Events: Student project presentation, Arkitektur, Bygg- och miljöteknik, Data- och informationsteknik, Energi och miljö, Kemi- och bioteknik, Matematiska vetenskaper, Material- och tillverkningsteknik, Mikroteknologi och nanovetenskap, Produkt- och produktionsutveckling, Rymd- och geovetenskap, Signaler och system, Sjöfart och marin teknik, Teknikens ekonomi och organisation, Fysik, Tillämpad IT, Tillämpad mekanik, Space, Earth and Environment, Architecture and Civil Engineering, Bioteknik, Electrical Engineering, Industrial and Materials Science, Informations- och kommunikationsteknik, Mechanics and Maritime Sciences events at Chalmers University of TechnologyWed, 12 May 2021 21:21:10 +0200 Sjösten, Complex Adaptive Systems, Physics<p>Online via Zoom</p><p>​​Title of Master&#39;s Thesis: &quot;Deep Learning for Nanofluidic Scattering Microscopy - Tracking and characterizing single biomolecules from microscopic images using deep convolutional neural networks&quot; Follow the presentation online Password: Python</p>​​<strong>Abstract</strong>:<div><span style="background-color:initial"><br /></span></div> <div><span style="background-color:initial">A novel technique for label-free, real-time characterization of single biomolecules called </span><span style="background-color:initial">Nanofluidic Scatter Microscopy (NSM) has recently been developed by a research group at Chalmers. We have created a machine learning (ML) framework consisting of deep convolutional neural networks such as U-nets, ResNets and YOLO in order to characterize single biomolecules through kymographs collected through NSM, as an alternative approach to a standard data analysis method (SA). As a laser irradiates visible light onto single biomolecules freely diffusing in solution inside nanofluidic channels, the biomolecule and the nanochannel scatter light coherently into the collection optics, such that the nanochannels improve the optical contrast of the imaged biomolecules by several orders of magnitude. A video of the total scattering intensity is then recorded with a high frame rate camera (capturing 200 fps) in order to capture the movement of the molecules as well as the optical contrast of the biomolecules with respect to the nanochannel. From the movement of one single biomolecule, it is possible to predict its diffusion constant, which can then be used to infer the hydrodynamic radius of the biomolecule. Additionally, the predicted optical contrast of one single biomolecule can in turn be used to infer its molecular weight. From the combination of hydrodynamic radius and molecular weight, information about the conformal state of single biomolecules can be inferred. In this thesis, we show that the ML approach yields results comparable to the SA which was developed independently of the ML technique for biomolecules in the weight span 66-669 kDa, and we also show that the ML technique is superior to the SA in other regards, such as computational speed and potential to characterize smaller molecules. The results of the data analysis performed with the ML framework will also make an appearance in the first paper on the NSM technique which has recently been sent in for publication.</span></div> Zhang and Kunyang Huang, MPEPO<p>Web seminar</p><p>​ Interaction analysis for converters in 5G power system at the -48VDC bus</p><div>​</div> <div>Conducted at Ericsson supervised by Xiaolong Yue.</div> <div><br /></div> <div><a href="">Join the seminar via Zoom.</a></div> <div>Password: Thesis</div> <div><br /></div> <div>Examiner: Torbjörn Thiringer, Dept of Electrical Engineering</div> <div>Opponent: Beibei Zheng</div> <div><br /></div> <h2 class="chalmersElement-H2">Abstract</h2> <div><br /></div> <div>The DC-DC converters, known as the dc voltage regulator, are widely applied to several power supply applications. Among these applications, in this thesis work, the research is conducted on the dc part of an in-serve 5G power system at Ericsson, where the dc part circuit mainly includes the front-end converters and PWM dc-dc converters. In a telecom system, especially in the 5G telecom system with cutting-edge technology, some critical performances, known as the high stability and anti-interference capability of the power supply, are required to guarantee a reliable telecom signal transmission. To meet these requirements, the dynamic analysis, mainly referring to the small-signal model analysis, needs to be adopted to analyze the interactions for converters.  <br /><br />According to the control schemes of converters, the thesis work is mainly distributed into two parts: the LLC resonant converter with voltage mode (VM) control and PWM dc-dc converters with peak current mode (PCM) control. For the LLC resonant converter, the extended describing function method is mainly employed to derive and establish the small-signal model. Moreover, the application of the VM control simplifies the closed-loop small-signal model establishment. For the PWM dc-dc converter, the circuiting averaging is an efficient method to derive the small-signal model. The PCM control in this converter, as the popular control scheme, consists of two loops: the inner current loop and outer voltage loop. <br /><br />In general, the research process of these two converters is formed almost in the same way. To be specific, the physical meanings behind the small-signal model are illustrated and used to explain the relationship between the Bode plots and equivalent circuits. Furthermore, the comparisons and mutual verification of analytical results from the mathematical model and the simulation results from the simulation model, as the principle line, run through the whole study. Eventually, the verified mathematical model and simulation model can serve as a tool to improve the performance of stability and anti-interference capability in the power system. <br /></div>'s thesis presentation<p>Online</p><p>​Douglas Molin: Effective quasiparallelogram laws on elliptic curves</p><p>​<br />Abstract: The quasiparallelogram law is a theorem governing in a quantitative way the group structure on elliptic curves. In short, the law rules that the so called height functions on the curve are almost quadratic forms. In this presentation, I will introduce the classical height functions and discuss effective quasiparallelogram laws on elliptic curves over number fields with short Weierstrass equations. This work is motivated by a problem about uniform estimates on nonsingular cubic curves.</p> <p>Supervisor: Per Salberger</p>'s thesis presentation<p>Online</p><p>​Lorents Landgren: Optimal Graphs for Connectedness under Random Edge Deletion</p>​<br />Supervisor: Jeff Steif's thesis presentation<p>Online</p><p>​Niclas Johansson and Jonathan Gildevall: Automatic Emergency Detectionin Naval VHF Transmissions</p><p>​<br />Investigating the feasibility of self-supervised speech-to-text models for complex domains with large amounts of unlabelled data.</p> <p>Supervisor: Marina Axelson-Fisk</p>'s thesis presentation<p>Online</p><p>​Veronika Chronholm: Numerical Analysis of Yield Curves Implied by Two-Factor Interest Rate Models</p>​<br />Supervisor: Simone Calogero<br />Examiner: Moritz Schauer Damercheli, MPBME<p>Web seminar</p><p>​ Designing treatments for neuropathic pains based on sensorimotor training and brain modulation</p><div>​</div> <div><a href="">Join the seminar via Zoom.</a></div> <div>Password: 227927</div> <div><br /></div> <div>Examiner: Max Jair Ortiz Catalan, Dept of Electrical Engineering</div> <div><br /></div> <div><h2 class="chalmersElement-H2">Abstract</h2></div> <div><br /></div> <div>Neuropathic pain is a complex medical condition that is often chronic and difficult to treat. As a consequence of having chronic neuropathic pain, the patient’s quality of life is negatively affected. Although, various types of treatment have been tested, no standard clinical solution has yet been found for neuropathic pain, and its underlying mechanism has not been fully understood. Here, different methods of treatment were designed for various target groups suffering from neuropathic pain with the purpose of pain reduction, as well as for the understanding of the underlying mechanisms causing neuropathic pain. <br />In this thesis, two new guided plasticity therapies were designed based on Phantom Motor Execution (PME) therapy for the treatment of phantom limb pain. First, a protocol for Mindful Sensorimotor Therapy (MSMT) was designed and implemented by adding sensory trainings to the PME. Primarily, MSMT was considered to be applied for alleviation of PLP in patients with amputation. This was then extended to cover individuals with disarticulation or nerve injuries. In addition and in order to improve the efficacy of the MSMT, non-invasive brain modulation techniques, specifically transcranial Direct Current Stimulation (tDCS), was added to MSMT. In summary, these novel therapies are expected to improve the treatment of neuropathic pain.  However, larger and more rigorous randomized control clinical trials could be highly valuable to better prove the efficacy of these therapies.<br /></div> thesis presentation, Jakob Andreasson and Carl Emme<p>Online</p><p></p><div><div><span style="background-color:initial">Jakob Andreasson och Carl Emme in the </span><span style="background-color:initial">Master programme Sustainable Energy Systems </span><span style="background-color:initial">present​their</span><span style="background-color:initial"></span><span style="background-color:initial"> master’s thesis work “</span><span style="background-color:initial">Carbon capture and utilization at LKAB</span><span style="background-color:initial">”, at the Department of Space, Earth and Environment.