SINTEF Open
Not a member yet
    11374 research outputs found

    Pitting Detection and Characterization From Ultrasound Timelapse Images Using Convolutional Neural Networks

    Get PDF
    Pitting corrosion, a localized form of corrosion leading to cavities and structural failure in metallic materials, requires early detection for effective mitigation. While ultrasonic inspection techniques can readily detect uniform wall thinning, they often struggle to identify pitting corrosion. This study proposes a time-lapse ultrasound inspection method to detect early-stage pitting using pulse-echo sensors. By recording multiple ultrasonic traces over time, 2-D timelapse images of ultrasonic reflectivity can be generated and fed into a trained neural network for pitting diagnostics. In general, training a machine-learning model requires a large training dataset. This work used data from a drilling experiment to generate a suitable dataset. Dataset construction by random time-ordered combinations of ultrasonic measurements was conducted to create a diverse set of time-lapse image samples to generalize the resulting machine-learning model adequately. A classification neural network was trained to detect the presence of drilled holes, and a separate regression network was trained to estimate the hole depth. Based on drilling data from an independently acquired test dataset, results demonstrate a mean absolute error of 0.163 mm for hole depth estimations. All holes are successfully detected when 0.1 mm deeper than the defined pitting threshold of 0.5 mm. This suggests that the proposed method generalizes well and can be deployed to any similar acquisition system.publishedVersio

    Practical and Ethical Considerations for Generative AI in Medical Imaging

    Get PDF
    Generative Artificial Intelligence (AI) has the potential to transform medicine. It is helpful to clinicians and radiologists for diagnosis, screening, treatment planning, interventions, and drug development. It benefits the clinical flow with real-time decision-support systems. While generative AI can potentially improve healthcare, it also introduces new ethical issues that require careful analysis and mitigation strategies. This work emphasizes the ethical aspects of generative AI in medical imaging, aiming to ensure that advancements in this field align with established ethical principles and societal values. We delve into the ethical implications surrounding bias, fairness, patient privacy, consent, transparency, explainability, intellectual property, and data ownership. Furthermore, we discuss regulations governing the use of synthetic medical data. To promote equitable application of these powerful tools, we also propose clear guidelines for promoting fairness, mitigating bias, and ensuring diversity within generative AI models.acceptedVersio

    State-of-the-art report – NLFEA of deteriorated concrete structures

    Get PDF
    State-of-the-art report for performing nonlinear finite element analysis to assess the structural integrity of deteriorated reinforced concrete structures. In accordance with the scope and objects set out in work package H3.5.publishedVersio

    Helsehjelp til barn i barnevernet - Behov, barrierer og helsetjenestebruk

    Get PDF
    Hovedmålet med rapporten er å undersøke om barn og unge som enten er i barnevernsinstitusjon, i fosterhjem eller som får tiltak i hjemmet, får den helsehjelpen de trenger. Dette undersøkes i tre delprosjekter. I rapportens del 1 undersøker vi behov for, tilgang til, og erfaringer med helsehjelp - slik det oppleves av barn og unge samt foreldre og fosterforeldre. Datagrunnlaget er kvalitative intervjuer og en survey. I del 2 undersøkes barrierer for helsehjelp, og hva som bidrar til helsehjelp. Datagrunnlaget er en studie av saksmapper i fire kommuner og fire institusjoner, og intervjuer med saksbehandlere og ledere i barnevern, helseansvarlige i barnevernsinstitusjoner og fagpersoner i samarbeidende tjenester utenfor barnevern. Del 3 undersøker helsetjenestebruk hos barn i barnevern sammenlignet med barne- og ungdomsbefolkningen generelt ved hjelp av en kobling av data fra ulike registre. I del 4 gis en oppsummerende diskusjon av samtlige problemstillinger, og anbefalinger for veien videre.Helsehjelp til barn i barnevernet - Behov, barrierer og helsetjenestebrukISBN: 978-82-14-07239-6publishedVersio

    Nærnettverksmetodikken. Systematisk arbeid med nærnettverket til yngre personer med demenssykdom

