This thesis consists of studies of the economics of mineral resource exploitation and their environmental impact, reported in four papers. Paper I replicates two seminal papers from the early 2000s which indicated that the correlation between genuine savings and future consumption was weaker than theory predicted, at least when genuine savings were measured using the World Bank estimates. The aim of this paper was to examine whether the correlation has become stronger over time, on the back of policy changes in resource-rich countries and of revisions to the World Bank estimates. The results indicate that the correlation between genuine savings and future consumption growth may be stronger for poorer countries than for richer, and for sub-Saharan Africa, the theoretical predictions appear to hold. Paper II estimates resource rents for Mozambican coal mining using company-level data employing the residual value method devised by SEEA-Energy (the multi-purpose conceptual framework for organizing energy related statistics) and compares the findings with the World Bank’s estimates of coal rents. The latter estimates are often used in the resource curse literature and guide the World Bank’s policy notes, forming the baseline of their policy advice on resource exploitation. On average, the results show unit coal rents for the 2011–2020 period that are less than half of the World Bank estimates, suggesting that the World Bank overstates coal rents for Mozambique considerably. The main driver of this discrepancy is the World Bank’s underestimation of extraction costs. The results suggest that studies employing resource rent estimates should consider sensitivity analyses and greater use of local data, and that the World Bank’s policy advice should be interpreted cautiously to avoid unreasonably high expectations. Paper III implements a choice experiment to examine the willingness to pay for improved water supply and forest use currently degraded due to mining activities, with a specific focus on how different household groups near coal mines value these services. The survey sample comprised 419 households in Moatize, Mozambique. Overall, the estimation results obtained via a latent class conditional logit model indicate that four classes of households exist, with different preferences across classes. The main drivers of class membership include gender, income, education and age. All classes express dissatisfaction with the status quo, and improvements in water supply is generally highly valued. However, the four groups in some cases express quite different valuations of the proposed improvements. Development interventions attempting to address the environmental impacts of mining should therefore consider the heterogeneous preferences of the intended beneficiaries, and how experiences from previous interventions are likely to affect attitudes toward new interventions. Paper IV uses the contingent valuation method to estimate displaced and resettled Mozambican households’ willingness to pay to restore the landscape where they used to live before mining began. The study results indicate that on average households are willing to contribute about 9 working days per month. Results further indicate that resettled respondents have been adaptive and used the monetary compensation they were given to buy more productive land to offset the land lost due to resettlement, but still see themselves as worse off compared to before resettlement. A possible explanation is that they now are far away from the marketplaces and the river, making it difficult to develop new income sources and to get access to water. Mitigation interventions and future resettlements should think carefully about the resettlement site selection.

In light of the increased demand for raw materials to support the transition away from a fossil-fuel society, mineral exploration for future raw material sources is imperative, requiring a holistic mineral systems approach. This thesis focuses on the Pahtohavare area in the Kiruna mining district where three epigenetic Cu ± Au deposits of unknown age and structural setting occur within 0.5 km of a pre-orogenic stratiform Cu (Fe ± Zn) deposit, 2 km of the Rakkurijärvi iron oxide-copper-gold (IOCG) deposit, and 5 km of the giant Kiirunavaara iron oxide-apatite (IOA) deposit, which may have shared mineral system ingredients during ore formation. The aim of the thesis is to characterize the mineral system ingredients of the epigenetic Pahtohavare deposits utilizing an up-to-date regional tectonic framework and to assess parallels between other mineral systems in the district. The following ingredients were investigated in this study: 1) driving force, using U-Pb zircon geochronology of local igneous intrusions, 2) transport pathways and structural traps, utilizing structural analysis and field mapping, 3) metal sources, based on sulfide trace element and carbonate REE geochemistry, and 4) ligand and fluid sources, using S, Sr, and Nd-isotope analyses.

U-Pb zircon geochronology of intrusions in the district where Cu-mineralization is spatially proximal yields ages unanimously from the early phase of the Svecokarelian orogeny (ca. 1923-1867 Ma including age uncertainties). These ages suggest magmatism provided a thermal energy drive for the mineral systems developing during this time such as the Kiirunavaara IOA and Rakkurijärvi IOCG.

Structural analysis and petrographic investigation from the Pahtohavare area show that the host rocks for the Pahtohavare deposits are folded into a noncylindrical, SE-plunging anticline and that cleavage measurements mirror the fold geometry. Foliation trails, mylonitic fabrics, and porphyroclasts with pressure shadows characterize the S0/S1 fabric as tectonic and constrains the fold as F2, indicating the folding event occurred during the late phase of the Svecokarelian orogeny (ca. 1.81-1.78 Ga). The epigenetic Pahtohavare deposits sit in brittle-ductile second-order structures that cross cut earlier foliation and the F2 fold, indicating that the deposits formed ca. 80 m.y. after the IOA and IOCG deposits in a distinct tectonic regime.

Petrographic investigation of microstructures identifies remobilization textures linking the pre-orogenic stratiform Cu (Fe ± Zn) and late-orogenic Pahtohavare Cu ± Au deposits. Comparison of Co/Ni ratios between structurally constrained early and late sulfide generations (0.1-10 vs. 0.1-100) indicate the late-orogenic fluids were capable of efficiently transporting and concentrating Co, implying hot and saline conditions. However, a weak Co/Ni ratio mixing trend between the two generations suggest the late-orogenic fluids primarily sourced metals from the pre-existing host rocks and sulfides, rather than an external reservoir. This is supported by flat REE (chondrite normalized) patterns from ore-related carbonate, reflecting a source from the mafic host rocks.