</span><br /></div> <div></div> <div><br /></div> <div>Supervisor: Elin Fahrman</div> <div>Examiner: Fredrik Normann</div> <div>Opponent: Adam Fasth &amp; Kumail Merchant​<span></span></div></div> <div><br /></div> thesis presentation<p>Online</p><p>​Lovisa Sedin: Predicting playlist curation behaviour</p>​<br />Supervisor: Rebecka Jörnsten<br />Examiner: Aila Särkkä Johansson and Karl-Fredrik Zingaropoli, MPSYS<p>Web seminar</p><p>​ Detection, semantic segmentation and generation of engine drive cycles using machine learning</p><div>​</div> <div> <a href="">Join the seminar via Zoom.</a></div> <div>Password: 570251<br /></div> <div><br /></div> <div>Examiner: Erik Agrell, Dept of Electrical Engineering</div> <div>Supervisor: Martin Nilsson, Volvo Penta</div> <div>Advisor: Maryam Lashgari, Dept of Electrical Engineering</div> <div>Opponents: Zacharias Hultman, Adrian Fahlström Svärd</div> <div><br /></div> <h2 class="chalmersElement-H2">Abstract</h2> <div><br /></div> <div>Data that are consequently stored within data lakes is extremely valuable to identify future abnormalities or behaviours for developing the products constantly. The data analysis can be done using supervised machine learning to perform predictions of user defined behaviours in time series. The goal of this thesis is to explore real engine data and use applied machine learning as a proof of concept for advanced analytics. The engines may vary in size, area of application, and operation. A general form of an engine operational cycle for a given area of application is referred to as an engine drive cycle. We have proposed three approaches to explore the potential of the engine data. These approaches utilize an Artifical Neural Network (ANN) as a classifier. The dataset is based on a sequentially windowed feature extraction of statistical, temporal, and spectral features. The first approach, called detection, shows the separability of the features between the engine size and applications with 98% accuracy. The second approach, called semantic segmentation, proved that by adopting a similar ANN architecture as used for the detection approach, we can provide an automated way to semantically classify manual annotated data. With an imbalanced dataset, the ANN obtained 98% accuracy of the classifications. The third approach, called generation of engine drive cycles, utilizes the ANN as a fitness function for the genetic algorithm, which illustrated the lack of robustness of the ANN. Only 90% accuracy was achieved because the number of feature representations was lower by using only one input signal to the ANN. A realistic drive cycle could not be generated, which was concluded to be due to the lack of feature representations in the ANN and few constructional constraints in the genetic algorithm during generation process.<br /><br /><br /><br /></div> thesis presentation<p>Online</p><p>​Rikard Martin and Tomas Lundberg: AI-Alignment with Evolutionary Algorithms and Artificial Life - a Concrete Approach to the Abstract Problem of Controlling Advanced AI</p>​<br />Supervisor: Torbjörn Lundh and evaluation of a test rig for measuring gear related losses within a Volvo Trucks gearbox<p>Online in Zoom</p><p>​Sebastian Eliasson and André Fransson present their Master thesis titled &quot;Design and evaluation of a test rig for measuring gear related losses within a Volvo Trucks gearbox&quot;.</p>​<div>Students: Sebastian Eliasson and André Fransson<br /></div> <div><br /></div> <div>Supervisors: Valery Chernoray, Lars Jacobsson, Henrik Hagerman and Maria Petersson</div> <div><br /></div> <div>Examiner: Valery Chernoray</div> <div><br /></div> <div>Opponents: Daniel Hård and Henrik Svensson</div> Mohammadi, Physics<p>Online via Zoom</p><p>​​Title of Master thesis: &quot;Protein-corona formation and structural changes of lipid nanoparticles investigated using QCM-D and waveguide microscopy&quot; Follow the presentation online Password 685385</p>​<strong>Abstract</strong>:<div><br /></div> <div><span style="background-color:initial">Lipid nanoparticles (LNPs) are being increasingly considered for use in drug delivery.However, understanding the interactions of LNPs with a biological environment is crucial for successful deliver of their cargo, such as for example mRNA, across cellular membranes. Upon exposure to a biological fluid it is believed that the surface of LNPs is spontaneously covered by a protein corona, a layer of adhered biomolecules, prior to being taken up by the targeted cells. A protein of interest in this context is Apolipoprotein-E (ApoE), both because it is known to be part of the protein corona and because it may control endocytic uptake via specific interactions with receptors in the cell membrane. Furthermore, ApoE-coated LNPs are expected to undergo a structural change in the acidic environment of the endosome during cargo release to the cytosol.