Context. Fast and slow solar wind have distinct properties linked to their solar sources. The Alfvénic slow wind complicates the usual classification by solar-wind speed, especially at intermediate speeds. The Solar Wind Analyzer (SWA) on Solar Orbiter offers unique capabilities to investigate how Alfvénic slow wind differs from classical fast wind, and relates these differences to their solar origins.

Aims. In September 2022, Solar Orbiter moved inbound toward the Sun from 0.59 to 0.38 AU, and observed several Alfvénic streams: one fast wind (F), three Alfvénic slow wind intervals (AS1, AS2, AS3), and a moderate fast wind (MF); a non-Alfvénic slow wind (S) interval was included for comparison. We analyzed these streams to highlight their similarities and differences.

Methods. We combined plasma parameters from all SWA sensors with magnetic field measurements from the magnetometer (MAG). A spectral analysis of magnetic and velocity fluctuations was used to characterize the Alfvénicity. The magnetic connectivity of each stream to its solar source was examined using potential field source surface extrapolation combined with ballistic backmapping from the spacecraft.

Results. Proton velocity distribution functions exhibit anisotropies and field-aligned beams typical of Alfvénic streams, while electron strahl populations become progressively broader and weaker from fast to non-Alfvénic slow wind. Heavy-ion composition shows a similarly ordered transition, where the Alfvénic slow wind displays intermediate charge-state ratios between fast and non-Alfvénic slow wind. Magnetic connectivity analysis associates the fast wind with a large coronal hole, whereas the Alfvénic slow wind intervals connect to open-field regions near pseudostreamers and overexpanded coronal-hole boundaries. Spectral analysis indicates that most Alfvénic streams remain close to energy equipartition despite their different heliocentric distances and bulk speeds. Altogether, these results support the interpretation of Alfvénic slow wind as the low-speed extension of coronal-hole-associated Alfvénic plasma undergoing a distinct acceleration and transport history.

Conclusions. The combined SWA observations offer key insights into how turbulence, particle transport, and coronal source conditions jointly shape the evolution of Alfvénic solar wind streams, revealing a coherent transition between fast, Alfvénic slow, and non-Alfvénic slow wind regimes.

Vehicle braking releases particles that could negatively affect both the environment and public health. With the introduction of Euro 7, emission limits became more stringent, increasing the need for reliable measurement methods. The aim of this project was to further develop KTH's existing pin-on-disc machine to better replicate real braking events and thereby improve the representativeness of the results.

This report investigated particle emissions generated during braking, their environmental and health impacts, and the requirements introduced by Euro 7. Existing methods for estimating emissions, such as pin-on-disc testing and brake dynamometer testing, were also presented, together with their respective advantages and disadvantages.

The study showed that KTH's pin-on-disc equipment could be improved by implementing dynamic loading, as the existing system used a static load applied through a dead weight. Various existing solutions were examined, and a final design was presented in the study. The system's maximum vertical force, actuator precision, and response time were calculated to be 528 N, 0.15 N, and 0.16 s, respectively, thereby meeting the specified system requirements. Structural analyses demonstrated that the design possessed sufficient durability. The implementation cost was kept below the budget limit of 10,000 SEK. The proposed design remained at the conceptual stage and was not implemented or experimentally validated on physical equipment.

The powertrain manufacturing industry requires efficient resource allocation to handle complex and stochastic environments. Traditional heuristic rules, such as First-In-First-Out (FIFO), often operate reactively and fail to account for spatial limitations in 3D physical layouts, leading to suboptimal throughput and high cognitive load on operators. The purpose of this thesis was to investigate the application of Deep Reinforcement Learning (DRL) to optimize tool change scheduling in a simulated manufacturing line subject to stochastic disturbances.

A stable integration framework was developed, connecting an industrial-grade Discrete Event Simulation (DES) environment (FlexSim) with a Python-based DRL agent utilizing the Proximal Policy Optimization (PPO) algorithm. Communication was handled via a synchronous TCP/IP socket connection operating under the Gymnasium standard (a widely used API for reinforcement learning), which proved to be perfectly stable over 20 million training steps. The agent’s performance was evaluated against the FIFO baseline over 100 independent deterministic simulation runs.

The empirical results show that the DRL agent outperformed the baseline, achieving a 6% increase in total throughput while requiring a 4% increase in tool changes. This performance gain results from a behavioral shift; the agent transitioned from reactive logic to proactive resource allocation, reducing operator idle time and increasing active utilization from 40% to 45.77%. The higher production rate was achieved without increasing operational variance, proving that the policy is highly robust against stochastic machine breakdowns (MTBF/MTTR).

The study concludes that DRL is a practical solution for dynamic manufacturing optimization. By relieving operators from stressful decision-making, this framework aligns with the human-centric principles of Industry 5.0 and provides a verified foundation for future implementations of fully operational digital twins.

Engineering systems are often analyzed and operated using models that are incomplete, uncertain, or, in some cases, unavailable. In control engineering,vthe analysis of important system properties commonly relies on approximate parametric models, typically obtained through a prior system identification step. This thesis develops data-driven methods for characterizing behaviorally significant perturbations in dynamical systems and machine learning models using experimental observations rather than complete parametric descriptions.

The first part of the thesis concerns robustness analysis under model uncertainty. Given an unknown dynamical system, the objective is to estimatecontrol-relevant quantities directly from input-output data. In particular, we develop methods based on sequential input design to characterize structured perturbations that may destabilize a system and to quantify the mismatch betweena nominal model and the true plant from a closed-loop perspective. This information makes it possible to assess whether stability and performance conclusions obtained from the nominal model remain valid for the actual system.

In the second part of the thesis, we study the complementary problem of identifying minimal perturbations capable of producing a desired change in the behavior of a black-box model. Inspired by explainable artificial intelligence,we develop an adaptive algorithm for identifying feasible and actionable modifications using a limited number of model evaluations. We further apply this method to controller fine-tuning, where the objective is to identify small changes in controller parameters that lead to satisfactory closed-loop behavior while respecting operational constraints.

Together, these contributions show how task-relevant information about system sensitivity can be extracted directly from data, without reconstructing acomplete model of the underlying system.

Torque measurement is essential in many industrial applications, particularly in high-speed tightening systems where torque varies rapidly over time. Inline Rotary Torque Transducers (IRTTs) are commonly used for torque measurement in such applications; however, conventional calibration methods are primarily performed under static or quasi-static conditions. This creates challenges when evaluating transducers intended for dynamic operating environments.

Within this project, the mechanical design and development of a voice-coil-driven dynamic calibration test rig for an Inline Rotary Torque Transducer (IRTT) was carried out. The work focuses on improving an existing prototype by addressing mechanical limitations in the excitation and transmission systems.

The study primarily focuses on reducing the moving mass of the excitation system, improving mechanical interface conditions, and minimizing motion loss caused by play in critical joints and connections. Several design concepts were developed through CAD-based design iterations and evaluated using preliminary simulations based on the Finite Element Method (FEM). The proposed improvements included mass reduction and geometric modification of the arm, development of new connector and pin configurations, application of appropriate fit tolerances in critical mechanical interfaces, reduction of the number of bearings in the excitation system, and selection of a more compact and torsionally stiff coupling for the torque transmission system.

The developed configuration achieved a substantial reduction in total moving mass while maintaining acceptable structural performance under the applied loading conditions. In addition, the improved interfaces and implementation of appropriate fit tolerances contributed to reduced mechanical play and motion loss within the excitation system. The proposed design improvements resulted in a more compact excitation system with improved assembly conditions, providing a foundation for further development of the dynamic calibration test rig.

The lime and cement industries account for roughly 8–-9% of global annual CO2 emissions, but the decarbonisation of the sector is challenging due to the liberation of CO2 during chemical reactions. Traditional kilns also rely on fossil fuel combustion, and available carbon capture technologies incur severe energy penalties. To enable effective carbon capture, SaltX Technology AB has developed the Electric Arc Calciner (EAC), utilising an electric plasma torch operating in a pure CO2 atmosphere to enable rapid particle conversion and serve as an inherent direct-capture mechanism. However, this high-agility operating window eliminates standard structural thermal buffers, introducing highly coupled mass-energy transients. The thesis addresses this by building and validating a dynamic, first-principles Simulink/Simscape model. Discretised into a finite network of modular volumes, the framework tracks multi-domain mass and energy balances. Steady-state cross-validation against pilot SCADA logs demonstrated high accuracy of temperature and flow rate predictions with global error scores below 5% and a high goodness of fit (R2 > 0.97). Pressure metrics exhibited greater deviations, MAPEs reaching 23.62%, but also with a high statistical fit (R2 > 0.95), mostly due to sensor noise and idealised abstractions. Conversely, transient validations revealed severe thermal inertia anomalies, overshooting start-up temperatures by 91.9% with a 14.4% faster rise time. This poor transient validation indicates that the homogeneous wall abstraction drastically underestimates physical heat capacity and entirely omits internal conductive resistance. OFAT mapping showed that power responses were highly asymmetric due to the S-shaped calcination curve; minor under-powering collapsed the calcination degree to 89.5%, risking severe product deficits. Scaling the technology reduces the pilot SEC from 10.73 MJ/kg to 5.5 MJ/kg by reducing heat losses relative to production capacity. When conventional kilns are penalised with a 3,000 MJ/t CO2 amine-scrubbing load to meet net-zero requirements, the scaled EAC becomes highly competitive. Evolving this framework into an industrial Digital Twin requires updating homogeneous wall blocks to multi-node conductive layers to eliminate transient errors and building a communication infrastructure between the digital and physical plant.