    Get PDF
    Prosjektet Nærnettverksmetodikk for yngre personer med demenssykdom har hatt som mål å utvikle og teste ut verktøy for å skape bedre vilkår for målgruppen gjennom å fokusere på relasjoner i pasientens nettverk. Nærnettverksmetodikken er utviklet i samarbeid med Utviklingssenter for sykehjem og hjemmetjenester Trøndelag og åtte kommuner. Metodikken retter seg mot yngre personer med demenssykdom, og er utviklet som et verktøy for kommunehelsetjenesten. Formålet er å utvikle systematikk i kommunikasjonen mellom kommune, pasient og pårørende, noe som er etterlyst fra kommunene. Det er gjennomført enkelte uttestinger som viser at verktøyene kan bidra til positive virkninger for pasient og pårørende på kort sikt.Nærnettverksmetodikken. Systematisk arbeid med nærnettverket til yngre personer med demenssykdomISBN: 978-82-14-07064-4publishedVersio

    Morgendagens HMS-opplæring for verneombudene i norsk arbeidsliv. Verneombudenes erfaringer med dagens opplæring og innspill til fremtidens HMS-opplæring

    Get PDF
    Denne rapporten har som formål å kartlegge verneombudenes erfaringer med dagens HMS-opplæring og identifisere behov for fremtidig opplæring i rollen. Studien adresserer hvordan opplæringen kan forbedres for å møte utfordringer i et arbeidsliv preget av endringer i lovkrav og arbeidsmiljøforhold. Basert på en omfattende spørreundersøkelse med 2247 verneombud fra ulike sektorer og bransjer, fremkommer det at dagens HMS-opplæring i hovedsak vurderes positivt, men med rom for forbedringer. Fysiske, dialogbaserte og bransjespesifikke kurs med lengre varighet gir spesielt høyt læringsutbytte, tilfredshet og opplevd relevans, og anbefales prioritert. Samtidig foreslås hybride, modulbaserte og fleksible kursmodeller for å øke tilgjengelighet og tilpasning til ulike behov. Studien peker også på behovet for styrket opplæring innen endringsprosesser og psykososialt arbeidsmiljø, samt systematisk videreopplæring og kvalitetssikring av kursinnhold for å sikre at verneombudene har nødvendig kompetanse til å møte morgendagens krav i arbeidslivet.Morgendagens HMS-opplæring for verneombudene i norsk arbeidsliv. Verneombudenes erfaringer med dagens opplæring og innspill til fremtidens HMS-opplæringISBN: 978-82-14-07091-0publishedVersio

    Årsrapport for 2023 - SINTEF Digital - Senter for Jernbanesertifisering

    Get PDF
    Denne rapporten beskriver SINTEF Digital sin virksomhet som teknisk kontrollorgan for 2023. SINTEF, ved Senter for jernbanesertifisering – SJS, ble 16. desember 2003 teknisk kontrollorgan (Notified Body - NoBo) under EU direktiv 96/48/EC (samtrafikkevnen i det europeiske jernbanesystem for høyhastighetstog). Tilsvarende ble SINTEF i 2005 utpekt teknisk kontrollorgan under EU direktiv 2001/16/EF (samtrafikkevnen i det transeuropeiske konvensjonelle jernbanesystemet), samt 2008/57/EC i 2010 (samtrafikkevnen i jernbanesystemet). 17. februar 2021 ble SINTEF, ved SJS, utpekt som teknisk kontrollorgan under EU direktiv (EU) 2016/797 (samtrafikkevnen i jernbanesystemet). Utpekingen dekker delsystem styring, kontroll og signal. SINTEF, ved SJS, er også utpekt organ (DeBo - Designated Body) for delsystem styring, kontroll og signal i Norge, Sverige og Finland, og akkreditert som inspeksjonsorgan Type C (ISA – Independent Safety Assessor, AsBo - Assessment Body og ICA – Independent Cybersecurity Assessor) i henhold til ISO/IEC 17020:2012. Gjennom vår deltagelse i NB-Rail er vi til stede der de faglige anbefalingene blir utarbeidet og akseptert for rollen som teknisk kontrollorgan.Årsrapport for 2023 - SINTEF Digital - Senter for JernbanesertifiseringISBN: 978-82-14-07241-9publishedVersio