</span><div> </div> <div>In this master thesis, we have studied protein-corona formation and the reaction of the LNPs covered/not covered with ApoE in an acid environment that mimics the endosomal environment using label-free surface-based bioanalytical tools such as quartz crystal microbalance with dissipation monitoring (QCM-D) and waveguide microscopy. QCM-D provides information based on ensemble averaging of the biological interactions taking place on the surface of the sensor, while waveguide microscopy provides both ensemble-averaged and information based on inspection of individual LNPs. Further, waveguide microscopy is capable of detecting both light scattering and fluorescence signals, which provides the possibility to observe both the labeled interior of the LNPs and binding of unlabeled protein.</div> <div> </div> <div>We found that ApoE bind to PEG-modified LNPs with a lag-time of tens of minutes unless mixed with bovine serum albumin (BSA). This is attributed to BSA-induced PEG shedding which promotes ApoE binding. Inspections using combined label-free and fluorescence based waveguide microscopy releveled a weaker than expected dependence between cargo fluorescence and scattering intensity, suggesting that the self-assembly process utilized for LNP fabrication may vary with LNP size. It is also concluded that further liquid handling must be improved to follow ApoE binding to LNPs using waveguide microscopy. </div></div> <div><br /></div> compatibility between machines and operations for aerospace engine products<p>Online</p><p>​Roberto Costanzo &amp; Vedanth Limbayyaswamimath present their master thesis project. Online presentation.</p>​<span style="background-color:initial">Program: Production Engineering</span><div>Department: Industrial and Materials Science</div> <div>Examiner: Kristina Wärmefjord</div> <div>Supervisors: Stefan Cedergren &amp; Sunney Fotedar</div> <div>Opponents: Rahul Gowda Bangalore Rudre Gowda &amp; Max Lundström</div> <div><br /></div> <div><strong>Abstract</strong></div> <div>In today’s globalized marketplace, businesses of all kinds experience a constant change in customer needs and demands. For a company to stay competitive, it is essential to have flexible production systems, which allow for rapid production changes and ensure that new products are effectively and quickly introduced into the factory. Before a new product, or a change in product specifications, is introduced into the production system, two main areas must be carefully evaluated: the identification of the critical features of the product and the categorization of the machines suitable for performing the machining operations required for that product.</div> <div>GKN Aerospace, as the whole aerospace industry, manufactures an enormous amount of product variants, with different design and mechanical specifications. On the other side, the company has a large number of resources to select from. As a result, when it comes to introducing new products into their production system, selecting the right resources is a complex operation. Hence, it is crucial to understand how product features and machine capabilities relate to each other and how this relationship can be used to effectively allocate resources. The purpose of the project is to examine a specific number of both products and machines for identifying the critical parameters that have a direct effect on resource allocation.</div> impact of optimizing construction fleet operations - A case study of productivity services at construction sites<p>Online</p><p>​Ina Wollmer &amp; Julia Ferreira Sandberg present their master thesis project. Online presentation.</p>​Examiner: Johan Malmqvist, IMS<div><br /><div><strong>​</strong><span style="background-color:initial"><strong>Abstract</strong></span></div> <div>Co-Pilot is an on-board display, supported by sensors and cloud services, that delivers real</div> <div>time data to construction equipment operators in order to support their daily work. The</div> <div>purpose of this thesis is to analyze how the changes in performance and user behavior on</div> <div>construction sites implied by Co-Pilot impact the CO2 emissions of the equipment. The</div> <div>product is developed by CPAC Systems AB, for Volvo CE. It is sold as a productivity</div> <div>service, increasing the efficiently of the site. Furthermore, it is believed that Co-Pilot</div> <div>increases the site sustainability by reducing the CO2 emissions.</div> <div>The research approach is focused on understanding how Co-Pilot contributes to sustainability.</div> <div>This is done by comparing user behavior and equipment performance before and</div> <div>after implementation of the product, and translate these differences into CO2 emissions.</div> <div>Three different use cases where Co-Pilot is currently used are analyzed and compared in</div> <div>order to receive diverse perspectives. The applied user study methodology is a quantitative</div> <div>approach supported by qualitative practices in order to help explain the findings.