This thesis explores flexibility strategies for low to medium volume, high-variety battery pack assembly, focusing on module-to-pack production. As battery technologies, customer requirements and demand levels continue to change, assembly systems need to handle new product variants without creating excessive disruption, cost or quality risk. The aim of the study was to define flexibility in this industrial context and develop a qualitative assessment approach to support future assembly system design when quantitative production data is limited.

The study used a qualitative case-based methodology, including current-state analysis, stakeholder interviews, scenario development, qualitative SMED and multi-criteria decision analysis (MCDA). The methodology itself forms part of the thesis contribution, as it provides a structured way to assess flexibility concepts under uncertainty before detailed production data or full implementation is available.

The findings show that flexibility in battery pack assembly is multidimensional and depends on product flexibility, volume and mix flexibility, material flow, digital support, verification, traceability and operator capability. Two assembly concepts were developed and evaluated. Concept 1 focused on cellular and local reconfiguration logic, while Concept 2 focused on dual-line and scalable flow logic. Concept 2 scored slightly higher, mainly due to stakeholder weighting on capacity and scalability. However, the small difference suggests that a staged hybrid approach is the most practical industrial direction.

The use of low-carbon concrete is increasing as the construction industry aims to reduce its environmental impact. These concretes often replace part of the cement with materials such as ground granulated blast furnace slag. However, there is limited knowledge about how this affects the risk of plastic shrinkage cracking at an early age. The aim of this thesis was to study how different slag contents and the timing of surface covering influence the formation of plastic shrinkage cracks. An experimental laboratory study was carried out using three concrete mixtures and two curing conditions, with covering and without any covering. Crack formation was evaluated based on the number, width and time of occurrence. The study has been conducted in collaboration with Swerock. The results have increased the knowledge on the risk of plastic shrinkage cracking in low-carbon concrete. 

The purpose of this study is to examine how architectural design in nursing homes affects the well-beingand the everyday life of elderly residents in Swedish nursing homes. The study also aims to investigatehow architectural interventions can help to improve existing nursing homes and evaluate how designstrategies can help care environments respond to crises such as pandemics.The study is based on a literature review and interviews with architects and municipal developmentspecialists working in architecture and elderly care. In addition, a building analysis was conducted basedon the study’s findings to illustrate and compare different design principles.The results indicate that nursing homes in Sweden are moving away from institutional, corridor-basedlayouts toward more home-like environments with a greater focus on the needs of the individual. Severalarchitectural factors were identified as particularly important for residents’ well-being, such as daylight,material selection, lighting, and clear spatial organization. The study also identifies recurringshortcomings existing buildings, such as long corridors and poor accessibility, which can affect both theworking conditions of staff and the well-being of residents. Experiences from the COVID-19 pandemicalso show the need for layouts that enable zoning and reduce the spread of infection.The study concludes that nursing homes should be designed to meet the needs of both residents andstaff, since the physical environment affects both quality of care and daily working conditions.Architecture plays an important role in supporting both well-being and staff efficiency. Targetedmeasures such as improved lighting, clear spatial organization, color coding and centralized socialspaces can help improve the quality of existing nursing home environments.However, the results should be interpreted within the confines of the limitations of the study, as thestudy is based on a limited number of interviews and does not include direct interviews with residents.

This licentiate thesis examines how students in compulsory school express their understanding of technological systems in technology education, with a focus on the wastewater system. The thesis analyses how students in Grades 6 and 9 express their understanding of system structure, function and relationships, with primary emphasis on their oral descriptions.

The licentiate thesis is based on two empirical studies presented in separate articles. The first study investigates how students in Grade 6 describe and model a wastewater system through self-drawn system models and oral descriptions. The second study investigates how students in Grade 9 express their understanding of the same system, with particular attention to aspects that are less visible in everyday use, such as control and system boundaries. To enable a comparative analysis, the students’ descriptions from both studies were analysed using the same analytical framework.

The comparative analysis draws on a classification model (Hallström et al., 2022) in combination with the SOLO-taxonomy (Biggs & Collis, 1982), enabling an analysis of both which aspects of the system students express and how these aspects are integrated in their descriptions. The results show that students in Grade 9 generally use a greater number of subject-specific concepts than students in Grade 6 and express their descriptions with greater linguistic precision. At the same time, there is no clear relationship between increased concept use and more relational ways of expressing systems understanding, in which relationships between parts and the system as a whole are made explicit.

In both grades, students’ expressions range from fragmented and linear descriptions to more relational descriptions in which different aspects of the system are related to one another. Variation in systems understanding is greater within each grade than between the two grades. Some students in Grade 6 express descriptions corresponding to relational levels of understanding, while some students in Grade 9, despite using more developed language, express a limited understanding of the system as a whole.

The results show that expressions of systems understanding do not follow a linear pattern associated with grade level. Concept use and ways of expressing relationships between parts and the system as a whole emerge as partly distinct aspects in the material. The use of concepts does not always indicate an integrated systems understanding. The findings highlight the need for teaching that explicitly supports students in using concepts to reason about relationships, structure, function and system boundaries in technological systems.

This thesis explores the potential of adaptive reuse within modern urban development, using the northern ticket hall at Hornstull, Stockholm, as a case study. As the construction industry currently accounts for a significant portion of society's total climate impact, transitioning toward circular processes and increased reuse has become a critical priority. Current urban plans for the site prescribe the total demolition of the existing ticket hall to make way for a new, larger building intended for retail and office use, including an extension of Högalidsparken ontop of the new volumes.

The purpose of this study is to investigate whether these urban and functional goals can be achieved by preserving the building's existing concrete frame instead of demolition. Employing a qualitative methodology—including site visits, analysis of architectural drawings, and interviews with industry professionals from Tham & Videgård and Arcona—a conceptual design proposal was developed.

The results demonstrate that it is technically and architecturally feasible to transform the ticket hall by retaining the concrete structure and adding an extension made of cross-laminated timber (CLT). By activating the ground floor with public functions such as an art gallery, café, and restaurant, while adding office spaces and a rooftop bar, the proposal creates a more vibrant and secure environment through "eyes on the street". Furthermore, the design introduces an open passage and a new staircase to strengthen the physical and visual connection between Hornsbruksgatan and Högalidsparken. A simplified carbon calculation indicates that preserving the 556 m³ of concrete in the existing structure results in a climate saving of approximately 157 tonnes of CO2e, equivalent to roughly 448 round-trip flights per passenger between Stockholm and London.

This study examines how the design of service contract areas affects the procurement, planningand operation of tendered bus services in Region Stockholm. The study focuses particularly onthe spatial aspects of the contracting structure, including the size of contract areas, theirgeographical boundaries and their relationships to one another. The aim is to understand howbus service contract areas in Stockholm have changed over time and which factors influencetheir size and design.

The study is based on a qualitative research approach, combining a targeted literature reviewwith semi-structured interviews with representatives from both the regional public transportauthority and bus operators. In addition, data on contracts, operators and geographicalboundaries has been used to visualise changes in the contracting structure between 2019 and2026.

The results show that contract areas in Region Stockholm are not static, but change inconnection with new procurement processes, political decisions and strategic considerations.Between 2019 and 2026, the number of contract areas increased from 11 to 12 as one previouslylarger area was divided into two separate contracts. The study also shows that operator changesoccurred in several areas during this period. The design of contract areas is influenced byseveral factors, including depot location, municipal boundaries, traffic areas, route networks,previous contract structures, market conditions and procurement strategy.

A central conclusion is that contract boundaries mainly function as organisational boundariesrather than physical barriers to bus services. They rarely prevent bus routes from crossinggeographical areas, but they do affect responsibility, planning, compensation, depot use and theability to implement changes during the contract period. The study also shows that the designof contract areas involves a trade-off between economies of scale and competition. Larger orgeographically connected areas may enable more efficient service operations, while smallerareas may support competition and market entry. The study thereby contributes to theunderstanding of how the organisational structure of public transport affects the conditions foran efficient, accessible and sustainable transport system.

Sensor drift is a major obstacle to reliable long-term measurements of low-cost sensors. It is commonly treated as a calibration problem, although its definition, causes, and temporal behavior are often left implicit. Focusing on non-dispersive infrared (NDIR) CO₂ sensors, we develop a systematic approach to defining, understanding, characterizing, and addressing drift.

We first describe sensor measurements and measurement errors as stochastic processes. Within this framework, drift occurs when the statistical properties of the measurement error process vary with time. We then introduce an additive drift decomposition model (ADDM) to separate two principal sources of drift: reversible effects caused by environmental conditions and irreversible intrinsic changes in the sensing instrument. Experiments using long-term field data demonstrate the feasibility of estimating the individual components and support the additive formulation.

