution of wood at the pixel level that entail continuing researching wood drying processes with an improvement in the accuracy of the MC determination. It represents an improvement regarding the first method because the calculation is fast and highly automatized in a single piece of software.The third method studied is the application of dual energy CT (DECT) to moisture. DECT would provide means for MC calculation at the pixel level and, potentially, in real time, which would mean an important breakthrough in wood drying research. Previous research shows promising results, but its implementation in medical CT, the tool used throughout this work, has shown poor predicting ability. Nevertheless, further research is encouraged.The work done in this thesis proves that it is possible to measure local distribution of MC in wood using CT with accuracy and precision. It also shows that further research could potentially provide a means for MC estimation in real time.
|
|
| 3. |
- Huber, Johannes A. J., 1989-
(författare)
-
Numerical Modelling of Timber Building Components to Prevent Disproportionate Collapse
- 2021
-
Doktorsavhandling (övrigt vetenskapligt/konstnärligt)abstract
- An increasing number of multi-storey buildings are being constructed with engineered wood products, such as glulam or cross-laminated timber (CLT). Multi-storey timber buildings can be safely designed for foreseeable loads, but knowledge is limited concerning their ability to survive unforeseeable events, e.g. accidents, natural disasters or terrorism. Multi-storey buildings with many occupants are required to be able to resist a disproportionate collapse due to an unexpected event. Collapse resistance consists of three lines of defence: I) decreasing the probability of the event, II) decreasing the structural vulnerability and III) increasing the structural robustness. The focus of the present thesis is on defence lines II and III, since they can be affected by engineering considerations.Robustness requires the availability of alternative load paths (ALPs) after an initial structural damage, e.g. the removal of an element. The activation of an ALP, e.g. catenary action, usually happens as the result of a larger displacement than that for which the components are designed, and with the participation of the surrounding structure. Physical tests of removal scenarios are expensive and they are often unable to represent realistic building situations. Numerical models can replace physical tests, e.g. by introducing parameter variations or changed boundary conditions, and can deliver an insight into the underlying mechanisms. Vulnerability depends on the ability of individual components of the structure to withstand loads greater than their intended design loads. To reduce vulnerability, so-called key elements can be be made overly strong. If the uncertainty concerning the material properties is high, e.g. for timber, both nominally stronger and larger amounts of material are required, resulting in inefficient material utilisation. Automated strength grading of sawn timber can narrow the uncertainty, but, even with the current technologies, the variations in the graded material remain large.The predictive power of computerised models for sawn timber offers a great potential for integration with traditional strength grading based on testing combined with statistical models. So far, surface data of sawn timber has been used for numerical models, but X-ray computed tomography (CT) scanning equipment now being installed in sawmills has made it possible to measure the inner structure of logs. Using CT data could make it possible to develop high-fidelity numerical models for predicting the mechanical properties of sawn timber, possibly even before sawing, and this could reduce the uncertainty for structural components and enable the production of high-strength timber. However, attempts to develop CT-based models for timber have been scarce.The objective of the work presented herein was to advance the research front regarding the prevention of disproportionate collapse in multi-storey timber buildings. The work has focused on numerical modelling aspects and on subsystems and components, rather than on entire buildings. The goals were: 1) to describe the state of the art regarding the prevention of disproportionate collapse and its application in timber buildings, 2) to develop models to identify and quantify the ALPs in subsystems and components of CLT buildings, and 3) to develop models of sawn timber based on X-ray CT scanning data, to reduce the uncertainty regarding the mechanical properties of the timber.For goal 1, the literature was reviewed and a survey was conducted among practitioners and researchers in the field. The results provided an extensive overview of the topic and the status quo in the industry, and identified a scarcity of guidelines for multi-storey CLT buildings.For goal 2, non-linear finite element (FE) models were developed for quasi-static pushdown analyses. A study of a platform joint first validated some modelling assumptions. The ALPs in single storeys in a corner bay of an 8-storey CLT building were then studied after the removal of bottom-storey walls. In subsequent parameter variations, the full bay was studied in dynamic analyses. The results identified six different ALPs, which were dependent on the connection capacities and the shear capacity of the floor panels, and indicated that collapse was likely after a double wall removal, but unlikely after a single wall removal. Furthermore, the ALPs in a platform-type CLT floor system were studied in parameter variations of calibrated FE models. The results showed how three different ALPs can develop, depending on the storey, the floor geometry and the connectors.For goal 3, a method was developed for the generation of continuum and FE models from CT scanning data of sawn timber, in which the knots, pith and local fibre orientations were reconstructed. The models gave realistic impressions and they could predict the bending stiffness, strength and initial failure location for Norway spruce sawn timber. The predictions improved, if the eigenfrequency of the sawn timber was also considered for modelling.