Both the pre- and late-orogenic sulfide generations have overlapping δ34S values ranging between ca. -20‰ and ca. -5‰ to near-zero values. The sulfur source for both mineral systems likely originated from local graphite schist horizons (ca. -22‰ to -21‰) and the volcanic host rocks. 87Sr/86Sr values for ore-related carbonates from the epigenetic Pahtohavare deposits are radiogenic (ca. 0.7095-0.7154) and form a mixing trend with strongly radiogenic (ca. 0.7134-0.7740) non-ore-related syn-sedimentary carbonate in biotite-altered rocks. This suggests a metamorphic fluid derived from crustal and Rb-rich rocks was responsible for the formation of the ore-related carbonate. Therefore, the Pahtohavare epigenetic deposits are suggested to have formed from metamorphic fluids during prograde metamorphism modified by the pre-enriched and altered host rocks in the Pahtohavare area.

The results of this thesis suggest that late-orogenic tectonism played a key role in reactivating mineral system ingredients, highlighting the potential of second-order structures near pre-existing mineral deposits as a strategic exploration target for Cu-mineralization in the region.  

Self-supervised representation learning (SSL) has emerged as a fundamental paradigm in representation learning, enabling models to learn meaningful representations without requiring labeled data. Despite its success, SSL remains constrained by two core challenges: (i) lack of robustness against real-world distribution shifts and adversarial perturbations, and (ii) lack of domain-awareness, limiting its usability beyond natural scenes. These limitations arise from the generic invariance assumptions in SSL, which rely on predefined augmentations to learn representations but suffer to generalize when exposed to unseen environmental distortions, adversarial attacks, and domain-specific nuances. Existing SSL approaches—whether contrastive learning, knowledge distillation, or information maximization—do not explicitly account for these factors, making them vulnerable in real-world applications and suboptimal in specialized domains.

This thesis aims to enhance both robustness and domain-awareness in a modular, plug-and-play manner, ensuring that the advancements are applicable across different joint embedding architecture and method (JEAM)-based SSL approaches and adaptable to future developments in SSL. To achieve this, this thesis follows a guiding principle-leveraging invariant representations to improve robustness and domain-awareness in a modular and plug-and-play manner without altering fundamental SSL objectives. This principle guides that improvements can be seamlessly integrated into existing and future SSL approaches.

To systematically address the above-stated core challenges, this thesis begins with a foundational study of SSL approaches, identifying the common schema that underlies different SSL approaches. This unification provides a conceptual view of SSL methods, allowing us to isolate the domain-sensitive and domain-agnostic components across approaches. This conceptual outcome set the stage to establish precisely where improvements are needed to enhance robustness and domain-awareness across methods as current SSL methods fail under real-world challenges.

Next, the thesis conducts a large-scale empirical evaluation of existing SSL methods against relevant robustness benchmarks, uncovering their failures under distribution shifts caused by real-world environmental challenges. This evaluation reveals a significant decline in the robustness performance of existing SSL methods across different SSL approaches. It establishes the fundamental research gap and motivates the advancements introduced in this thesis.

The first advancement focuses on robustness against distribution shifts, particularly geometric distortions such as perspective distortion (PD), which are prevalent in real-world environment but not addressed by existing SSL methods. Since PD introduces nonlinear spatial transformations, standard affine augmentations fail to model these effects, leading to degraded representation stability. To address this, this thesis introduces Möbius-based mitigating perspective distortion (MPD) and log conformal maps (LCM), mathematically grounded transformations that enable robustness without requiring perspective-distorted training data and estimation of camera parameters. These methods are additionally adapted to multiple real-world computer vision applications—including crowd counting, object detection, person re-identification, and fisheye view recognition—showcasing their effectiveness. Further, addressing the non-availability of dedicated perspectively distorted benchmark, ImageNet-PD robustness benchmark is developed to fill the gap.

Beyond environmental challenges, another critical real-world challenge is adversarial attacks. SSL methods are highly susceptible to adversarial attacks, as the learned representations lack perturbation-invariant constraints. Existing adversarial training approaches in SSL rely on brute-force attack strategies, which fail to adapt dynamically. To address this, this thesis introduces adversarial self-supervised training with adaptive-attacks (ASTrA), where attack strategies evolve dynamically based on the model’s learning dynamics and establish a correspondence between attack parameters and training examples, optimizing adversarial perturbations in a learnable manner. Unlike conventional adversarial training, ASTrA ensures robustness while maintaining SSL’s efficiency and scalability.

While robustness, in this thesis, focuses on real-world challenges in natural scenes, domain-awareness focuses on specialized visual domains beyond natural scenes. Standard SSL augmentations are designed for variations in natural scenes, making them ill-suited for specialized fields such as medical imaging and industrial mining material inspection. This thesis introduces domain-awareness in SSL that incorporates domain-specific information into SSL’s view generation process. Particularly, (i) magnification prior contrastive similarity (MPCS) makes learned representations invariant to magnifications for histopathology images by inducing varying magnifications in the view generation process, improving breast cancer recognition. (ii) depth contrast explicitly enforces modality alignment between material images and attained height of materials on conveyor belt, ensuring that the learned representations become aware of physical properties, thereby improving material classification.

Beyond robustness and domain-awareness, SSL’s ability to generalize with limited data is advantageous for its practicality. While the loss objective in SSL is generally domain-agnostic, its effectiveness relies on large-scale data. In this direction, this thesis explores functional knowledge transfer (FKT), where self-supervised and supervised learning objectives are jointly optimized, enabling SSL representations to adapt dynamically to supervised tasks. This approach enhances generalization in low-data regimes.

In conclusion, this thesis provides a foundation for robust and domain-aware self-supervised representation learning in a modular manner, highlighting its applicability to existing and future JEAM-based SSL approaches, which can inherit these advancements and adapt to emerging challenges.