After environmental correction, most sensors exhibit mean-reverting instrumental errors rather than a common deterministic trend. We model these dynamics using an Ornstein–Uhlenbeck process, whose parameters characterize the sensor-specific equilibrium error, mean-reversion rate, and stochastic variation. These characteristics provide theoretical guidance for designing sensor-specific one-point calibration strategies.

Finally, we address calibration without a co-located reference instrument by extracting pseudo-reference information from three external sources: same-type sensors, distributed reference stations, and local sensor networks. First, nonlinear environmental dependencies estimated from tested sensors are transferred to other sensors of the same type for environmental compensation. Second, periodic minima are interpreted as order-statistic estimators of atmospheric background concentrations, and observations from distributed reference stations are fused using a spatio-temporal generalized additive model to construct dynamic baselines for one-point calibration. Third, normalized pairwise differences are used to identify rendezvous events within a local sensor network; adaptive piecewise constant approximation and a two-state explicit-duration hidden semi-Markov model are proposed to detect rendezvous suitable for the downstream network calibration. These methods are evaluated using long-term field data and dedicated indoor deployment experiments.

Transition to Circular Manufacturing Systems (CMS) is crucial for sustainable growth,resource conservation, and mitigation of environmental impacts. This shift requiresa systemic approach that integrates value propositions, product design, and supplychains, supported by digital technologies. Circular supply chains (CSCs) emerge as acore pillar of CMS, demanding deliberate design to close the loop beyond simply addingend-of-life recovery to linear models. CSCs are inherently complex; however, currentresearch often fails to capture these aspects, overlooking the systemic approach to interdependenceamong the value proposition, product design, and supply chain. Furthermore,most studies are theoretical, lacking the necessary quantitative and empiricalanalysis to evaluate CSC’s performance. This research aims to conceptualize CSCswithin the CMS context, identifying their complexity aspects, and develop an analysistool to capture these complexities and evaluate CSCs’ technical, economic, and environmentalperformance to support informed decision-making. This thesis proposesan analysis framework with two modules. The first, Conceptualization, provides anoverview of CSCs implementation challenges, stakeholders’ roles and relationships,and how value proposition, product design, and transparency influence CSCs’ structure,uncertainties, and stakeholders’ footprints. Building on this, the second module,Analysis and Implementation, quantifies CSCs’ performance through a multimethodsimulation model, combining agent-based, discrete-event, and system-dynamics. Theframework’s applicability is illustrated through two distinct case studies.

Temperature development in tyre-road contact is an important factor for understanding wear and particle emissions, but it is difficult to measure experimentally in short-duration sliding contacts. In this work, a method for measuring the surface temperature of a rubber specimen after contact in a Wessex pendulum test rig under dry conditions was developed and evaluated. An infrared temperature sensor was integrated into the rig using an adjustable 3D-printed mounting solution designed to enable repeatable sensor positioning without affecting the function of the rig. Temperature measurements were performed during repeated pendulum strikes using different sampling frequencies and sensor positions. The results show that the solution enables consistent temperature measurements after sliding contact and that the method can be used to complement pendulum-based friction testing with temperature data. Since the measurement is performed after contact, the results should be interpreted as surface temperature after sliding contact rather than as the maximum temperature in the contact zone. The study therefore demonstrates that an existing pendulum test rig can be adapted for investigating contact-related temperature development, which may support further studies of particle formation in tyre-road contact.

  This report addresses the development of an improved enclosure for a laboratory-based measurement system, with a focus on functionality and manufacturability. The work began with a review of the state of the art, where existing solutions were analyzed to identify the strengths and limitations of the current enclosure at KTH. The work was based on a newly developed, predefined shape that had been developed in accordance with the technical regulation of GTR24 from the United Nations.

This was followed by a needs analysis based on interviews with laboratory users, which formed the basis for the development of a requirements specification. The specification defined, among other aspects, functionality, safety, and mechanical requirements that the enclosure should meet. A function tree was established to structure the main and sub-functions of the system, serving as a basis for concept generation. Based on this, three concepts were developed and subsequently evaluated through a systematic concept evaluation.

After selecting the final concept and materials, a virtual concept was detailed in a CAD environment. Then, a structural analysis was performed in Ansys to evaluate whether the enclosure could withstand the loads that may occur in a laboratory environment. Finally, a technical drawing and a bill of materials were produced.

The result of the thesis was a manufacturable, user-adapted conceptual enclosure design that is more manageable for one-person handling compared to the current enclosure. The design also provides a basis for further development and future testing in the laboratory. The final design has approximate outer dimensions of 120 cm in length, 45 cm in width, and 40 cm in height, with a total weight of 11.3 kg.

This thesis was conducted as a Minor Field Study (MFS) funded by the Swedish International Development Cooperation Agency (Sida). The study examines teaching practices in a Vietnamese university mathematics classrooms and explores how digitally designed lectures can support more interactive classroom practices. Using a design-based research (DBR) approach, the study combines Think-Pair-Share (TPS), digital tools, and sociocultural perspectives on learning to develop a lecturers’ guide for interactive mathematics teaching. Data were collected through students’and lecturers’ surveys, semi-structured interviews, and iterative feedback cycles. The findings show that teaching remains largely lecture-based, while opportunities for interaction are more limited. Across the DBR cycles, the instructional design was refined toward clearer TPS structures, stronger scaffolding, and more explicit use of digital visualization and multiple representations. The study suggests that interactive mathematics teaching can be developed within existing lecture structures through carefully designed forms of participation and pedagogically integrated digital tools. However, since the design was not fully implemented in authentic classroom practice, the study cannot determine its effects on student learning outcomes.

Within control theory, a question of great interest is to understand when a stabilizing feedback exists. Building upon Krasnosel’ski and co-workers, necessary conditions for the existence of a continuous stabilizing feedback have been pioneered by Brockett and Coron, providing so-called adversarial and homological conditions, respectively. Kvalheim recently generalized their conditions from equilibria to compact attractors and—in line with earlier work by Coron—showed that when a control system can be identified with a vector bundle, the homology condition is stronger, that is, the homology condition implies the adversary condition. Kvalheim also showed that this comparison breaks when extending the case of vector bundles to fiber bundles. In this work, we propose and characterize an appropriate adversarial condition for control systems represented by fiber bundles. The homology condition, on the other hand, is already well-defined, showcasing its general applicability. With the generalized conditions at hand, we show that by appropriately relaxing the regularity of the feedback under consideration, that is, by allowing for mild discontinuities, the homology condition implies the adversarial condition for fiber bundles. Thereby, in this work, we extend these earlier comparative results.

In light of the new EU regulation, Euro 7, a stronger focus on brake wear and brake energy has emerged. Compared to commercial vehicles, Heavy-Duty vehicles have the advantage to possess different auxiliary brakes which help to decelerate the vehicle without emitting particles. This study proposes an initial approach to show the effect of auxiliary brakes on particulate matter PM10. This was done by developing a theoretical model using the simulation tool MATLAB based on the newly developed heavy-duty brake cycle, by calculating braking energy and temperature and connecting them to wear and PM10 emissions. The brake cycle was processed twice to achieve two different routes by changing the order of sections representing the driving environments: urban, rural, and motorway, while maintaining the same total brake energy. The results suggested that the level of PM10 emissions may vary depending on route, even though it is the same distance, driving environments and total brake energy. For a Heavy-Duty vehicle assisted with a hydraulic retarder, the results showed that the average temperatures can be reduced by 40%, and max temperatures can be reduced by 28%, and that PM10 emissions overall can be reduced by 53%.

Partiklar från vägburen trafik har länge utgjort en betydande risk för människors hälsa. Majoriteten av trafikrelaterade partiklar kommer från avgassystem men striktare reglering av avgasrelaterade partikelutsläpp i kombination med ökad elektrifiering av fordonsflottan har medfört att icke-avgasrelaterade utsläpp, så som från däck och bromsar, utgör en allt större andel av de totala partikelutsläppen. År 2026 införs EURO 7 inom Europeiska unionen, vilket blir det första regelverket som omfattar icke-avgasrelaterade partikelutsläpp. Detta skapar ett behov av tillförlitliga metoder för att mäta dessa utsläpp under verkliga förhållanden.

Genom en litteraturstudie och en iterativ designprocess har detta arbete undersökt hur partikelutsläpp från bromsar och däck kan mätas på en elektrisk lastbil under körning. Arbetet har fokuserat på att utvärdera tillgängliga mätinstrument samt att utveckla koncept för en lämplig mätanordning. Vidare har potentiella störkällor som kan påverka mätresultaten identifierats och hanterats.

De utvecklade koncepten inkluderade inkapslingar för både bromsar och däck med syftet att samla in partiklar. Arbetet identifierade kontamination från omgivande partiklar, värmeeffekter samt störningar av det naturliga luftflödet som stora utmaningar för ombordmätning. Resultatet presenterades i form av konceptuella CAD-modeller, avsedda att fungera som en grund för utveckling och vidare arbete.