|
|
| 4. |
- Myronycheva, Olena, 1975-
(författare)
-
Effectiveness and Evaluation of Wood Protection against Biological Deterioration Caused by Filamentous Fungi
- 2021
-
Doktorsavhandling (övrigt vetenskapligt/konstnärligt)abstract
- Developing a holistic understanding of the biological deterioration of wooden material by fungi in the laboratory and outdoor conditions requires the development of new assessment procedures and tools that allow describing the process with maximum precision and accuracy. Environmental biodeterioration is a complex process including a combination of physical, chemical and biological changes, with many uncertainties limiting the predictability and effectiveness of selected preservatives after laboratory tests. Therefore, in the current thesis, the investigation of the effectiveness of selected wood process parameters and protection systems against fungal growth and evaluation of the applicability of near-infrared spectroscopy for wooden surfaces assessment under fungal attack were accomplished. The mould attack on copper impregnated Scots pine sapwood regulated to a greater extent by planing depth than by the infection method. Air-borne contaminants can heavily occupy the unplaned surfaces, but the extent of such occupation could be reduced with planing and impregnation solutions. Despite the vulnerability of the differently planed and copper-impregnated wood towards mould fungi, mass loss of that wood degraded by white-rot Trametes versicolor was less than 5%. The distribution, quantity, and nature of lipophilic substances beneath the surface in the air- and kiln-dried Scots pine sapwood boards significantly influenced mould fungi attack. It was found that the concentration of total extractives was significantly higher in kiln-dried than in air-dried samples and was higher close to the surface than in the layers beneath. During kiln-drying, a migration front is created at a depth of 0.25 mm with a thickness of about 0.5 mm. The evidence from the previous study is committed to understanding the influence of extractives and other migrating compounds on the unplanned surface and, consequently, on mould growth on that surface of Scots pine sapwood subjected to air and kiln drying. Therefore, a multivariate regression model was developed. The thermal modification at different temperatures of exotic African wood influenced the chemistry. Iroko wood demonstrated stabilisation of pH and different patterns of chemical changes compared to padouk. The open process of wood treatment like heating-and-cooling (i.e. fully soaking heated wood in cold liquor allowing the liquor to penetrate wood partially) can improve wood performance by developing a protective layer beneath the surface on heat-induced curing. However, the applied methacrylic resin demonstrated effectiveness during laboratory testing for biodeterioration but did not perform efficiently during outdoor tests. The test of available commercially of generally recognised as safe (GRAS) compounds and biocidal treatment in laboratory conditions revealed a moderate inhibition effect on protection against biodeterioration. Hyperspectral imaging in the NIR region could be applied to classify thermally modified wood but not for air/kiln-dried Scots pine wood. The use of a portable microNIR spectrometer efficiently demonstrated the separation of no mould and mould specimens in laboratory tests of Scots pine and allowed classifying boards treated with commercial biocides after outdoor weathering.
|
|
| 5. |
- Neyses, Benedikt, 1986-
(författare)
-
Surface Densification of Solid Wood : Paving the Way Towards Industrial Implementation
- 2019
-
Doktorsavhandling (övrigt vetenskapligt/konstnärligt)abstract
- Surface densification of a piece of solid wood results in an increase in density and in hardness in the whole or a part of the densified material, and is one of the ways of improving the properties and value of low-density wood species. Despite efforts for many years, mass commercialisation of either bulk- or surface-densified wood products has not yet been achieved. Most of the previously tested densification methods have limitations in terms of processing speed and integration into the largely continuous wood processing chain, which leads to high production costs. Established methods to eliminate the set-recovery rely either on technologically complex close-system methods or on open-system methods that require relatively long periods of high energy input. For this reason, impregnation with adhesives is used in almost all commercially available densified wood products, and none of them have risen above their status of being niche products.Based on this background, three objectives for this project were formulated: (1) the development of a method for selecting the most suitable wood species for surface densification, (2) showing that surface densification can be carried out in a continuous manner at high process speeds, (3) and researching a fast open-system method to reduce the set- recovery.The method developed for selecting the most suitable wood species for surface densification was based on Lean principles, and it confirmed the suitability of previously studied wood species, such as Scots pine, spruce and poplar. In addition, several suitable alternatives from different parts of the world and from different types of forest were identified. This suggests a high potential for establishing such wood products on a global market level.Two studies using a continuous roller press showed that solid wood can be successfully surface-densified at process speeds of up to 80 m min-1, and that some defects, such as knots, are acceptable in the raw material, but the problem of set-recovery could not however be solved.A screening experiment testing different open-system approaches to reduce the set-recovery highlighted the potential of a novel method using ionic liquids as a plasticiser prior to the surface densification of solid Scots pine. The set-recovery could be reduced to 10%, with the time of high energy input being less than 10 minutes. The Brinell hardness was increased by a factor of 2.7 over that of undensified wood. A study with thermo-gravimetric analysis and digital image correlation showed that the set-recovery almost exclusively happens in the transition zone between the densified and undensified wood cells, where there is less penetration of the ionic liquids.The work accomplished in this project has successfully addressed several gaps in the field of wood densification, firstly, by employing a continuous surface densification process using a roller press, and secondly, by developing and studying a fast open-system pre-treatment with ionic liquids, which greatly reduces the set-recovery. Research will continue on a new band press, facilitating a swift transfer of knowledge between small- scale studies and the continuous surface densification of production-size wooden boards.