Solar photovoltaics (PV) is one of the most promising renewable sources of energy. Crystalline and multi-crystalline silicon solar cells hold most of the market share (up to 95%) of the photovoltaic industry. However, they require high-purity silicon and high production costs. Thin film technologies, including a-Si, Cu(In, Ga)(S, Se)2 (CIGS), and CdTe, have been thoroughly researched due to their minimal material consumption and scalability, yet they have struggled to achieve significant commercial success. These technologies face challenges such as low technological flexibility (e.g., use of flexible substrates), use of critical or toxic raw materials, and long-term stability. Furthermore, semitransparent photovoltaic technologies (STPVs) can harness previously unused spaces like windows and facades to produce on-site electricity. Buildings represent 40% of overall energy use and are responsible for 36% of total greenhouse gas emissions. STPVs will play a crucial role in meeting the energy requirements of a ‘zero-emission building’. The shortcomings of current PV technologies and the potential of STPVs incentivize the search for alternative PV technologies utilizing absorber materials that can effectively address these issues sustainably and at a reduced cost.

Antimony sulfide (Sb2S3) is an emerging light absorber material with favorable properties, such as a high absorption coefficient, wide bandgap (1.7–1.8 eV), earth abundance, and nontoxic constituents. Its low melting point (~550 °C) allows for obtaining high-quality crystalline thin films at low temperatures. The Sb2S3 solar cells in this thesis use a conventional planar n-i-p heterojunction with a configuration: glass/bottom contact/electron transport layer (ETL)/Sb2S3/hole transport layer (HTL)/top contact. A commercial glass/fluorine-doped tin oxide (FTO) substrate acts as the bottom contact. ETL and HTL help in the efficient and directional collection of electrons and holes. The top contact is a high-work function metal such as Au thin film (>60 nm for opaque devices). For semitransparent solar cells, the top electrode can be ultrathin Au (<15 nm) or indium tin oxide (ITO).

N,N-di(4-methoxyphenyl)amino]-9,9′-spirobifluorene (spiro-OMeTAD), an organic material, has emerged as the predominant choice for HTLs; however, it is costly and requires additives that render it vulnerable to moisture and elevated temperatures. Therefore, finding suitable inorganic stable HTLs is of utmost importance. The first part of this thesis (papers I and II) utilizes copper thiocyanate (CuSCN), an inexpensive and easily processable inorganic material that is highly transparent, as the HTL to realize fully inorganic Sb2S3 solar cells. CuSCN was first evaluated in the solution-processed solar cells using Sb2S3 films deposited through chemical bath deposition (CBD) (paper I). The opaque and semitransparent planar solar cells using all-inorganic layers obtained power conversion efficiencies (PCEs) of 1.75% and 1.67%, respectively. Subsequently, a hydrothermal deposition technique was utilized to enhance the quality of the Sb2S3 thin films, resulting in all-inorganic hydrothermally deposited Sb2S3 solar cells (paper II). A direct comparison between solar cells using CuSCN as the HTL and those lacking an HTL underscored the importance of the HTL in these devices. The HTL-free solar cells achieve a modest PCE of 1.54%, which improves to 2.46% when CuSCN HTL is included. These findings were corroborated by a one-dimensional numerical simulation. A semitransparent device is fabricated with a PCE of 2.13% and an average visible transmittance (AVT) of 13.7%.

Additionally, cadmium sulfide (CdS), derived from established CdTe PV technology, has solely served as the ETL for high-efficiency Sb2S3 solar cells. Nevertheless, cadmium's toxicity raises concerns that hinder the broader acceptance of these solar cells. Additionally, it possesses a low bandgap of 2.4 eV (with its characteristic yellow color), resulting in absorption-related current loss and diminished device transparency. Therefore, in the next section of the thesis (paper III), CdS was substituted with a wide bandgap, nontoxic TiO2 as the ETL. The PCE of the cadmium-free device using TiO2 was 5.1%, which was comparable to that of the CdS-based device (5.2%). However, the hydrothermal deposition of Sb2S3 on TiO2 results in non-uniform, island-like growth, which is unsuitable for semitransparent applications that require pinhole-free thin films of less than 100 nm. This island-like growth, caused by dewetting issues, is mitigated by applying an ultrathin ZnS layer (1–3 nm) on TiO2 using the successive ionic layer adsorption and reaction (SILAR) deposition method. By utilizing the resulting excellent film morphology of Sb2S3 on TiO2-ZnS ETL, semitransparent solar cells were fabricated with an ultrathin Au (<10 nm) electrode, achieving a PCE of 3.3% and an AVT of 11.2%.

In the last part (paper IV), a highly scalable radio frequency (RF) magnetron sputtering deposition was developed to obtain high-quality, uniform, and impurity-free Sb2S3 films directly on TiO2. Unlike the island-type growth of Sb2S3 on TiO2 seen in solution-based depositions, sputtering with the binary target enables the formation of smooth and dense Sb2S3 films, even at thicknesses less than 100 nm. A thorough optimization of the post-deposition annealing parameters yielded a record PCE of 4.6%. Semitransparent solar cells with varying degrees of transparency were developed through precise thickness control via sputter deposition. The semitransparent solar cells utilizing ultrathin Au as the top contact achieved PCEs of 3.2% (AVT: 10%), 2.6% (AVT: 13.5%), and 2.0% (AVT: 16.5%) for Sb2S3 layers with thicknesses of 80 nm, 60 nm, and 40 nm, respectively. Next, ITO is employed instead of ultrathin Au as the top transparent electrode to enhance transparency (AVT: 14.9% for the 60 nm Sb2S3 layer). Finally, the highly transparent CuSCN replaces the low bandgap polymer poly(3-hexylthiophene-2,5-diyl) (P3HT) as the HTL, further increasing the AVT to 20.5% for the same Sb2S3 thickness of 60 nm.

This thesis highlights how the thin-film deposition conditions, the ETL/HTL interfaces, and the device structure significantly influence the AVT and PCE of Sb2S3 semitransparent solar cells. In conclusion, the results of this thesis will greatly contribute to future studies on high-performance semitransparent Sb2S3 solar cells.