    Fracture network characterization through fractal dimension and Gutenberg–Richter parameter: Decatur open-source dataset as a study case

    Get PDF
    The fractal formalisms are well known for providing new understandings regarding the geometrical, spatial, and temporal behaviour of seismicity. Particularly, the fractal dimensions give information about the seismic events self-organization and self-similarity. On the other hand, the Gutenberg–Richter value, known as the b-value, has shown through the years to give handy information regarding the statistical distribution of earthquakes, on-site physical parameters, and geomechanical inputs. The Gutenberg–Richter value (b) and the capacity and correlation fractal dimensions, (D0 and D2), of the spatial distribution of earthquake hypocentres interact mathematically for micro- and macro-events. From this interaction, it is possible to obtain new insights into the fracture network development and the microseismicity source characterization in terms of single fractures, fault planes, or densely fractured volumetric spaces. Here we show this interaction for the open-source Decatur CO2 project seismicity catalogue, comparing it with the results obtained for a natural earthquake catalogue of Illinois, in the United States. The fractal dimension D0 is calculated using two different methodologies: box-counting and correlation integral partitioning. This last method is also used to calculate D2. The results presented in this study allow us to describe how the fracture network geometry influences the earthquake complexity. Together with the calculation of the b-value, we present clear indications which show that seismicity recorded in the Illinois tectonic environment partially follows the Aki relationship D0 ∼ 2b, which is not the case for induced events. In addition, the induced earthquake dataset shows that D2 > D0, an anomalous behaviour in terms of the fractal formalisms. All these facts might be used to establish spatial fracture network control techniques and seismicity-type distinctions in CO2 injection sites located in highly active tectonic areas, respectively.publishedVersio

    Sizing and optimization of a cold thermal energy storage (CTES) for a dairy: A case study

    Get PDF
    In the food industry, particularly in the dairy industry, significant amounts of thermal energy are required on different temperature levels. The thermal demand is linked to the production schedule and varies during the day. Hence, solutions are required to avoid part load operation and high peak loads on the refrigeration system. The subject of this case study is a dairy in central Norway, where combined CO2 heat pumps/chillers are used to cover the main heating and cooling loads. This study focuses on improving the current operation of the CO2 system by implementing a CTES with a phase change material. A Modelica model was used as a tool to investigate the process integration and evaluate the impact of the CTES on peak load reduction and a secured supply of process cooling load. The results show that the CTES can successfully smooth the provision of cold process water cooling, as supply and demand are decoupled from each other and sufficient process cooling can be provided at the same time.acceptedVersio

    Improving Flow Balancing: Employing GNF-X for Predicting Flow Profile

    Get PDF
    Flow balancing is the process of changing the geometry of an extrusion die to achieve a uniform flow distribution at the die-exit. This can be an iterative and costly process. Computational Fluid Dynamics (CFD) is often used to aid in flow balancing of extrusion dies by predicting the flow profile without the need for experiments. The numerical models used in literature simplify the behavior significantly by using a Generalized Newtonian Fluid (GNF) approach. This study investigates the effects of applying the recently introduced GNF-eXtended (GNF-X) model to the flow balancing problem for molten polymers. This aims to include the effect of extensional flow on the materials apparent viscosity and increase the accuracy of numerical predictions. The model was implemented into ANSYS CFX and the effects on the flow balance of an extrusion die evaluated. A substantial impact was observed on the flow distribution at the die-exit when using the GNFX model instead of the GNF version. The prediction of the flow balance with the GNF-X model differs by a factor of up to two with the GNF model, underlining the importance of including extensional flow behavior.publishedVersio

    9,869

    full texts

    11,374

    metadata records
    Updated in last 30 days.
    SINTEF Open
    Access Repository Dashboard
    Do you manage Open Research Online? Become a CORE Member to access insider analytics, issue reports and manage access to outputs from your repository in the CORE Repository Dashboard! 👇