|
|
| 6. |
- Olofsson, Linus
(författare)
-
Machine Learning for Appearance Grading of Sawn Timber using Cameras and X-ray Computed Tomography
- 2021
-
Doktorsavhandling (övrigt vetenskapligt/konstnärligt)abstract
- This doctoral thesis deals with a new approach for the appearance grading of sawn timber adapted to the requirements of modern sawmilling industries and timber market situations. Appearance grading of sawn timber allows wood products to be made with a specific visual style due to wood features such as knots. Identifying and grading sawn timber by its visual style is a holistic-subjective task that is inherently suitable for humans. However, with the ever-increasing demand for a faster and more consistent grading operation, humans have been replaced by automatic systems during the past few decades. However, the human perception of the appearance of sawn timber is not something easily defined coherently and concisely for use in automatic systems, resulting in automatic systems struggling to perform appearance grading using conventional rule-based grading. As shown in this thesis, machine-learning methods can be used to teach an automatic system to perform holistic-subjective grading in a way that emulates manual grading while still performing the fast and consistent grading associated with automatic systems. This thesis introduced machine learning for product-adapted appearance grading of sawn timber and studied the use of machine learning to appearance grade sawn timber according to standardised quality grades, using an X-ray computed tomography (CT) scanner and a camera-based board scanner.In the studies presented in this thesis, measurement data from the CT scan-ner and the board scanner was used to create a set of variables only regarding knots. The variable sets and the grades of the sawn timber were modelled by projection to latent structures (PLS) models. The grade of the sawn timber was determined in three ways; firstly, manual grading according to standard-ised quality grades; secondly, called the product grade, the sawn timber was delivered to a wall-panelling customer, and the grade of the sawn timber was determined by the quality yield at the customer; and thirdly, called the image grade, images were extracted from the board scanner and used to estimate the quality yield of the wall-panelling customer manually. The grading in each scanning system was performed using a machine-learning method and a conventional rule-based approach, and their performances were compared.Seven data sets were collected in the studies presented in this thesis, each with a combination of variable sets from the scanners and quality grades as described above. In each study, one or more PLS models were trained to model the relationship between a variable set and a quality grade and used to predict the quality of the sawn timber. A PLS model predicts a score for each piece of sawn timber, and if that score passes a classification threshold, the model assigns a quality grade. This classification threshold could be tuned manually to introduce a bias in the model and thereby change the sorting outcome.When performing standardised appearance grading of dried sawn timber, both a PLS model and rule-based grading achieved about 80% grading accuracy, while a manual grader agreed to 95% with the PLS model and to 81% with the rule-based grading in a verification test. Furthermore, when performing customer-adapted grading of the standardised grades, a PLS model managed an 84% grading accuracy compared to 64% of the rule-based approach. These results show how a conventional rule-based ap-proach struggled with performing customer-adapted grading compared to a PLS model. When performing standardised grading, however, both meth-ods achieved similar grading accuracy, but only the grading performed by the PLS model could not be significantly distinguished from the targeted standardised grades.Using a PLS model to perform product-adapted grading of dried sawn tim-ber resulted in a grading accuracy of about 70%–80% for di˙erent scenarios. These gradings resulted in a quality yield, pass or fail, of about 80% for the wall-panelling customer. According to the customer, rule-based grad-ing did not yield impressive product-adapted results, and no metric was given. Furthermore, this thesis showed that the image grade was as useful as the product grade for training the PLS models, which greatly simplifies the logistical process of creating a data set for training a product-adapted machine-learning model. Had a traceability method been used to collect the data from the scanners automatically, the image grade would allow for completely software-based data collection, which is very much in line with the industry 4.0 concept.A CT scanner enables the appearance grading of virtual sawn timber in the 3D images of the scanned logs, which allows the logs to be sawn for maxi-mum value or quality yield. The CT scanner was made to perform a primary product-adapted grading using either a PLS model or a rule-based approach. In addition to this primary grading, the CT scanner and board scanner were programmed to perform a small secondary grading by limiting a small set of measurements that the CT scanner could not suÿciently account for. For example, large pith deviations were limited in the CT scanner, and rotten knots were forbidden by the board scanner, as these measurements were associated with a high risk of resulting in poor quality wall panels for the customer. With this setup, a dataset of 300 pieces of virtual sawn timber was studied. Using rule-based primary grading, the sawmill delivered about 200 pieces of sawn timber with a product yield of 77% for the customer, after the board scanner rejected 28 pieces (12%). Then, by controlling the classification threshold of a PLS model to make the primary grading very strict, meaning that the log was sawn to only yield very likely high-quality pieces of sawn timber, the