The growth in cyber-physical systems (CPS), the industrial internet of things (IIoT) and integrations of machine learning (ML) models have enabled Industry 4.0 automation and intelligence in industrial System of Systems (SoS). However, scalable automation frameworks, dynamic interoperability for smooth communication among heterogeneous systems, and the integration of ML models particularly for online collaborative intelligence in distributed architectures, are open challenges that must be addressed. This thesis structures its research as a progressive investigation, with each identified challenge leading to the next. First, local cloud-based automation is explored using the Eclipse Arrowhead framework to propose digitalization frameworks for industrial use cases, such as predictive maintenance in wind energy systems and smart manufacturing. The primary objective is to bridge the digital divide in SoSs environments, ensuring seamless intercommunication between IoT-connected devices. Significant engineering effort is dedicated to developing dynamically scalable automation solutions that integrate heterogeneous CPS, providing real-time adaptability and efficiency. This leads to the second research challenge addressed in this thesis. To investigate the challenge of semantic interoperability among heterogeneous IIoT devices, this research explores ontology alignment techniques through Natural Language Processing (NLP) and deep learning models. Extension of an existing language model (BERT_Intereaction) is proposed for ontology graphs to facilitate seamless communication between heterogeneous IIoT devices. It is designed using a language encoder to develop knowledge of the text to understand the labels or entities and a structural encoder to understand the context or semantics behind the text. This proposed model consistently outperforms cross-lingual tasks over the state-of-theart techniques with an error reduction of 2.1% on benchmark datasets DBP15KZH−EN, DBP15KJA−EN, and DBP15KFR−EN.

The third challenge involves scaling collaborative intelligence across distributed IioT systems. To address this, a local cloud-based collaborative learning (CCL) model is designed for the service-oriented architecture (SOA) and a decentralized ML model to digitalize IIoTs while enabling ML-based optimizations for IIoT tasks across cloud and edge nodes. The CCL model integrates machine learning-as-a-service (MLaaS) into the distributed cloud architectures. CCL offers scalable, privacy-driven, self-contained local

clouds for every CPS in the system of systems model. The local clouds enable distributed ML deployment across the IIoT SoS, where devices collaborate to share their knowledge representations. The model uses unsupervised dictionary learning, allowing IIoT nodes to share compressed, optimized learning representations. Furthermore, this thesis also highlights a culminating issue for designing decentralized ML-enabled IIoT solutions that is the information overload and redundancy at the edge and cloud. To mitigate this challenge, the CCL+ model is proposed, integrating coherence-based dictionary refinement

with Bayesian optimization. The model is tested on the condition monitoring task using data from an automated farm of six wind turbines using the CCL model. Implementing redundancy-aware strategies in the CCL+ optimized bandwidth usage and reduced communication overhead, especially for the resource-constrained IIoT devices. In the simulation experiments, over one year, the propagated learned dictionary size at a single wind turbine exceeded 1 petabyte. In contrast, in comparison, using the CCL+ model, for the same duration, the learned dictionary remained at 18 MB, significantly enhancing communication and computational efficiency without losing essential information.

Various potential future research directions accompany the findings presented in this thesis. For instance, to strengthen the ablation study on the semantic interoperability among heterogeneous SoS challenge, a generalized IIoT ontology that is designed for any IoT device (beyond sensors), such as the smart applications reference ontology (SAREF) can be tested for ontology alignment. This work provides a step towards enabling translation between heterogeneous IoT sensor devices. The proposed BERT Intereaction model can be further extended to a translation module using the generalized ontology graphs. Then investigations can be conducted to test if the model can interpret the messages transmitted across ontologically different devices in two scenarios: a) where the ontology graphs of both devices are a subset of the generalized ontology graph, and b) they are overlapping graphs and may contain different nodes and relations but they are semantically the same. Furthermore, exploring the utility of CCL+ model for extended

large-scale SoS with multiple parallel tasks to test the collaborative learning concept across the heterogeneous cyber-physical system of systems (CPSoS).

Railway electrified systems are one of the most popular and essential forms of transportation globally, and the performance of those systems impacts society. The electric railway power systems (ERPS) comprehend the infrastructure and apparatus that aims to deliver power for the rolling stocks in different types of railway transportation. Due to the broad application of static power electronics, ERPS is characterized by several sources of waveform distortion. Waveform distortion is a critical power quality (PQ) issue and a challenge to managing electromagnetic compatibility (EMC) in railway systems. It englobes harmonics (disturbances synchronous with the fundamental power frequency up to 2 kHz), interharmonics (disturbances asynchronous with the fundamental power frequency up to 2 kHz), and supraharmonics (synchronous and asynchronous disturbances between 2 and 150 kHz).

The ERPS has several system complexities that should be taken into consideration when assessing waveform distortion related to the characteristics of the phenomena: extensive distribution system with intricate circuit arrangements and moving single-phase loads; multiple voltage levels and electromagnetic environments, including railway grid and subsystems, as well as public grid; waveform distortion has time-varying behavior dependent on operating states of rolling stock, traffic plan, grid balancing, and spatial position of the vehicles; a mix between traditional equipment or infrastructure and population of new power electronic conversion stages with a lack of guidelines and standardization; and variety of waveform distortion sources and signatures.

The objective of this research is to gain knowledge and a better understanding of waveform distortion, including not only harmonics but also interharmonics and supraharmonics in railways systems, to characterize emission sources, propagation, the impact of the operation on time-varying behaviors in several scales, interaction among systems and subsystems, and adverse effects. The focus of the work is alternating current (AC) electrified railways, with a deeper assessment of, but not limited to, the railway system solution of Sweden (15 kV 16 ⅔ Hz). The development and scope of this work provide a comprehensive literature review of waveform distortion assessment for electrical railway power systems and build up a framework for future contributions, characterization of waveform distortion for electrical railway power systems using measurements, conduct detailed measurements on waveform distortion in a traction converter station, modeling waveform distortion propagation for ERPS considering complexities of the system, application of unsupervised deep learning (DL) methods to find patterns in waveform distortion data and investigation of the impacts related with those issues. The research contributions from those defined scopes are summarized below.

