Filetype
Bakgrund: Otillräcklig postoperativ smärtlindring förlänger återhämtningen och ökar opioidförskrivningen med risk för biverkningar och beroende. Medicinsk akupunktur hardärför föreslagits som ett opioid-besparande komplement. Syfte: Att undersöka om medicinsk akupunktur reducerar postoperativ opioidförbrukning hos vuxna patienter som opererats i generell anestesi, samt att värdera effekten i rela
In recent years, machine learning models have changed the field of video denoising. Most state-of-the-art models utilize multiple frames as input to denoise a single frame, which limits the temporal context window and often requires more computational power. Another, less explored, approach is to use recurrence to gain a hidden memory state in which temporal information can be stored. While the fo
Syfte: Studiens syfte var dels att kartlägga utformning av språkligt landskap i klassrumsmiljö för elever med språkstörning, dels att undersöka skolpersonalens attityder kring flerspråkig representation i skolorna. Metod: Tre skolor från två olika kommuner deltog i studien. Samtliga skolor bedrev vid tidpunkten för studien specialundervisning för elever med språkstörning. Datan samlades in genom
Stora språkmodeller (LLM) har uppvisat anmärkningsvärda förmågor inom en rad olika uppgifter inom naturlig språkbehandling (NLP). Deras prestanda är dock mycket känslig för den inledande prompt som används för att fråga dem. Manuell promptkonstruktion, även om den är effektiv, är ofta tidskrävande och kräver betydande expertis. Denna avhandling undersöker promptoptimering, med fokus på en automatiLarge Language Models (LLMs) have demonstrated remarkable capabilities across various Natural Language Processing (NLP) tasks. However, their performance is highly sensitive to the input prompt used to query them. Manual prompt engineering, while effective, is often time-consuming and requires significant expertise. This thesis investigates Prompt Optimization, focusing on a closed-loop, automated
In the pursuit of more sustainable urban energy systems, recovering industrial waste heat presents a promising opportunity to enhance energy efficiency and reduce emissions. This thesis explores the technical and economic feasibility of capturing excess heat from an industrial process and repurposing it to supply a new low-temperature district heating network planned for a new residential district
This study investigates the dynamic response of a glass pane under impact loading utilizing modal analysis and Newmark’s method. The benefits and drawbacks of modal analysis as well as Newmark’s method in this context are also investigated. The study aims to simulate the testing of glass panes according to the testing method described in SS-EN 12600 where an impact device is swung in a pendulum mo
In the context of the current climate crisis and a growing need for electricity, renewable energy plays an important role. Repowering means upgrading existing renewable energy production with new, more efficient technology to increase production in older sites. Wind-power technology has developed significantly in the last few years and upgrading wind parks in the most suitable locations is relevan
Detta projekt avsåg att optimera ett automatiskt backspolande filter för emulsionspolymerer på the Dow Chemical Companys anläggning i Landskrona, samt att jämföra det med det befintliga påsfiltersystemet. Då filtret använder färdigprodukten för att backspola filtret fokuserade optimeringen på att minska mängden produkt som spolas ut, så kallat reject, då detta medför avfallskostnader och förlust aThis project aimed to optimize an automatic backwash filter for emulsion polymers at the Dow Chemical Company's facility in Landskrona, and to compare it with the existing cartridge filter system. Since the filter uses the finished product to backwash the filter, the optimization focused on reducing the amount of product that is flushed out, so-called reject, as this entails disposal costs and
Global plastic waste production has more than doubled in the past two decades, with plastic packaging accounting for 40% in 2019. Due to low recycling rates, pollution and limited fossil resources, increasing circularity in the plastic industry has become increasingly important. Chemical recycling through pyrolysis, unlike mechanical recycling, can treat contaminated, mixed plastic waste and incre
In paediatric hospital wards, parents or caregivers who are physically capable are already present with their child when a fire emergency occurs. Yet existing approaches to hospital evacuation have either overlooked their potential role or treated them as additional occupants for staff to manage. If these caregivers could evacuate independently with their child, staff could direct their efforts to
Interrogation leaders in high-stakes investigative settings must convert long interview sessions into accurate written protocols under time pressure, stress, and cognitive bias. While AI has the potential to streamline this work, generic large-model assistants risk replacing one form of cognitive effort with another by inviting passive acceptance of generated output. To address this, we de- signed
Cleanrooms in the pharmaceutical industry are energy-intensive environments where HVAC systems account for roughly 50-75% of energy use due to high air changes per hour and strict conditioning requirements for indoor humidity and temperature. This thesis presents a techno-economic analysis comparing two of the most common HVAC system types. These systems are centralized, using one large AHU to pro
Abstract: To meet the future demand for sustainable and environmentally friendly “green” chemical cleaning solutions for rust removal from naturally oxidized iron alloys, this study showed how an optimized DES (Deep Eutectic Solvents) can enhance descaling performance. A series of experiments were conducted using different combinations of acetic acid, malonic acid, and choline chloride in aqueous
Large language models (LLMs) have demonstrated exceptional capabilities in complex mathematical reasoning; however, their massive parameter scales impose prohibitive computational costs, limiting their deployment in resource constrained or local environments. To address these challenges, this thesis investigates advanced knowledge distillation methodologies to enhance the mathematical problem-solv
The assessment of sunlight access in buildings is constrained by methods that rely on simplified point-based evaluation. EN 17037 sunlight exposure evaluates sunlight at a single point on the aperture under restricted temporal conditions. This limits the representation of how sunlight interacts with the aperture over time. These characteristics reduce its applicability at early design stages. The
Unmanned ground vehicles (UGVs) operating in uneven and unstructured terrain are often limited by fixed sensor mounting, which can reduce mapping quality, situational awareness, and sensor performance when the vehicle body experiences pitch, roll, and vertical motion. This thesis presents the development of a compact four-degree-of-freedom (4-DOF) gimbal prototype for a terrain-mapping module on a
Den svenska byggsektorn står för en betydande del av samhällets energianvändning och miljöpåverkan, varav installationer och uppvärmning står för majoriteten av energianvändningen. Ventilationssystem i våra bostäder behövs till vår hälsa, komfort och byggnadens hållbarhet men bidrar även till kontinuerliga energiförluster under byggnadens livslängd. För att minska dessa energiförluster och utsläppThe Swedish construction sector accounts för a significant share of energy use and greenhouse gas emissions. Ventilation systems are necessary for indoor comfort and air quality but also contribute to continuous energy losses during a building´s lifetime. Proper insulation of ventilation ducts is therefore important to reduce both energy use and climate impact. The purpose of this study was to inv
Fifth-Generation New Radio (5G NR) base stations produce detailed scheduling logs that record per-slot decisions, channel measurements, and Hybrid Automatic Repeat Request (HARQ) feedback at sub-millisecond granularity. Diagnosing anomalies in these logs currently requires domain experts to spend approximately one hour of manual inspection per file, and no standardised tooling exists for automated
Chirality is the property of an object that cannot be superimposed onto its mirror image. Chiral materials exhibit circular dichroism (CD) which can selectively populate one spin state over the other upon optical excitation, creating a spin-polarized carrier population. In chiral hybrid organic-inorganic perovskites, chirality is transferred from the organic spacer to the inorganic framework, maki