sawmill could deliver 114 pieces of sawn timber with a product yield of 90%, after the board scanner rejected 9 pieces (7%). These results show that a PLS model achieved higher grading accuracy and higher quality yield than a rule-based approach. Furthermore, the classifica-tion threshold of the PLS model allows for easy and intuitive control over the sorting outcome, something that the rule-based approach does not support.This thesis showed that a PLS-based machine-learning model could be used to perform holistic-subjective appearance grading by both a CT scanner and a board scanner, where a rule-based approach struggled in all but the most familiar case of standardised grading. Once a framework for a machine-learning method such as PLS has been implemented, this thesis showed the ease of customising and fine-tuning the grading performance to be in line with customers needs. A customer or product adaptation could conceivably be initiated and finalised completely in software by automatically collecting the data using a traceability method, collecting the reference grades needed for training by grading images of sawn timber, and using the intuitive clas-sification threshold to fine-tune the sorting outcome.
|
|
| 7. |
- Shankar, Priyamvada, 1991-
(författare)
-
Data driven crop disease modeling
- 2022
-
Doktorsavhandling (övrigt vetenskapligt/konstnärligt)abstract
- The concept of precision farming deals with the creation and use of data from machinery and sensors on and off the field to optimize resources and sustainably intensify food production to keep up with increasing demand. However, in the face of a growing amount of data being collected, smarter data processing and analysis techniques are needed and have prompted the evaluation and incorporation of artificial intelligence (AI) and machine learning (ML) techniques for multiple use cases right from seeding to harvesting. One such use case that has yet to fully gauge the propositions of AI and ML is crop disease prediction. Since multiple biotic and abiotic factors could be responsible for the occurrence of a disease, modeling requires finding suitable data associated with these factors from multiple farms for an extended time frame and developing smarter models able to capture underlying relationships between them. This thesis presents research conducted to develop data-driven methodologies and optimization approaches for building crop disease models. The objective is realized by breaking down the task into three modules: (i) data collection; (ii) data processing and model building; and finally, (iii) the maintenance of models in production. The traditional data collection approach for disease modeling is through setting up of trials which is expensive and labor-intensive which prompted the evaluation of other novel and free to access data sources. Therefore, in module one two studies were conducted to assess the suitability of social media platforms and remote sensing products. The results show that social media is not a viable option yet due to limited geo-referenced data and ambiguity in categorizing the discussions. On the other hand, vegetation indices derived from multispectral satellite imagery, despite their high spatial granularity, are an interesting addition to the modeling pipeline. Moving on to module two, a study was conducted to demonstrate the process of fusing and preparing data from multiple sources with different formats collected in an extended time frame to be used for model building. The study establishes the relevance of using advanced machine learning models such as deep learning in the prediction of crop diseases. The results show that given the appropriate data preparation process at the right data granularity and the use of some smart tricks, neural network-based models hold the potential to outperform widely used models such as XGBoost. Since neural networks offer advantages such as multimodal learning, transfer learning, and automated feature engineering, which are crucial in building scalable models with heterogeneous data and reduced human effort, the observations of this study led to a follow-up study. This study investigates neural network-based algorithms specifically designed for tabular data and compares them against popular tree ensemble-based models. Apart from acting as a comprehensive analysis of the two families of techniques the results showed that although neural network-based models were not able to outperform tree-based models, they achieved comparable results and allowed for the creation of easier and more accurate models for new diseases by application of transfer learning. Climate change leads to unexpected weather events and modified disease occurrence patterns that cause static models to drift rapidly. Models need to be maintained to ensure they are performing as required. Capturing real-time data and triggering retraining when enough new data has been collected can help maintain models by acting as a feedback loop for model improvement. This was attempted by collecting crowd-sourced data from a disease recognition app, but it was not usable in its current form and required further annotation. Since annotations are expensive and time-consuming, a study for real-life agricultural data retrieval and large-scale annotation flow optimization based on similarity search technique is presented which significantly optimizes the annotation process. The results derived from these studies are highly relevant for progressing the United Nations Sustainable Development Goal of Zero Hunger. It is also expected to ease farmers' anxiety related to yield loss due to crop diseases and enhance their capability of planning and scheduling management practices by giving them an early warning of disease occurrence. The results have been verified through comparison with traditional crop disease prediction methods and interaction with experienced agronomists working for a major AgTech company.