Free return trajectories have been used ever since the Apollo era. This type of trajectory aims to achieve a close approach to the Moon and utilize the Moon's gravitational pull to safely return to Earth without additional impulses during flight. This requires the use of three-body dynamics. This thesis investigates the generation and navigation of such trajectories using reinforcement learning (RL), particularly Proximal Policy Optimization with Long Short-Term Memory (PPO-LSTM). Reinforcement learning stems from Markov Decision Processes (MDPs) and utilizes exploration and exploitation to solve problems governed by a defined reward structure, where the agent seeks to maximize the expected reward for a given set of actions. The actions available to the agent consist of a full two-dimensional impulse and a variable drift time. This work investigates two particular use cases of RL implementations for free return trajectories. First, PPO-TLI focuses on the generation of the Trans-Lunar Injection (TLI) starting from Low Earth Orbit (LEO), followed by the use of multiple short-burst impulses to generate a ballistic free return trajectory. Second, PPO-MCC focuses on Mid-Course Corrections (MCC), where the agent must correct a near free return trajectory approximately 30 minutes after TLI. The results show that an RL agent can both find and correct these types of trajectories with varying stability. The main challenge arises from the nonlinearity and sensitivity of the Earth–Moon system, where small early deviations can lead to large changes later in the mission, which the agent struggled to adjust for. This work identified a favorable reward structure and curriculum for the agent, where hyperparameters and reward coefficients are gradually adjusted to help the agent learn to effectively explore and exploit the environment in order to generate and correct free return trajectories.

Modern vehicles have generally increased in size over time, while parking dimensions have only changed marginally. This has led to recurring discussions about narrow parking areas, limited space for door opening and parking damage.

The purpose of the study was to investigate how existing parking garages relate to today's changed vehicle dimensions and how the parking design can be developed to improve usability and function. The study was based on literature studies, vehicle data, site studies, interviews, user tests and modelling of larger parking dimensions in existing parking garages.

The results indicate that modern vehicles occupy a larger part of the parking space than previous vehicle generations, which can reduce the available space for entry and exit. Observations, interviews and user tests also indicate that users in certain situations adapt their parking behaviour to create more doorway space.

The modelling experiments indicate that the possibilities of introducing larger parking spaces can improve usability by creating more space for entry and exit. At the same time, the results indicate that the possibilities of introducing larger parking spaces in existing parking garages are often limited by garage design, column placements and available manoeuvring space. In the garages studied, larger parking spaces also resulted in a reduced parking capacity.

The study contributes with a basis for further discussion on how parking garages can be designed with regard to vehicle development and user needs. However, the results should be interpreted with some caution as the study is based on a limited number of case studies, interviews and user tests and therefore mainly provides exploratory indications rather than generally valid conclusions.

Artificial Intelligence (AI) is increasingly entering procurement environments through applications supporting supplier discovery, evaluation, monitoring, and decision-making. However, despite growing technological development and organizational interest, AI adoption within procurement remains uneven, particularly in complex B2B environments where supplier relationships, risk management, and human judgment remain central. This study explores the drivers and barriers influencing the adoption of AI-enabled procurement solutions and examines how organizations perceive, evaluate, and engage with AI-supported decision-making processes. The research adopts an exploratory and interpretive qualitative single-case study approach centered on the procurement ecosystem surrounding Find My Factory (FMF), an AI-enabled supplier discovery platform. Empirical data were collected through eleven semi-structured interviews involving internal stakeholders, procurement practitioners, and procurement industry experts. Data were analyzed using a hybrid analytical approach combining deductive and abductive reasoning through thematic analysis.

The findings indicate that AI adoption in procurement is driven by operational efficiency needs, information accessibility, supplier visibility, and increasing procurement complexity. At the same time, adoption is constrained by trust concerns, validation requirements, data quality limitations, organizational readiness differences, and procurement process realities. The study further identifies procurement maturity as an important contextual factor shaping AI value perception and adoption behavior. The research contributes to AI adoption literature by positioning procurement AI adoption as a workflow-embedded and socio-organizational process extending beyond technological implementation.

Warehouse order picking remains one of the most physically demanding operations in modern logistics, yet the AI vision systems increasingly deployed to monitor this work have been designed without accounting for the workforce heterogeneity that determines whether workload is fairly distributed. Adopting a sociotechnical systems theory perspective, this thesis examines what workload equity considerations relevant to workforce heterogeneity are absent from existing frameworks for the use of AI vision systems in warehouse order picking, and how such considerations can be integrated into existing frameworks for the use of AI vision systems to support worker well-being. A qualitative exploratory design was adopted, combining a systematic literature review conducted across two streams addressing workload equity in warehouse order picking and AI vision system design respectively, and five expert interviews with specialists in human-centered AI design, human factors, and workload assessment. The study derives the Workload Equity Lens from the human-side literature and applies it analytically against the AI vision system literature using the Computer Vision-based Hybrid Intelligence System (CV-HIS) framework developed by Zschech et al. and the Human-Centered AI for Industry 5.0 (HUMAI5.0) framework developed by Passalacqua et al. Four structurally absent considerations are identified: a strain-sensitive workload metric, a comparative equity function, workforce heterogeneity representation, and a distinction between equal task allocation and equitable harm outcomes. The thesis specifies how these considerations can be integrated across the CV-HIS design mechanisms and HUMAI5.0 process steps, and presents the Workload Equity Lens in prescriptive form as its primary theoretical contribution.

Pharmaceutical plastic waste is difficult to manage because it can contain active pharmaceutical ingredient residues, and is connected to quality, safety, and regulatory requirements. In most cases, incineration is used because it provides controlled destruction of contaminated waste. However, incineration of fossil plastic also generates direct and intense carbon dioxide emissions. This study assessed decarbonization opportunities for rejected pharmaceutical plastic ampoule waste, with focus on a fossil low density polyethylene (LDPE) based product stream containing both plastic and liquid fractions.

The study used an empirical case study approach. First, the current waste pathway was mapped, and a screening level global warming potential (GWP) calculation was used to identify climate impact hotspots. Then, possible decarbonization pathways were identified using production data, exploratory literature review, waste hierarchy logic, and pharmaceutical waste handling constraints. Finally, five scenarios were assessed under the same functional unit and system boundary: baseline incineration, prevention, carbon capture, pyrolysis, and mechanical recycling. The GWP calculation for each scenario separated production burden, waste packaging, transport, end-of-life (EoL) treatment, and avoided burden credits.

The baseline result showed that the production burden of the rejected ampoules was the largest contributor to climate impact, while incineration was the second largest contributor. Waste packaging and transport had small contributions. All alternative scenarios reduced total GWP compared with the baseline. Mechanical recycling showed the largest calculated reduction, decreasing the total climate impact to 74.11 percent of the baseline. Carbon capture and pyrolysis reduced the total impact to 87.04 percent and 88.04 percent of the baseline, respectively. The prevention scenario reduced the total impact to 94.89 percent of the baseline because it represented a partial reduction of rejected products rather than complete prevention.

The results show that downstream treatment substitution can reduce the climate impact of rejected pharmaceutical plastic waste, but it cannot remove the production burden already embedded in the rejected product. Therefore, deeper decarbonization requires both upstream prevention and improved treatment pathways for remaining rejects. Mechanical recycling appears most beneficial from a GWP perspective, but its feasibility depends on API residue control, washing performance, and acceptable external end use. Carbon capture is more compatible with the current safe destruction logic, but requires capture and carbon dioxide handling infrastructure. Pyrolysis provides a chemical recovery route, but requires separation and validation of API fate and degradation products. The study provides screening level decision support and identifies which pathways require further technical, environmental, and economic validation before implementation.

Tire and road particles generated by mechanical wear during contact between tires and the road surface constitute a growing environmental and health problem. Compared to exhaust pollutants, these emissions have received less attention, while previous studies have largely focused on finer particle fractions. As a result, coarser wear particles risk being neglected.

The purpose of this degree project was to investigate particle generation from friction between tires and the road surface during tests in a laboratory environment using a pendulum setup. It also aimed to describe the current state of particle generation from tire-road interactions and to investigate the possibility of establishing a relevant size interval for these emissions. The study specifically aimed to test the hypothesis that tire-road interaction gives rise to particles with a diameter larger than 10 μm, meaning particles outside the PM10 fraction.

The methodology was based on a literature review combined with quantitative measurement data from controlled experimental trials. The tests were conducted using a Wessex British pendulum tester, where a retreaded truck tire rubber sample struck an asphalt sample of the ABT 11 type. The generated particles were collected and analyzed using the Fidas Frog optical measuring instrument, which has the capacity to measure particle size distribution within the 0.18–93 μm range. The analysis was also supplemented with Cumulative Distribution Function (CDF) to describe the cumulative distribution of particle number and estimated particle volume.

The results showed that most of the measurement points fell within the particle size range of 0.2–1.1 μm. Regarding total particle number, about 40 % were found in the 0.2–0.3 μm range, while roughly 60 % fell within 0.3–5 μm. In terms of particle volume, approximately 90 % was found in the 1.1–5 μm interval. No particles larger than 5 μm were recorded during the tests. Consequently, the hypothesis of this study could not be confirmed.