</div> <div>To conclude, the results are considered positive as more is produced per tonne emitted</div> <div>CO2 at a majority of the studied sites. The fuel consumption and all production</div> <div>measures included in the study decrease after implementation of Co-Pilot. In order to</div> <div>maximize the improvement on sustainability, the fuel consumption must be the most decreasing</div> <div>measure out of all production categories, which is the case for two out of three</div> <div>use cases. In order to improve Co-Pilot and its sustainability even further, development</div> <div>should focus on decreasing idling and queuing times and strive to make Co-Pilot even</div> <div>easier to use and understand in order to promote full utilization.</div></div> <div><br /></div> Element Modelling of Cell Swelling for Traction Batteries<p>Online</p><p>​Albin Knutsson &amp; Jonas Trombati present their master thesis project. Online presentation.</p>​Supervisors: N. Mysore, VCC and D. Carlstedt, IMS<div>Examiner: Leif Asp, IMS<br /><div>Opposition by Kristian Ivancic followed by an open discussion</div></div> Sun Simulation of Polymer Materials in Automotive Applications<p>Online</p><p>​Kristian Ivancic presents his master thesis project. Online presentation.</p>​S<span style="background-color:initial">upervisor: R. Gutkin, VCC</span><div>Examiner: Leif Asp, IMS<span style="background-color:initial"> </span></div> <div><span style="background-color:initial">Opposition by Albin Knutsson and Jonas Trombati followed by an open discussion</span></div>'s thesis presentation<p>Online</p><p>​Martin Gulliksson: A genetic programming approach to finding discrepancies in log files</p>​<br />Supervisor: Martin Raum Remanufacturing Barriers through Refining the Information flow at Volvo Cars<p>Online</p><p>​Oskar Kader &amp; Ida Nykvist present their master thesis project. Online presentation.</p><span style="background-color:initial">Program: MPPDE, MPPEN</span><div>Examinier: Mélanie Despeisse, IMS </div> <div>Supervisor: Mélanie Despeisse, IMS, Charlotta Nedström, Volvo Cars</div> <div>Opponent: Adarsh Vasudevan &amp; Jenny Duan</div> <div><br /></div> <div><div><span style="font-weight:700">Abstract</span></div> <div>Value creation in today's economy is heavily based on linear model, thus products become discarded at their end of life. The linear approach is a strong contributor to climate change and this is no exception for the automotive industry, which is currently facing new challenges to be more sustainable. In contrast to this model, circular economy represent a regenerative and restorative model, which differ in value creation and instead aims at decouple economic growth from finite resource consumption. This report will mainly address how remanufacturing can be emphasized to increase circularity. This is easier said than done. The study aims to provide knowledge for industrial processes that support Volvo Cars in designing electric passenger car components for remanufacturing. This will further provide action plan for the R\&amp;D unit, where in the process of remanufacturing Volvo Cars can improve to work towards a more circular mindset. More specifically, two questions will be answered: 1) What are the potential enablers and barriers to remanufacture the electric transmission component in a circular economy context? And what aspects to take into consideration with components when designing for remanufacturing? and 2) What are the challenges Volvo Cars faces with the current remanufacturing strategy for vehicles? The findings of the study are as following: First, internal and external enablers and barriers to remanufacture are identified and put into the context of Volvo Cars. Further, three main aspects to take into consideration when designing components for remanufacturing have been identified and been presented, which are technical, environmental and economical aspect. Secondly, implications are presented based on each problem category for Volvo Cars and possible improvement areas identified. Tertiary, the information between design and remanufacturing supplier was found to be the main barrier to overcome, thus a workshop was performed with the purpose of verifying this, as well as establish an action plan to overcome this.​</div></div> <div><br /></div>'s thesis presentation<p>Online</p><p>Sebastian Oleszko and Alexander Wölfinger: Detection of Evasive Maneuvers for Surrounding Vehicles</p>​<br />Supervisor: Konstantinos Konstantinou<br />Examiner: Aila Särkkä's thesis presentation<p>Online</p><p>​David Elinder: Investigating Abelian categories in univalent type theory</p>​<br />Supervisor: Martin Raum's thesis presentation<p>Online</p><p>​Sara Nordin Hällgren: Reading Numbers from Annual Reports</p>​Join the presentation via zoom: <a href=""></a>'s thesis presentation<p>Online</p><p>​Felix Augustsson: Symmetries in Biomathematical Models</p>​<br />Zoom link: <a href=""></a> Assurance in Composite Manufacturing - Identification of Variation Sources<p>Online</p><p>​Vilma Fernström presents her master thesis project. Online presentation.