·         Identification of the challenges of waveform distortion assessment in ERPS and categorizing the available literature to address some of those challenges.

·         Characterization and screening of the waveform distortion time-varying dependencies in different time scales.

·         Providing a methodology for assessing time-varying waveform distortion in railway systems, adapting traditional methodologies, advanced statistical analysis, and machine learning approaches.

·         Modeling waveform distortion interaction within the ERPS in Sweden, incorporating challenges such as moving loads, meshed grid analyses, and a wide range of disturbances propagation in ERPS.

·         Addressing the different mechanisms affecting waveform distortion at the catenary and public grid sides.

·         Investigation of the impact of waveform distortion performance on associated equipment.

The work provides crucial steps for better establishing a PQ framework and future standardization for waveform distortion in ERPS by exploring multiple aspects and directions on the assessment side.

Gradually more stringent environmental and safety regulations in the transport sector have made third generation Advanced High Strength Steel (3rd-gen AHSS) grades and new generations of press hardening steels (PHS) cost-effective and natural substitutes in the automotive industry. Increasing the strength of steel allows for potentially downgauging the sheet thickness while maintaining or improving structural performance, and thus reducing the weight of the vehicle. 3rd-gen AHSS and PHS grades have been continuously adapted by the automotive industry for body-in-white parts and energy-absorbing safety components. However, the limited ductility of these higher-strength materials can make them more prone to cracking, which in turn has a negative impact on the folding behaviour of safety structures in a crash. For further introduction of new high-strength steel grades in the design and production of safety parts, proper calibrated material models are needed, and their crash behaviour must be investigated and quantified. Plane stress fracture toughness measured with the Essential Work of Fracture (EWF) method has recently emerged as a viable material parameter to rationalise edge crack resistance and crashworthiness. EWF offers a small-scale laboratory methodology capable of characterising important fracture characteristics of modern automotive steel grades. Hence, EWF together with well-instrumented crash tests in the laboratory are powerful tools for estimating the crashworthiness and quantifying energy absorption. However, much of the published fracture toughness data is based on quasi-static conditions, which do not reflect the conditions in a crash typically involving high deformation rates. To characterise the material for crash scenarios and validate simulation models, further investigation is necessary at higher deformation rates. In this PhD thesis, the crashworthiness and fracture characteristics of 3rd-gen AHSS and PHS grades at higher deformation rates were investigated. The crashworthiness and energy absorbing capacity were evaluated by studying dynamically loaded axially crushed crash boxes both experimentally using full-field deformation measurements and numerically by finite element analysis using a commercially available damage model. Stereo high-speed imaging allowed for more efficient evaluation of crash performance with fewer components and aided in model validation. Furthermore, the rate dependence of fracture toughness and the underlying mechanisms were explored, revealing that crack propagation resistance after crack initiation significantly influences fracture toughness at higher loading rates. It was also experimentally shown that there is significant adiabatic heating in the fracture process zone using the EWF methodology at higher loading rates, which can influence the value of fracture toughness.

Hälso- och sjukvården behöver bli mer hållbar, proaktiv och personcentrerad. För att uppnå detta behöver digitalisering nyttjas, nya innovationer utvecklas och personers egna resurser tillvaratas. Det innebär även omställning på flera organisatoriska nivåer och utveckling inom varje profession för att möta olika klientgruppers behov. Forskning och riktlinjer visar på ett behov av att utveckla stöd för self-management i aktiviteter i vardagslivet för personer med stroke. Det övergripande syftet med denna avhandling var att utveckla kunskap om införande av internetbaserad intervention och mer specifikt att utvärdera en ny arbetsterapiintervention ”Strategier för att nå ett aktivt vardagsliv” [SEE] för personer med stroke. 

Avhandlingen består av fyra delstudier. Studie I utforskade hur personer med stroke och deras arbetsterapeuter (AT) upplevde interventionsprocessen i SEE och huruvida SEE hade potential att främja ett aktivt vardagsliv. Studien var en explorativ deskriptiv fallstudie där fyra personer med stroke och två AT deltog i en första prövning av prototypen under utvecklingsfasen av SEE. Data samlades in genom en kombination av metoder från interventionsstart samt löpande till interventionens avslut. Data analyserades med mönstermatchning. Resultatet indikerade att SEE var genomförbar och accepterbar avseende innehåll och leverans. SEE kunde stödja utveckling av självreflektion och strategianvändning, vilket inverkade positivt på engagemang i aktiviteter i vardagslivet och livstillfredsställelse för personer med stroke. Arbetsterapeuterna upplevde en omställning i sin yrkesroll när de började leverera arbetsterapi digitalt och anpassade sig till att stödja klienternas förändring på ett nytt sätt. Utbildningsprogrammet och interventionsguiden för SEE var ett stöd i omställningen. Studien visade på behovet av mindre justeringar av prototypen och gav stöd för fortsatt forskning om genomförbarheten av SEE. 

Studie II utvärderade genomförbarheten av SEE för personer med stroke. Studien var utformad som en genomförbarhetsstudie med en pre-test-post-test-design utan kontrollgrupp inbäddad i en mixad metod ansats. Data samlades in genom en kombination av metoder vid interventionsstart, efter fyra och 12 månader samt löpande under interventionstiden. Data samlades från 15 personer med stroke, fyra AT och tre chefer. Data analyserades med deskriptiv kvantitativ analys och kvalitativ innehållsanalys. Resultatet visade att SEE var genomförbar för personer med stroke och kan stärka deras self-management samt förbättra engagemang, balans och värde i aktiviteter i vardagslivet. Studien visade också att SEE var genomförbar när det gällde dosering och accepterbarhet samt värdet av SEE skattades högt. Brister i följsamhet framkom relaterat till saknade formulär, tidsavvikelser i modulleveranser och utebliven uppföljning. Arbetsterapeuternas erfarenheter visade att det var utmanande att lära sig SEE, men samtidigt att SEE möjliggjorde ett nytt sätt för dem att ge stöd samt främjade klienternas engagemang och ansvarstagande. Chefernas erfarenheter visade att SEE har potential att förbättra och utveckla utbudet av rehabiliteringstjänster. Justeringar behövs i studiedesignen, avseende rekryteringsstrategier, bedömningsinstrument och utvärderingsmetoder av följsamhet. Utbildningsprogrammet för AT bör utvecklas inför framtida forskning för att bättre stödja implementeringen av interventionen.