This thesis investigates the expansion of Additive Manufacturing (AM) as a strategic tool for enhancing supply chain resilience within the defense industry, with specific focus on spare part production at BAE Systems Hägglunds in Örnsköldsvik. Against a backdrop of increasing geopolitical volatility and mounting pressure on global logistics networks, traditional just-in-case supply chains face compounding challenges: long lead times, high inventory costs, legacy part obsolescence, and deep subcontractor dependency. AM presents a pathway toward decentralized, on-demand fabrication that directly addresses these vulnerabilities.

The study employs a dual-framework methodology. Technology Readiness Levels (TRL), defined according to the European Commission's Horizon 2020 program, were applied to establish a commercial maturity threshold across all seven ISO 52900:2021-standardized AM process categories and their subtechnologies. The Analytic Hierarchy Process (AHP) was subsequently used to derive weighted procurement criteria from an internal expert survey distributed across R&D, engineering, and manufacturing departments at Hägglunds. The survey combined pairwise comparison questions evaluating five criteria, material properties, post-processing needs, operational reliability and knowledge, scalability, and product economics, with open-ended questions identifying organizational use cases and material preferences.

TRL proved effective only as a binary eligibility filter, since all commercially relevant AM technologies operate at TRL 8-9, making differentiation between candidates impossible on maturity grounds alone. The AHP results, while directional rather than mathematically precise due to an aggregated consistency ratio of approximately 27%, identified material properties and operational reliability as the dominant procurement criteria for both metal and polymer applications. Product economics ranked consistently lowest across both material groups.

Six technologies are recommended based on their alignment with identified use cases and AHP criteria weights: PBF-LB/M, PBF-EB/M, and Metal Binder Jetting for metals; Fused Filament Fabrication, Multi Jet Fusion, and PBF-LB/P for polymers. Together, these cover all seven use cases identified in the survey, ranging from structural spare parts and complex geometries to jigs, fixtures, and rubber-like components. No single technology covers all cases, and a coordinated deployment strategy is required rather than isolated procurement decisions.

AM shortens supply chains, reduces subcontractor dependency, and enables the transition from physical to digital inventory management. However, this digitalization introduces cyber-physical risks and new organizational knowledge requirements that must be managed alongside hardware investments. Entry-level technologies, particularly FFF, carry strategic value beyond their technical capability: early adoption builds internal AM competence and generates organizational demand for higher-specification systems over time. The study concludes that realizing AM's supply chain benefits at Hägglunds depends not on technology availability, but on how the organizational transition is structured and managed. Future work should conduct targeted AHP evaluations at the business unit level to translate these findings into concrete procurement decisions.

This study presents a framework for time-causal wavelet analysis. It targets real-time processing of temporal signals, where data from the future are not available. The study builds upon temporal scale-space theory, originating from a complete classification of temporal smoothing kernels that guarantee non-creation of new structures from finer to coarser temporal scale levels. We construct temporal wavelets from the temporal derivatives of a special time-causal smoothing kernel, referred to as the “time-causal limit kernel”—arising from the classification of variation-diminishing smoothing transformations with the complementary requirement of temporal scale covariance—to guarantee self-similar handling of structures in the input signal at different temporal scales. This enables the decomposition of the signal into different components at different scales while adhering to temporal causality. The study establishes theoretical foundations for these time-causal wavelet representations and maps structural relationships to the non-causal Ricker/Mexican hat wavelets. We also describe how efficient discrete approximations of the theory presented can be performed in terms of first-order recursive filters coupled in cascade, enabling numerically well-conditioned real-time processing with low resource usage. We characterize and quantify how the continuous scaling properties transfer to the discrete implementation, demonstrating how the proposed time-causal wavelet representation can reflect the duration of locally dominant temporal structures in the input signal. We propose this notion of time-causal wavelet analysis as a generic multi-purpose tool for signal processing tasks, where streams of signals are to be processed in real time—specifically for signals that may contain local variations over a rich span of temporal scales or more generally for analyzing physical or biophysical temporal phenomena, where a fully time-causal analysis needs to be physically realistic, by not accessing data from the future, as is often done for prerecorded data.

This study explores the conditions required for the successful implementation of Digital Workers in the Swedish manufacturing industry. The research is based on a qualitative case study consisting of two semi-structured interviews representing both a vendor and a user perspective. The empirical findings are analyzed through a thematic analysis based on three identified dimensions: technological, strategic, and organizational factors.

The study shows that technological factors such as data quality, ERP infrastructure, and clearly defined processes are fundamental prerequisites for a successful implementation. At the same time, strategic factors including clear business value, gradual implementation, and risk management influence organizations’ willingness to invest in the technology. Organizational factors are found to be particularly significant, where project champions, change management, employee acceptance, and “human-in-the-loop” emerges as critical prerequisites of successful implementation.

The study concludes that the implementation of Digital Workers should not merely be understood as an IT-project, but rather as an organizational transformation process. Digital Workers in the Swedish manufacturing industry are driven not by technological curiosity, but by operational necessity linked to efficiency and increasing complexity that traditional methods can no longer adequately address. The technology’s potential is realized only when technological, strategic, and organizational factors interact effectively. Consequently, this study contributes to a deeper understanding of how Digital Workers can be implemented within Swedish manufacturing companies and which challenges must be addressed for the technology to generate long-term value.

Additive manufacturing (AM) has evolved over the past few decades from being primarily used for prototyping to becoming an increasingly important part of modern manufacturing processes. The technology enables complex geometries and great design freedom, creating the conditions for more flexible and resource-efficient production. At the same time, the implementation of AM is associated with high investment costs, limitations in materials and processes, and challenges associated with series production. This has led to an interest in investigating under what conditions the technology becomes economically viable compared to conventional manufacturing methods.

The purpose of the study is to analyze when it is profitable for companies to implement additive manufacturing in their manufacturing process. The study focuses in particular on the advantages and disadvantages of AM compared to conventional manufacturing methods, for which types of components the technology is economically advantageous, and how new technologies affect the profitability of implementing AM.

The method for the work consisted of a literature study with a review of scientific articles and previous research. To complement the literature study, empirical data was collected through an interview with Sasan Dadbakhsh, researcher and lecturer in additive manufacturing at KTH. The combination of theory and interview enabled a more nuanced analysis of the technology's economic and technical conditions.

The results show that AM is primarily profitable in the manufacture of complex components, low-volume production, spare parts with uncertain demand, prototypes and in iterative product development. The technology creates particular economic benefits through reduced inventory, shorter lead times and the possibility of local and demand-driven production. The study also shows that Design for Additive Manufacturing (DfAM) and the development of existing AM methods are essential to fully exploit the potential of the technology. At the same time, it emerges that AM still has limitations linked to high production costs, material limitations and lack of competitiveness in mass production. The conclusion is therefore that additive manufacturing is unlikely to fully replace conventional manufacturing methods, but rather be integrated as a complement in specific parts of the manufacturing process where the technology creates clear economic and functional added value.

Many low-traffic railway lines currently lack level crossings with active protection in the form of warning lights, audible signals, or barriers. Upgrading these crossings is often limited by the absence of access to the local power grid, which complicates the installation of conventional protection systems. At the same time, the development of energy-efficient components and alternative energy solutions, such as solar power and battery systems, has created new opportunities for implementing level crossings with local energy supply.

The aim of this thesis is to investigate how existing unprotected, low-traffic level crossings can be upgraded with warning functions through the use of alternative energy solutions while maintaining the safety requirements of the Swedish Transport Administration. The study includes an analysis of the energy demand for different protection levels, for example warning lights with or without barriers, as well as an evaluation of which energy solutions are technically and operationally realistic.

The method is based on a literature study of the Swedish Transport Administration’s regulations and technical documents, combined with analytical modelling of the level crossing’s energy demand divided into base load and event-driven load. The results show that the continuous base load is the dimensioning factor for the energy system, while the event-driven load has limited impact in low-traffic cases. The study also shows that the choice of component technology is crucial, as LED-based solutions significantly reduce the energy demand compared with incandescent lighting. Solar cells and battery storage appear to be a feasible main solution for the most energy-efficient level crossings. Wind power may contribute as a complement, but is not considered sufficiently robust as a standalone solution. The most robust solution is therefore assessed to be a hybrid system in which solar cells and battery storage form the main power supply, while a diesel generator or additional battery reserve is used as backup power to ensure operation under winter conditions.

What is new?

There is a widespread lack of clarity and consensus regarding the role of Research Project Managers (RPMs) in collaborative research projects, which this paper aims to address. This is the most extensive attempt to define the role of a Research Project Manager (RPM) in collaborative research projects.

What was the approach?

The paper draws on the authors' combined experience managing collaborative research projects, synthesising insights from theoretical frameworks and practical implementations in diverse research environments.

What is the academic impact?

The paper provides a much-needed definition of the Research Project Manager's role, distinguishing it from related roles in the Research Management and Administration (RMA) field. It contributes to RMA and project management by offering a framework that clarifies the RPM's responsibilities, which have largely been underexplored in academic literature.

What is the wider impact?