</p>​<span style="background-color:initial">Program: Materials Engineering</span><div>Examiner: Kristina Wärmefjord</div> <div>Supervisors: Kristina Wärmefjord, Johan Lööf</div>'s-programme-Naval-Architecture-and-Ocean-Engineering.aspx's-programme-Naval-Architecture-and-Ocean-Engineering.aspxPresentation of Master theses in the Master's programme Naval Architecture and Ocean Engineering<p>Online in Zoom</p><p>​The 4 June between 8:30 and 17:00 it is time for the annual presentations of Master theses by students on the international Master&#39;s programme Naval architecture and ocean engineering.</p>​<div><span style="background-color:initial">Every year, the international Master’s programme in Naval architecture and ocean engineering</span><div>at Chalmers University of Technology arranges a public seminar where the students’</div> <div>theses are presented. This year, due to COVID-19, the seminar is held online in Zoom.</div> <div><br /></div> <div>In<a href="/SiteCollectionDocuments/M2/Kalendarium/Save-the-date%20-%20Invitation%20to%20MSc%20thesis%20presentations%20at%20Chalmers%20210604.pdf" target="_blank"> the preliminary program​</a>​ you can see which of the projects could interest you.</div> <div><br /></div> <div>Each thesis is presented for 20 minutes followed by an opposition made by one/two students. The audience is invited after that to ask questions and discuss the work.</div> <div><br /></div> <div><div>Welcome to attend the seminar! <span style="background-color:initial">​</span></div></div></div> of Misconstraining during Manual Assembly at Volvo Cars<p>Online</p><p>​​Erik Arvidsson &amp; Vasanth Kumar present their master theis project. Online presentation.</p>​<span style="background-color:initial">Program: MSc Product Development - MPPDE </span><div>Examiner: Kristina Wärmefjord, Department of Industrial and Material Science (IMS) </div> <div>Supervisor(s): Kristina Wärmefjord, Department of Industrial and Materials Science (IMS), (David Brinkby, Volvo Car Corporation (VCC))</div> <div>Opponent: Philip Blom &amp; Felix Englund </div> <div><br /></div> <div><div><strong>Abstract</strong></div> <div>The rising need for increased robustness among products and processes worldwide has made it essential to research areas around geometrical variation and robust design. The presence of humans in the assembly process is inevitable due to the high levels of flexibility they can exhibit. However, the ways humans contribute to geometrical variation in a robust design perspective have not been significantly studied. As an attempt to break this novelty barrier, this thesis focuses on investigating and examining the concept of misconstraining in the context of an error phenomenon at Volvo Cars through a series of case studies, semi-structured interviews, and tests. Following a standard product development process, the main aim of the thesis is to provide proof of concept and develop a framework for analyzing misconstraining from a manual assembly perspective. The test cases present in the thesis have been designed by keeping in mind the parameters involved in the manual assembly process and the surrounding conditions. The results have been corroborated by analyzing graphical plots and extensive evaluation of the various case studies.  </div> <div>Ultimately, by utilizing all the information brought to light, the thesis resulted in developing a detailed mapping of the different causes and effects of misconstraining. Furthermore, the tests conducted revealed how the play between the guiding elements leads to geometrical variation in connection to human error contribution and how the change in locating scheme expedites misconstraining. The thesis concludes with a list of recommendations to tackle misconstraining and the envisioned forms of misconstraining. Also, a solid amount of information vital for future research has been presented with some exciting visualization about the concept moving forward. </div></div> <div><br /></div> of Low-Carbon Hydrogen and Small Modular Reactors<p>Online</p><p>​Ludvig Blomgren &amp; Balder Hagert present their master thesis project :Opportunities of Low-Carbon Hydrogen and Small Modular Reactors: Are small modular reactors a competitive way of producing hydrogen in the future hydrogen economy? Online presentation. ​</p>Supervisor: Johan Malmqvist, IMS​<div><strong><br /></strong></div> <div><strong>Abstract</strong><div>Hydrogen is being proposed as the fuel of the future and to be used for many applications in industry and transportation. Producing low-carbon hydrogen requires a lot of low-carbon energy. The