Studie III utforskade hur AT upplevde användningen av SEE, med fokus på att stödja klientens resurser för att hantera ett aktivt vardagsliv efter stroke. Denna studie hade en fokusgruppsdesign med periodiska upprepade diskussioner. Fyra AT deltog vid fyra tillfällen och data analyserades med fokusgruppsanalys. Resultatet visade att AT upplevde SEE som ett värdefullt komplement till befintlig rehabilitering. Implementeringen av SEE innebar en mångfacetterad transition som involverade kontext, interventionsprocess och leverans, vilket tillsammans förnyade arbetsterapin. SEE erbjöd AT ett nytt sätt att tillhandahålla personcentrerad och aktivitetsfokuserad arbetsterapi. 

Studie IV utforskade hur personer med stroke upplevde interventionsprocessen i SEE och huruvida deltagandet i SEE hade betydelse för deras upplevelse av vardagslivet. Studien designades som en kvalitativ semistrukturerad intervjustudie med nio personer med stroke. Intervjuerna genomfördes efter en månad respektive fyra till fem månader från interventionsstart för varje deltagare. Data analyserades med konstant komparativ analys. Resultaten visade att relevansen och beredskapen för SEE varierade bland deltagarna. Sju av nio deltagare upplevde att interventionsprocessen gav dem kraft att driva sin egen förändringsprocess i aktiviteter i vardagslivet för att främja hälsa. Två deltagare såg inget behov av att förändra sina aktiviteter och påbörjade därför ingen förändringsprocess. När SEE var relevant och genomfördes som förväntat, beskrev deltagarna det som en ögonöppnare för hur de själva kunde agera för att förbättra sin hälsa genom att omfördela sin tid och sitt engagemang i olika aktiviteter.

Den ökade kunskapen som denna avhandling bidrar med kan användas för att förstå och främja införande av internetbaserad intervention för self-management. Kunskapen kan även användas för att utveckla rehabilitering som tillvaratar internets fördelar för att stödja förändringsprocesser i personers ofta komplexa vardagsliv. Slutligen visar kunskapen hur SEE och dess forskningsdesign behöver utvecklas för fortsatt implementering och utvärdering. 

Det övergripande syftet med avhandlingen är att, inom ett praktiknära forskningsprojekt, utforska hur undervisning blir till i förskolans praktiker med utgångspunkt i framväxande lokala och situerade utvecklings- och forskningspraktiker. Genom att under närmare tre års tid arbeta i nära samverkan med förskollärare och barnskötare från tre olika förskolor, inom ett och samma förskoleområde, har frågan om undervisning i förskolan kunnat undersökas på en mängd olika sätt, med utgångspunkt i praktikernas frågor. Med inspiration från kritisk aktionsforskning har utvecklings- och forskningsprocesser kombinerats, med sikte på att producera kunskaper i och tillsammans med de förskolepraktiker som deltar. Teoretiska och metodologiska utgångspunkter hämtas även från posthumanistisk teoribildning, främst via Barad, Haraway och Stengers, samt tidigare genomförda studier från förskolefältet med liknande utgångspunkter.

Tillsammans med en mångfald av aktörer så som förskollärare, barnskötare, barn, andra forskare, dokumentationer, förskolemiljöer, litteratur, lera och tidigare forskning har undervisningens tillblivelse i en rad olika praktiker utforskats. Undervisning framträder i avhandlingen som ett halt och svårfångat begrepp för de förskollärare och barnskötare som deltar i utforskandet, eftersom det utmanar föreställningar och traditioner om vad förskolans utbildning ska vara och hur den ska göras.  Genom att låta det som först framstod som ett motstånd mot undervisning i förskolan bli en aktiv deltagare i utforskandet av undervisningen skapades möjligheter för flera berättelser att existera samtidigt. Rörelser, i olika riktningar och hastigheter, men även avbrott deltar i undervisningens tillblivelser. Undervisning blir då en praktik som sträcker sig bortom en specifik händelse och omfattar både det förberedande planeringsarbetet, de efterföljande reflektionerna och den fortsatta processen. Det elaborerande med metoder som gjorts inom ramen för avhandlingen bidrar med förslag på hur berättandepraktiker kan användas för att uppmärksamma framför allt det som benämns som spontan undervisning.

I avhandlingen framträder en mängd berättelser om/med/för/av undervisning i förskolans praktiker som tillsammans visar på vikten av att behålla komplexiteten i frågor som rör såväl syfte med som innehåll i förskolans utbildning och undervisning, för att inte begränsa barns möjligheter att lära sig, utvecklas och bli till.

The contamination of aquatic environments by uranium (U) is a global concern due to its chemical toxicity and radioactivity, both of which pose significant health risks. Although U occurs naturally in the earth’s crust and is commonly present at low concentrations in natural waters, elevated levels in the environment can result from both natural and human-induced sources. While previous research has focused primarily on U release from U mining and processing sites, less attention has been given to U behavior at non-U mining sites such as iron ore mines, where U may still be present as a trace element. Under certain conditions, U can become mobilized posing a contamination risk to recipient rivers and lakes downstream. This study investigated U sources and mobility at the LKAB iron ore mine in Svappavaara, Northern Sweden, where U concentrations exceeding the Swedish annual average guideline value of 0.17 µg/L have been detected in a river receiving excess process water from the mine site. 