For research and management practitioners, this paper highlights the critical role of RPMs in ensuring project success, from ideation to execution. It underscores the importance of well-defined project management practices in collaborative research to meet the expectations of funding agencies and ensure scientific outcomes.

PurposeTo investigate the motivational profiles of research project managers (RPMs) using self-determination theory (SDT) and explore their correlation with perceived organizational role value appreciation.Design/methodology/approachA quantitative cross-sectional online survey was conducted with 276 RPMs. Data were analyzed using exploratory factor analysis, Kruskal–Wallis tests and Spearman's correlation to identify underlying motivation profiles and their relationships with perceived role appreciation.FindingsThree distinct motivational profiles emerged: autonomous (predominant), controlled and outcome-driven. Although RPMs generally report low role appreciation from their organizations, higher perceived appreciation strongly correlates with outcome-driven motivation.Research limitations/implicationsLimitations include a cross-sectional design and self-reported RPM identification.Practical implicationsOrganizations must enhance perceived support by granting RPMs autonomy, development opportunities and visible role appreciation to mitigate demotivation and turnover, ultimately boosting project productivity.Social implicationsImproving the retention and motivation of RPMs fosters more stable and effective research and innovation ecosystems, reducing the administrative disruption and turnover that often troubles collaborative scientific endeavors.Originality/valueThis is the first study to empirically map RPMs' motivation profiles using SDT and link them to perceived organizational support. This study provides a contextual empirical application of existing SDT dimensions within research and innovation environments. It maps the established concepts of identified and integrated regulation into a practical “outcome-driven” profile tailored specifically to the realities of research project management.

Objectives:

This study compared longitudinal cognitive decline across major dementia subtypes using data from two national cohorts, the Swedish Registry for Cognitive/Dementia Disorders (SveDem) and the U.S. National Alzheimer’s Coordinating Center (NACC). We also examined how demographic and clinical factors influenced these trajectories.

Methods:

Longitudinal Mini-Mental State Examination (MMSE) scores were analyzed with generalized additive mixed models (GAMM-like mixed-effects). Each model included a natural-spline time term and random intercepts for individual patients. Fixed effects comprised baseline age, sex, education, civil status, comorbidity index, and baseline MMSE. The harmonized datasets included 22,937 trajectories from SveDem and 42,593 from NACC, each with at least two MMSE or MoCA-equated assessments. In NACC, diagnostic transitions such as MCI progressing to Alzheimer’s disease were modeled dynamically.

Results:

Distinct non-linear cognitive decline patterns were observed across subtypes. Annual adjusted MMSE change (mean [95% CI]) was greatest in Alzheimer’s disease (SveDem −1.51 [−1.76, −1.26]; NACC −1.73 [−1.93, −1.52]) and mixed dementia (SveDem −1.83 [−2.11, −1.55]). Moderate decline was found in vascular (SveDem −0.95 [−1.21, −0.69]; NACC −1.41 [−2.00, −0.82]), Lewy-body (SveDem −0.72 [−1.18, −0.25]; NACC −1.28 [−1.87, −0.70]), and frontotemporal dementia (SveDem −0.71 [−1.24, −0.19]; NACC −1.59 [−2.03, −1.14]). In NACC, Healthy and MCI groups showed minimal decline (−0.07 and −0.55 points per year). Fixed-effect heatmaps indicated that baseline MMSE was the main predictor of subsequent cognition, followed by smaller positive effects of education and negative impacts of comorbidity and single marital status.

Conclusions:

Across both registries, Alzheimer’s and mixed dementias showed the fastest cognitive decline, while vascular, Lewy-body, and frontotemporal dementias were more variable. Harmonized GAMM modeling with baseline adjustment and dynamic subtyping provides a robust framework for comparing dementia progression between national cohorts.

Cognitive impairment progressively limits independent living, yet scalable methods for identifying individuals at future risk remain limited. In the 2026 George B. Moody PhysioNet Challenge, our team (Matcha) developed a hierarchical framework for predicting cognitive impairment from overnight polysomnography (PSG). Modality-specific convolutional stems learned a 192- dimensional representation for every 30-s epoch from human and automated sleep annotations together with crossview self-supervision. Local and full-night Transformers then aggregated the epoch sequence into a patientlevel prediction, while a low-capacity record-wise residual represented whole-night sleep summaries and outcomeobservation timing. In the three-site leave-one-site-out (LOSO) validation, it achieved a macro age-conditioned AUROC of 0.804. On the official phase hidden validation set, the best Small- and Large-track entries achieved 0.847 and 0.844, respectively, ranking 1st out of 513 entries.

Trimble Sweden AB is an industry-leading manufacturer of land surveying and scanning equipment. Despite this, Trimble currently uses an aging, manual manufacturing process for its high-volume instruments. Therefore, Trimble aims to develop a new manufacturing platform that streamlines the product flow and reduces manual handling and intervention.

This thesis project addresses the critical cleaning process of optical components for cleanroom assembly. Proper cleaning is essential to minimize the discarding of semi-finished instruments due to contamination. Currently, this manual, two-step process, which consists of ultrasonic solvent cleaning followed by O2 plasma surface activation, acts as a major bottleneck in production. Aging equipment shared between departments and suboptimal storage environments lead to significant inefficiencies, frequent maintenance, and wasted time.

The goal of this thesis project was to investigate whether automating this process was feasible, or what trade-offs need to be made if not all requirements can be met. As a basis for this project, contact was made with manufacturers and vendors with expert knowledge in their respective technologies, i.e., ultrasonic cleaning and plasma surface activation.

The study has shown that automating Trimble's cleaning process is entirely possible using existing technology. The concept integrates an ultrasonic cleaner from Ecoclean GmbH and a customized plasma system from Diener Electronic GmbH, interconnected via automated conveyors. While the currently used ethanol effectively dissolves fats, as the primary contaminants at Trimble’s facility are dust and particles, Ecoclean recommends evaluating a water-based medium for optimal removal. Overall, the solution provides Trimble with a robust foundation to eliminate bottlenecks, enable just-in-time deliveries, and significantly improve the work environment.

This study is looking into capacity limitations in pre-assembly lines in paced assembly lines, focusing on Volvo's gearbox factory in Köping. In today's automotive industry, there is a high demand for both customization and high production efficiency, which means that assembly lines need to be able to handle large variations in the product mix while simultaneously maintaining a stable flow. This becomes problematic when there is a capacity mismatch between the feeder and the main line, resulting in the main line waiting for the slower pre-assembly line. As a result, there is a risk for bottle necks and disruptions in the production tact time.

The aim of this report is to concretize the problems concerning capacity limitations in pre-assembly lines, evaluate best practice based on previous studies and other companies, and come up with an action plan for how Volvo could handle their assembly line issue. The study is conducted as a literature study combined with a case study at Volvo's gearbox factory in Köping. For collecting data, a large part was through scientific articles, combined with a study visit at Volvo in Köping and interviews with some personnel involved in the production planning process and the development of the assembly system.

The final result shows that capacity limitations in pre-assembly lines affect the flexibility, planning and sturdiness of the assembly system. The study identifies several possible solutions, including overcapacity, offers and additional efforts in connection with forecasting, buffers and inventory management, flexible manual assembly, sequencing, improved line balancing and to accept limitations. Whichever solution is the most suitable depends on many different factors such as how demand varies, how reliable the forecast is, economical possibilities and the organization's strategic objectives. The results also show that different solution strategies means balancing between different properties in the system such as flexibility, customer adaptations, resource and cost efficiency.

For Volvo's factory in Köping, a combination of limited buffers, improved sequencing and continuous evaluation of line balancing seems to be the most suitable best-practice solution for a majority of the pre-assembly lines. However, in some cases it can still be more economical and strategic to accept capacity limitations, and sometimes it can be necessary to apply overcapacity to meet a varying demand. The study also identifies a potential for improvement in developing a more detailed and short-term forecasting linked to the relationship between customer orders and the need of specific modules in the pre-assembly. This could enable a more pro-Active planning and a more efficient management of variations in demand.

This thesis investigates how large-scale systems can make effective use of limited communication, computing, and observation resources when operating under uncertainty, incomplete information, or evolving conditions. Across the problems considered, a common conclusion emerges: reliable and efficient performance does not necessarily require continuously updated information or simply increasing capacity and computational power; instead, it depends on identifying the stochastic structure of the system and using it to determine when information is needed and where resources are most valuable. To this end, this thesis develops stochastic models, analytical methods, and optimization frameworks for three problems in edge intelligence and smart microscopy.

The first contribution considers load balancing in distributed computing systems in which users independently select servers without access to a continuously updated global system state. Instead, users construct local estimates from their own routing decisions and from occasional status updates broadcast by the servers. The analysis establishes conditions under which such decentralized operation remains stable despite sparse information, communication losses, and heterogeneous server capacities. Increasing the relative update rate generally reduces the stationary backlog and brings the distributed policy closer to the centralized benchmark. However, for sparse update rates, the backlog can be non-monotonic in the update reception probability: moderate communication losses diversify user beliefs and can mitigate synchronized routing based on outdated server states, whereas at higher relative update rates, more reliable reception improves performance.