|
|
| 8. |
- Torniainen, Petteri
(författare)
-
Colour – A Reliable Quality-Control Tool for Industrial Thermowood® Production
- 2021
-
Doktorsavhandling (övrigt vetenskapligt/konstnärligt)abstract
- The objective of the work presented in this doctoral thesis has been to correlate the commercial output of ThermoWood® process with a pre-defined, quantifiable, non-destructive measurable parameter - colour, and to show the feasibility of measurements to be the essential part of the internal and external quality-control system. Colour is the most evident property changed of thermally-modified timber (TMT). Colour can be thought of as describing the intensity of the process and as such is a sign of repeatability when working with same treatment methodology and materials. Intensity is usually related to the temperature and time, but elements like water or moisture, steam and pressure present on their own or when combined, have their own influence on the final properties of wood. There are several processes using these above-mentioned components as a modification method for wood. In addition, processes using oil as a heat-transferring element are included in this same modification category. Some of the methods are closely interrelated, with only subtle differences noted. Most of these methods are patented, introduced and processed in Europe, but the market and potential increase for these methods is worldwide. TMT is a new product, with a relatively short industrial scale history (approx. 25 years). During this time huge technical improvements have been made. Processes that are computer aided, remote controlled, and containing recordable systems have replaced the traditional manual processes. Building materials and technical parts regarding air circulation, ventilation, heating and measuring technology with much better thermal endurance has been introduced. All these, combined with several variety of wood species becoming commercially available, have set the frames for a successful enabling of the technology. TMT dominates modified wood. Nevertheless, it is only fraction of treated wood available when looking at preservative treated wood produced globally. Respective comparisons between thermally-modification processes ThermoWood® and WTT® have been undertaken. Due to different treatment atmospheres and wood moisture content, significant chemical changes of water-soluble compounds and degradation products, colour, acidity and strength properties were reported. All the material used in these studies was based on typical dimensions and lengths and they were industrially produced. There were several options for ThermoWood® sources, because the products are classified, and processes are certified. WTT® treatment process was lately presented and the initial experience and knowledge about process was minor. As a consequence. some unexpected reactions and observations were experienced. ThermoWood® represents the major part of TMT commercially produced in Europe. It has two treatment classes Thermo-D and Thermo-S, where the D describes “durability” and S “stability”, respectively. Thermo-D products are usually exposed to external applications and wood species such as Norway spruce and Scots pine are used. Most of the ThermoWood® producers have been audited by certification body since 2006. At present there are several producers from Finland, Iran, Latvia, Poland, Sweden, Turkey and latest Canada which are under continuous control system and certified. New treatment kilns have been supplied to several new countries and as a consequence more potential customers are expected to apply membership and certification. The importance of required treatment parameters in each class, repeatability and internal quality control is highlighted. Non-destructive colour measurements (CIELab) system from wood surface has been applied as one quality control method and extensive colour data has been collected during external audits and even much more in continuous internal quality control. Colour measurements have been entrenched as a daily routine in the production plants. Much of this data has been used in this study and a correlation between L* and b* has been observed in both treatment classes. There have not been a lot of studies published dealing with Norway spruce and Scots pine colour measurements, produced with a dry process under superheated steam, but all the studies found regarding these two wood species suggested that colour might be a reliable quality control tool and the studies done in this thesis strengthened this opinion.
|
|