The results show that the iron ore fed into the processing plant is not a significant contributor to U in the process water circulating in the mine’s water management system. This is due to the iron ore’s low U content and the occurrence of U in thorite, a mineral that remains stable under the alkaline pH conditions prevailing in the processing plant. However, the ore also contains gypsum and anhydrite which dissolve after being liberated during grinding releasing calcium (Ca) and sulphate (SO₄^2-). The presence of Ca promotes the formation of ternary calcium-uranyl-carbonate complexes, which increase U solubility and reduce its tendency to adsorb onto mineral surfaces. As a result, U mobility is enhanced within the water management system and toward the recipient river downstream. 

Mine water pumped from Leveäniemi open pit was identified as the primary source of U at the mine site. This U originates from both groundwater infiltration into the open pit and weathering of U-bearing minerals in the open pit walls. Minewall weathering stations revealed that among the different rock types present in the open pit, pegmatite had the highest U leaching rates per unit area. However, trachyandesite, the dominant rock type in both the hanging and footwall, exhibited lower U leaching rates but likely contributes more U overall due to its abundance. Groundwater entering the open pit through fractures and the drainage pipes was also identified as a key U source influencing U concentration in the mine water. However, tracing U back to specific rock sources is challenging due to evolving groundwater flow paths as mining progresses and limited access to rocks along these flow paths. As a result, elemental and isotopic tracers are needed to better understand U release from bedrock into circulating groundwater. To support this, further investigation into U release from pegmatite and trachyandesite rock under varying geochemical conditions was carried out. 

Acidic environments significantly enhanced U release, particularly from uraninite compared to more refractory minerals such as fergusonite, and thorite. Under neutral pH conditions, mineral dissolution was limited, and U was primarily mobilized through surface desorption and complexation with dissolved ligands. Carbonate was the most effective ligand in promoting U mobility under these conditions. Elemental correlations suggested that trace elements such as Pb, Th, Nb, and Y could serve as indicators of pegmatite-derived U, as they are commonly associated with the primary U-bearing minerals in pegmatite. In contrast, U leached from trachyandesite showed similar leaching behavior to elements typically found in silicate minerals such as Fe, Al, Si, Mg, and Mn. These elements are not effective tracers for trachyandesite-derived U because they are ubiquitous in silicate minerals, some of which may dissolve under the same geochemical conditions, making it difficult to attribute their presence to U-bearing phases. U isotope analyses added a valuable dimension to the study revealing differences in source behavior not captured by elemental concentrations alone. While δ²³⁸U overlapped between the rock types, ²³⁴U/²³⁸U activity ratios varied distinctly. The lack of overlap in ²³⁴U/²³⁸U activity ratios between pegmatite and trachyandesite leachates under neutral pH conditions both in NaHCO3 and groundwater solutions suggests that these ratios are more suitable than δ²³⁸U for distinguishing between U rock sources in groundwater at the mine site.

These findings highlight the importance of understanding both the sources and geochemical controls on U mobility in non-U mining environments. This knowledge is essential for implementing effective source-specific remediation strategies and improving water management practices to minimize U transport to downstream ecosystems. 

Due to the successful commercialization of lithium-ion batteries (LIBs), there is a growing interest in developing new battery materials with improved properties. The uneven distribution of natural resources, the low abundance of battery materials in the Earth’s crust, and the growing geopolitical concerns should also be considered and addressed. In this context, alternative battery technologies, such as sodium-ion batteries (SIBs) and lithium metal batteries (LMBs), are getting attention by researchers, due to the low cost of readily available sodium resources and the very high capacity of a lithium metal anode, etc. Conventional electrolytes of any battery technology are today heavily based on fluorinated salts and volatile organic solvents, posing serious safety issues all the way from synthesis to application and recycling. Additionally, the increasing concerns of per- and polyfluoroalkyl substances (PFAS) highlight the urgent demand to explore performant fluorine-free electrolytes, ideally also non-flammable.

In this study, novel fluorine-free ionic materials and electrolytes have been designed and their physical and electrochemical properties thoroughly investigated. In the first part (Paper I), fluorine-free “solvent-in-salt” (SIS) sodium electrolytes based on sodium bis(2-(2-ethoxyethoxy)ethyl) phosphate (NaDEEP) salt and tris(2-(2-ethoxyethoxy)ethyl) phosphate (TEOP) solvent are presented. The addition of TEOP increased the electrochemical oxidation stability of the SIS electrolytes and an unusual ionic conductivity behavior is observed – the ionic conductivities of the electrolytes increase with increasing salt concentration. In the second paper (Paper II), a series of new orthoborate-based ionic materials, containing the bis(glycolato)borate (BGB) anion and phosphonium/ammonium cations are prepared and compared with the popular bis(oxalato)borate (BOB) salts. Some of these ionic materials are room temperature ionic liquids (RTILs), while others are organic ionic plastic crystals (OIPCs). The tetrabutylphosphonium bis(glycolato)borate ([P₄₄₄₄][BGB]) OIPC displays much higher decomposition temperature than the structural analogous [P₄₄₄₄][BOB] IL, and multinuclear solid-state NMR spectroscopy indicated weaker cation-anion interactions in phosphonium-based salts than the ammonium-based ones.

Given the excellent moisture and thermal stabilities brought by the BGB anion, a family of BGB-based alkali and alkaline metal salts were synthesized and characterized (Paper III). The LiBGB-based electrolytes using dimethyl sulfoxide (DMSO), triethyl phosphate (TEP) and trimethyl phosphate (TMP) have excellent moisture stability, optimal ionic conductivity, better aluminum (Al) passivation and long-term Li plating-stripping performance. Sequentially, the next study (Paper IV) is focused on investigating the effect of additives on the performance of these electrolytes, such as vinylene carbonate (VC), fluoroethylene carbonate (FEC), etc. Finally, in the fifth paper (Paper V), two- and three-component eutectic electrolytes based on pyrrolidinium saccharinate [Pyrr][Sac], lithium saccharinate Li[Sac] and/or [P₄₄₄₄][BGB] salts were created. The physicochemical properties of these salts as well as the Li compatibility and cell performance are thoroughly investigated. Overall, these studies identified several new fluorine-free salts and electrolytes with beneficial properties that can potentially be used in next-generation batteries.