The second contribution addresses the joint dimensioning of wireless and computing resources in multi-cellular edge-intelligent networks. Wireless access, spatially distributed traffic, edge-server queueing, and AI inference are modeled as a coupled stochastic system subject to statistical requirements on latency, coverage, server load, and inference accuracy. Stochastic geometry is used to characterize the spatial properties of the cellular network and to derive tractable uplink-capacity expressions. These results are integrated with queueing models and empirically characterized inference workloads to formulate an end-to-end resource-dimensioning problem for bandwidth and computing capacity. The resulting analysis exposes trade-offs that cannot be identified when communication and computation are designed independently. In noise-limited systems, denser infrastructure improves wireless conditions but reduces statistical multiplexing at the servers, creating an opposing computing cost. In interference-limited systems, densification with fixed frequency reuse may provide little benefit because increased interference offsets shorter transmission distances; when frequency reuse is adapted with network density, densification can again reduce both communication and computing requirements.

The third contribution considers adaptive observation in automated microscopy, where many asynchronously evolving regions of interest compete for a limited imaging budget. The biological dynamics are modeled as large ensembles of independent Markov processes, for which exact and asymptotic characterizations enable a scalable representation of their aggregate temporal evolution. Building on this structure, the observation problem is formulated as a restless multi-process multi-armed bandit under partial observability, and a computationally efficient Whittle-index policy is derived. Numerical results show that the policy is particularly effective when temporal persistence makes past observations informative about future activity. Experiments on biological microscopy data further show that the proposed policy captures 93% more biologically relevant events than fixed cyclic scanning.

Taken together, these contributions show that reliable operation in resource-constrained stochastic systems does not necessarily require complete information, continuous coordination, or ever-increasing infrastructure.  Instead, the key is to identify which aspects of uncertainty and system history are relevant to decision-making, and to allocate communication, computation, and observation resources accordingly. The thesis provides analytical and algorithmic tools for determining when information is sufficient for distributed coordination, how communication and computation should be provisioned jointly, and how scarce observations can be directed toward the most informative parts of a dynamically evolving system. More broadly, the results illustrate how uncovering and exploiting stochastic structure can turn uncertainty into a source of actionable information, enabling more efficient and informed decisions even when resources are fundamentally limited.

In a market characterized by rapid technological advancement, the ability to efficiently execute New Product Introduction (NPI) is a critical competitive factor. This thesis investigates administrative preparation flows at Sandvik Coromant in Gimo, where challenges regarding long lead times and rework have been identified despite extensive digitalization. The purpose of the study is to identify and analyze bottlenecks in these information-intensive processes and evaluate how the integration of Theory of Constraints (TOC) and Lean tools can reduce lead times and improve process quality. Adopting an abductive research approach and a case study design, data was collected through semi-structured interviews, site visits, and observations, subsequently analyze dusing Value Stream Mapping (VSM) and supported by an exploratory Discrete Event Simulation (DES). The results indicate that the technical assessment stage constitutes the system’s primary constraint, mainly due to insufficient information quality (%C&A) from preceding steps, which necessitates time-consuming rework. By applying TOC principles to protect the constraint through standardised checklists, automated data transfer, and increased industrialization of work methods, the study indicates a potential to reduce lead times while improving process efficiency. The study suggests that administrative constraints are most effectively managed by focusing on the integrity and structure of the information flow rather than merely increasing resource capacity.

Factory location is a long-term strategic decision that influences the cost efficiency, service quality, and environmental performance of a manufacturing supply chain. This thesis addresses the location problem faced by Westlake Instruments, a Chinese manufacturer of laser slicing equipment for silicon carbide, diamond, and AR waveguide wafers. The company’s forecast demand for the 2026 to 2028 horizon exceeded the production capacity of its existing factory in Hangzhou by a wide margin, which motivated an evaluation of where one or more new factories could be located.

The location problem is formulated as a bi-objective uncapacitated facility location problem with two conflicting objectives: minimizing total cost (fixed investment in land and labor plus transport cost) and minimizing the maximum service distance to any demand point. The model is parameterized using company-supplied demand data covering 63 customers across 78 sites, public industrial land prices and recruitment wages for 33 candidate cities across China, and actual road distances retrieved from the OSRM routing service. The Pareto front of efficient trade-off solutions is generated using two methods, the epsilon-constraint method and the weighted sum method, which are then compared in terms of solution quality and computational behavior.

The epsilon-constraint method recovers the complete Pareto front of 15 distinct solutions, ranging from a cost-optimal configuration with two facilities, a total cost of 1,457 wan yuan, and a maximum service distance of 3,393 km, to a distance-optimal configuration with seven facilities, a total cost of 10,502 wan yuan, and a maximum service distance of 649 km. The weighted sum method identifies 12 distinct solutions but misses several intermediate alternatives in non-convex regions of the front. Across all Pareto-optimal solutions, fixed cost contributes between 91 and 99.9 percent of total cost, indicating that the location decision is dominated by fixed investment rather than transport cost. Sensitivity analysis confirms that this conclusion is robust across a wide range of cost and demand assumptions. The most attractive solution from a managerial perspective lies near the knee of the Pareto front, where opening one additional facility in central China reduces the maximum service distance by more than half at a limited cost increase. The results suggest that this type of central location provides the strongest compromise between cost efficiency and service responsiveness, while also offering potential environmental benefits and strategic advantages related to logistics infrastructure, regional coverage, and access to technical expertise.

This master’s thesis presents the development and evaluation of a methodology for predicting allowable defect sizes in highly stressed cast components subjected to fatigue loading. The work links material-level fatigue data to component-level assessment using Kitagawa–Takahashi diagrams and local stress states obtained from finite element and fatigue simulations. Material-specific fatigue crack growth thresholds, ∆Kth, are determined at several load ratios R for EN-GJS-500-7 ductile cast iron and EN AC-46000 cast aluminium. These measurements are used to generate R-dependent Kitagawa relations. Because real components exhibit continuously varying mean stress and stress amplitude, a numerical interpolation procedure is introduced between the discrete experimentally characterised R levels. The experimentally derived Kitagawa relations are implemented in a Python-based computational module. Local stress amplitudes and mean stresses from Abaqus and FEMFAT are used to calculate node-wise R values and allowable defect sizes. The calculated values are written back to Abaqus ODB files as field variables, enabling direct visualisation of component-level defect tolerance. A complementary validation route uses fatigue-limit testing of U- and V-notched specimens at R = 0.1. Fatigue limits are determined using the staircase method, while the notch geometries are converted to equivalent defect sizes using Murakami’s √area approach for comparison with the corresponding Kitagawa prediction. The methodology is applied to two industrial demonstrators: an EN-GJS-500-7 engine mount and an EN AC-46000 flywheel/transmission housing. The results show that local defect tolerance depends on both stress amplitude and the R-dependent threshold dataset used by the workflow. For negative R, a dedicated inverted-specimen test procedure was used and the physical load ratio was reconstructed during test processing. The resulting measured threshold values are then used consistently in the subsequent Kitagawa and component-level calculations. The developed workflow therefore provides a basis for local, physics-based defect requirements and can support simulation, material, design and casting-quality decisions for fatigue-critical cast components.

Stockholm University has conducted research activities as part of an EU-funded project nPETS (GA No 957377), aiming to identify health effects on lung cells from nanoparticles arising in the transport sector. This is done by exposing lung cells to ambient air from the transport sector in an ALI system (Air Liquid Interface). However, the concentration of nanoparticles in ambient air is too low to produce a clear effect, and this is therefore addressed using a particle concentrator, which increases the proportion of nanoparticles in the sample air. Stockholm University thus has a need to construct a particle concentrator to conduct further research in this field. A version of a particle concentrator has been designed and tested by students at the Royal Institute of Technology, but has not achieved sufficient nanoparticle concentration. The report summarizes the original concentrator's design, identifies deficiencies in its subsystems, and describes and validates the measures taken to increase concentration. Deficiencies were identified in the growth tube's water supply and drainage, the system's air-tightness, and particle losses in the system. Validation of the implemented measures shows that the water supply system has been improved, and that the system is now air-tight. Test results show that the growth tube enlarges particles, but not sufficiently. Test results also show that particle loss occurs in both the drying tube and the growth tube. Overall, an increase in concentration cannot be identified within the concentrator. To meet the requirement of a concentration factor of at least 5x, further development of the concentrator's design is required.

In recent years it has become increasingly important for companies to improve resource efficiency within their work processes. Toyota has spent several decades working to reduce waste and create a resource efficient production system through the Toyota Production System (TPS). The core pillars of TPS – Just-In-Time, Jidoka, and Kaizen – have contributed to Toyota’s success. Despite this, many companies outside Japan have struggled to achieve similar results, even after implementing comparable methods and tools.

The purpose of this report is to examine how Toyota in Japan has adapted and further developed TPS to ensure resource efficient production under changing demographic and global conditions, as well as to analyse why companies in other countries often fail to achieve equally successful results. For this reason, Scania, the Scania Production System (SPS), the Toyota-GM joint venture NUMMI, and GM’s Van Nuys plant have also been studied. The report is based on a literature review in which scientific articles, books, and company materials were analysed. The material was primarily collected from the databases Google Scholar, Scopus, and Web of Science, and was complemented with information from Toyota’s and Scania’s official websites.

The conclusion is that TPS has continuously needed to evolve mainly due to changing market demands, technological innovations, labour shortages, and supply chain disruptions caused by natural disasters. To address these challenges, Toyota has for example developed stronger collaboration and information sharing with suppliers, improved oversight of the supply chain, and in some cases increased the stockpiling of particularly important components following events such as the Tohoku earthquake and the COVID-19 pandemic. At the same time, the company has continued to develop working methods focused on flexibility, problem-solving, and continuous improvement to manage new situations. Despite these challenges, the fundamental philosophy behind TPS has remained relatively unchanged since the system began to take proper shape after the second world war. The study also shows that TPS implementation works best when organizations understand the system as a whole, where leadership, respect for employees, work practice, and tools interact, rather than focusing solely on introducing isolated tools and methods.

The development of Scania, SPS, together with the cases in NUMMI and the plant in Van Nuys demonstrates that principles from TPS can be successfully implemented outside Japan, provided that the systems are adapted to local conditions.

This study examines what types of machine data are available from two heavy truck manufacturers, Volvo and Scania, through the rFMS (remote Fleet Management System) standard, how the data is structured, and what potential it holds for creating value in Skanska’s construction operations. The study combines a technical comparison of available data points, a quantitative regression analysis, and qualitative interviews with representatives for Enterprise Architecture and data management within Skanska Sweden. The results show that Scania (rFMS 4.x) provides a significantly broader range of data than Volvo (rFMS2.1), including diagnostics, fuel level, brake air pressure, and service distance. Volvo shares no fuel consumption data for trucks, which substantially limits the scope for quantitative analysis. The regression analysis was conducted as a methodological demonstration with n = 5 driving segments and cannot produce statistically significant conclusions. It does, however, demonstrate a scalable analysis pipeline from raw data to model selection using best subset selection and leave-one-out cross-validation. The final model identifies average speed and cruise control usage as theoretically motivated predictors for fuel consumption.

The interview study indicates that Skanska is in an early stage of its digital transformation with regard to machine data. The technical foundation is in place, but organisational prerequisites including clear data ownership, shared conceptual models, and a communicated vision are lacking. Drawing on Kotter’s change management model and research on digital transformation in the construction sector, the study concludes that technology is rarely the limiting factor. Rather, it is organisational climate and information exchange that determine whether data can be translated into decision support.

The study recommends that Skanska prioritise organisational capacity building in parallel with expanded data collection, and begin with bounded use cases such as reducing idle time and improving service planning to generate early, tangible results.

This thesis was conducted at WesDyne Sweden AB, which primarily performed assignments in heavy industry, mainly in underwater environments. The company frequently used remotely operated underwater vehicles (ROVs) in its work. These could be commercially available or custom-built. When commercially available equipment was used, adaptations were often required to meet industry-specific requirements. In this thesis, a design study was carried out on the structural adaptation of a commercially available ROV. The purpose was to evaluate material selection and joining methods for the frame structure in order to better meet the requirements of the relevant industrial environment. The requirements mainly included a reduced risk of parts coming loose and improved possibilities for cleaning the equipment of residues from process products. A system analysis was performed to delineate which components were considered part of the frame structure, thereby excluding the propulsion system and electronics. The delineated frame structure was then analyzed to identify critical interaction points with the remaining subsystems. Based on a limited selection of manufacturing materials deemed suitable for the industry in question, a concept for a replacement frame structure was developed and modeled in a CAD environment. The model was evaluated through a design review at WesDyne, using a full-scale 3D-printed model and CAD visualization. As a next step, testing of selected materials was proposed to verify the functionality of a finished frame structure.

This thesis examines whether financial sponsorship from private equity (PE) or venture capital (VC) influences the characteristics, pricing, and post-listing performance of initial public offerings (IPOs) in the Nordic markets between 2016 and 2021. Using a sample of 565 IPOs across the Nordics, of which 144 were sponsor-backed, the study compares issuer scale, leverage,valuation multiples, offer structure, and three-year stock performance. The results show that sponsor-backed issuers are significantly larger at listing and more often EBITDA-positive, yet they do not command higher valuation multiples at the offer. Sponsorbacked deals are far more likely to include secondary share sales, while lock-up agreements are near universal with similar average durations. Return analyses reveal little sustained momentum beyond the first weeks and tentative reversal between months 3 and 12. Overall, sponsorship appears to act more as a quality-and-design factor, through larger, more mature issuers and structured exits, than as a driver of systematic pricing or long-run outperformance.

This study examines cross-functional coordination and information flow management during high-volume promotional planning processes within the fast-moving consumer goods (FMCG) sector. Promotional events introduce significant uncertainty into demand planning activities due to fluctuating customer requirements, operational constraints, and variable demand patterns, increasing the need for effective coordination across commercial and operational functions. While previous literature emphasizes the importance of Sales and Operations Planning (S&OP), demand planning, and cross-functional integration, empirical understanding remains limited on how information flows and coordination mechanisms operate in practice during uncertain promotional environments.

To address this gap, the study adopts a qualitative case study approach based on semi-structured interviews conducted with twenty internal stakeholders. Participants represent functions across demand planning, supply planning, transportation, warehousing, order management, and commercial roles, all involved in the end-to-end planning of promotional events. The findings reveal that, although collaboration between departments is generally perceived positively, organizations continue to face significant coordination challenges associated with fragmented and inconsistent information flows, leading to a reactive risk visibility and limited proactive planning capabilities.

The study finds that stakeholders require different types of information depending on their operational responsibilities, KPIs, and planning horizons, highlighting the importance of role-specific information flows and continuous feedback mechanisms. Based on the findings, a cross-functional information flow framework is proposed. This framework includes a risk matrix that categorizes events into three scenarios, along with stakeholder-specific information requirements and closed-loop feedback mechanisms to support proactive coordination and uncertainty management.

The study contributes to the literature by distinguishing collaboration from coordination in promotional planning environments and by emphasizes the critical role of information visibility and feedback loops for effective cross-functional alignment under uncertainty. he proposed framework offers organizations a structured approach to enhance coordination, improve information sharing, and strengthen operational decision-making during high-volume promotional events.

This report describes an investigation of the efficient use of resources in additive manufacturing compared to conventional methods. Conventional methods in the form of investment casting is characterized low casting yields and high rejection rates, aswell as extensive lead times and high cost. This study seeks to compare the manufacturing methods through an extensive review of literature to make an analysis focused on the material utilization, cost of production through labor aswell as lead time. The report processes the manufacturing of gas turbine blades both conventionally and by additively. Gas turbine blades involve complex geometries and strict strength requirements from intense physical strain, which limits the amount of feasible manufacturing methods. The literature review is supplemented by a interview of a person with expertise to reinforce the theory and highlights clear advantages with a transition to additive manufacturing such as Selective Laser Melting, from the conventional Investment casting regarding the utilization of materials and lead time, but should be revised and include a inquiry about the mechanical properties of the turbine blade produced from each manufacturing method.

The textile industry is one of the most resource intensive industries in today’s society. The industry stand for approximatley 10% of global carbon dioxide emissions. A large share of this waste arises already in the production, in the form of material waste during cutting of fabrics, process losses and incomplete use of yarn. Within the luxury segment this becomes even more complex because of the high priority of quality standards.

The purpose of this study is to examine how manufacturing companies within the luxury segment work to minimise material waste in production without affecting the product quality, and providing insights that similar companies may benefit from.

A case study was conducted at a weaving mill, based on an interview with a company representative and a literature review covering material waste, lean production, the circular economy, and quality management. The interview responses were analysed in relation to the theoretical framework.

The study shows that waste minimalisation within the luxury segment is possible, but should take place within the limits where product quality is not affected. Quality functions as a structural boundary for how far waste minimalisation can be driven, and companies should therefore adopt selective and carefully chosen measures to ensure that the stages where product quality is created remain unchanged.

The automotive industry faces increasing demands for efficiency and flexibility in production, where buffer management plays a critical role in maintaining a smooth production flow. This study investigates how optimization of buffer management, through the application of Industry 4.0 technologies, can improve production efficiency in the automotive sector.

The study is based on a literature review of existing research in buffer management and Industry 4.0, combined with a case study consisting of interviews with people working in the automotive industry with knowledge in buffer management and Industry 4.0 implementation.

The literature review shows that real time data analysis combined with artificial intelligence and digital twins offers theoretical potential for dynamic buffer optimization, reducing capital tied up in buffers and minimizing production disruption. However, the case study reveals that practical implementation of dynamic buffer sizing is deemed unnecessary, as production flows in modern automotive plants are already well planned. Furthermore, the cost of developing AI systems and digital twins are high, with limited short term return. Instead alternative Industry 4.0 applications were identified as more immediate improvements.

The conclusion drawn is that while theoretical opportunities exist to optimize buffer management through Industry 4.0, the practical impact is limited. The most promising applications don’t lie in direct buffer resizing but rather in other Industry 4.0 technologies.