Wood has been a crucial material for construction throughout history. However, due to poor natural durability of wood, it is difficult to use outdoors without any additional treatment. Conventionally, wood has been fully or partially impregnated with preservatives. However, some substances are harmful to mankind and environment, hence, regulated strictly. Therefore, methods for achieving sustainable protection of wood have been required and one method that has been investigated for achieving this has been through chemical modification. 

This doctoral thesis aims to develop a new modification system for solid wood in use class 3. The objective was to develop a wood modification system based on maleic anhydride (MA) and sodium hypophosphite (SHP) that enables exterior use without leaching by weathering. To meet this requirement, the modification should involve formation of stable cross-linking, altering the interaction between moisture and wood, consequently enhancing dimensional stability and biological resistance. 

To test the possibility of using MA and SHP, Scots pine sapwood (Pinus sylvestris L.) was treated with various ratio of chemical reagents, curing temperatures and durations. The treated wood was subjected to repetitive wet-dry cycle to assess its dimensional stability and leachability of chemical reagents. The result indicated formation of a stable cross-linking between wood constituents. 

To further investigate the formation of cross-link, solid-state 13C cross-polarization magic-angle-spinning (CP-MAS) nuclear magnetic resonance (NMR), 31P MAS NMR and X-ray photoelectron spectroscopy (XPS) were employed. The findings indicated that the cross-linking was likely to involve phosphonate (C-P-O) bonds. These results provided a deeper fundamental understanding of the reaction mechanisms between wood, MA and SHP, providing further scope for improved treatment systems in the future.

The impact of the modification on wood-water interactions was analyzed using low-field nuclear magnetic resonance (LFNMR) to study water in the wood at a saturated state. Additionally, the hydrophilicity of cell walls was studied via infrared spectroscopy after deuteration using liquid D2O. The results indicated that the modification reduced the affinity of the wood cell wall to water without altering the number of accessible hydroxyl groups.  

Finally, the modified wood was evaluated for fungal decay resistance, mechanical strength test (bending), and thermal stability. The modification significantly reduced mass loss caused by wood-decaying fungi by limiting the moisture uptake in wood and altering the chemical structure of wood. On the other hand, the modification did not improve resistance to fungal growth on the wood surface, suggesting that nutrient accessibility on surface was not influenced by the modification. A bending test showed that while the modulus of elasticity (MOE) was not affected, modulus of rupture (MOR) decreased to half that of untreated wood. Thermal resistance was improved due to the presence of phosphonate, which can promote the formation of a protective char layer and radical moieties. 

This study demonstrated the potential of modifying wood with MA and SHP to enhance durability, dimensional stability, and fire resistance. The modification formed stable cross-link within the wood components, reducing water interaction and improving resistance to biological degradation. However, the reduction in MOR limits its suitability for load-bearing applications. Despite this, the results suggest that the modified wood could be a viable alternative for non-load bearing exterior applications.

Future research should focus on optimising the modification process by reducing temperature, duration, and solvent use while maintaining performance. Investigating catalysts for the reaction may help address these challenges. Additionally, long-term field testing under real environmental conditions is needed to evaluate the durability and stability of the modified wood. Environmental impact assessments and life cycle analysis will also be crucial for ensuring commercial feasibility and sustainability.

Electric power transmission systems transfer large amounts of power (typically hundreds of MW) over long distances (typically hundreds of kilometres) at high voltage (typically hundreds of kilovolts). The operational security of the transmission system has always been high top priority for the transmission system operator (TSO); because of this supply interruptions originating in the transmission grid are very rare. 

To maintain the high reliability of a transmission grid, TSOs deploy the N-1 operational security criterion. The predominant shortcoming of this criterion is that all contingency cases are treated as equal; no differential is made concerning probability and impact of individual contingency cases. 

Operational risk assessment complements the N-1 security assessment method, by incorporating the probability of contingency cases and their impact, the latter in terms of severity factors.  Important elements of operational risk assessment are lead time, instantaneous component outage model, contingency definition, contingency list and filtering, probability of contingency cases, and severity factor. The existing literature on operational risk assessment concentrates on contingency filtration and ranking. Only a limited amount of literature exists on definition of severity factors. 

The main purpose of this thesis is to identify and summarise different existing and required research trends on the fundamental elements of operational risk assessment.  The contributions of the thesis include:

•   Identifying the fundamental elements of operational risk assessment and highlighting potential barriers against the practical implementation of operational risk assessment into the transmission system. Currently, TSOs are not deploying operational risk assessment, among others due to the absence of proper guidance and because of the high reliability resulting from the (N-1) criterion. Potential barriers against implementation of operational risk assessment, that were identified in the work, include absence of acceptable operational risk criteria, lack of a common and standardized set of severity factors, lack of sufficient knowledge on interpretation of operational risk results, and improper guidance on when and which types of measures are required to reduce the operational risk. 
• Introducing multi-state component models, including hidden failures, to operational risk assessment. In the power grid, major blackouts occur due to contingency cases involving protection failures. Including protection and protection failures in operational risk assessment results in several practical and mathematical challenges. Practical challenges include obtaining transition rate data; mathematical challenges include computing the time-dependent state probability of a large Markov model. This thesis addresses these mathematical challenges and provides a way to resolve them.  
• Clarifying the role of severity factor in operational risk assessment and proposing different deterministic and stochastic severity factors. The definition of the severity factor has a big impact on the way in which the results from operational risk assessment should be interpreted. A common set of severity factors is important for the interpretation of operational risk results and for the exchange of information and experience.

An important finding from this work is that operational risk assessment provides additional dimensions to the operational security planning, next to deterministic security criteria. However, several research gaps remain that need to be filled before implementation of operational risk assessment to existing transmission systems is possible.