study has investigated what applications hydrogen will most likely have a prominent role in, and what requirements different demands put on the supply. The development of small modular nuclear reactors (SMR) has been investigated as well as their potential as an energy source in hydrogen production. The research approach of the study consisted of a literature study, followed by a qualitative study. The qualitative study consisted of interviews with 12 experts in the areas of hydrogen applications and nuclear power. The study also included calculations deciding the delivered cost of hydrogen. The study found that there are applications where the demand for hydrogen is certain, including in the production of ammonia and methanol. Applications of hydrogen such as producing steel, as a fuel for trucks, busses, and ships are also found to be probable growing demands for hydrogen in the future. The study found that SMRs have many synergies with hydrogen production, including process heat, continuous production, less location-specific, and modularity. Finally, the results of the study include calculations of the delivered cost of hydrogen, proving small modular reactors as a competitive alternative under certain assumptions. To summarize, the study has found that there is a need for low-carbon hydrogen in the future and that it is important to consider the delivered cost of hydrogen, which includes production, transportation, and storage cost. It was found that it is important to be technology-neutral when considering energy sources for producing hydrogen and that nuclear and SMRs should be included as an option.</div></div> Nordanger and Christina Chau, MPBME<p>Web seminar</p><p>​ Guiding AI based classification: Can conventional functional neuroimaging analysis improve deep-learning methods for identifying risk for essential hypertension?</p><div>​</div> <div><a href="">Join the seminar via Zoom.</a></div> <div>Password: 215706</div> <div><br /></div> <div>Examiner: Paolo Monti, Dept of Electrical Engineering</div> <div>Supervisors: Alice Deimante Neimantaite/Justi Schneiderman, Syntronic/Medtech West, Mohammad Hossein Modhaddam, Dept of Electrical Engineering</div> <div>Opponent: Emelie Carlberg</div> <div><br /></div> <h2 class="chalmersElement-H2">Abstract</h2> <div><br /></div> <div>Hypertension is a highly prevalent disease and a major risk factor for cardiovascular diseases. Therefore, it is important to find methods to identify the risk of developing hypertension at an early stage in order to work in a preventive manner. It has been shown that the invasively-measured muscle sympathetic nerve activity (MSNA) response to stressful stimuli is likely to predict this risk. Furthermore, a strong correlation between MSNA and beta oscillations in the brain has recently been identified. With these known correlations, the neural activity in the beta band could potentially be used as a non-invasive biomarker recorded with magnetoencephalography (MEG) for evaluating the risk of developing hypertension.<br />The aim of this thesis is to investigate how the learning and classification of two previously developed neural networks can be improved with conventional MEG analysis methods. We attempted to give emphasis to the beta oscillations around 12-30 Hz by filtering out frequencies outside this beta band. Multiclass classification was also tested on different groupings of subjects in the MEG study based on the level of risk, i.e. high risk, medium risk etc. The resulting accuracies were generally lower for the filtered data, indicating that the filtering had removed some aspect in the data that the network had previously learned form, perhaps some characteristics of the MSNA-related brain response profile. Accuracies improved for multiclass classification but were still deemed to be essentially equivalent to previous results. Thus, pure frequency-based filtering does not appear to achieve the desired effect, which could be because the AI methods used are not designed to identify frequency-based changes in power. Dividing the subjects into multiple classes also failed to improve classification, which is likely due to the limitation of having a low number of individuals in the MEG study. We conclude with a framework for potentially overcoming such challenges by training the networks on the data after transforming it into the time-frequency domain.<br /></div>'s thesis presentation<p>Online</p><p>​David Lidström: Agent based match racing simulations</p>Supervisor: Torbjörn Lundh Daily Management of an Efficient Organization<p>Online</p><p>​Axel Magnusson &amp; Quintus Westberg present their master thesis project. Online presentation.</p>​<span style="background-color:initial">Program: </span><span style="background-color:initial">PRODUCTION ENGINEERING (MPPEN)</span><div>Examiner/Supervisor: <span style="background-color:initial">Torbjörn Ylipää, IMS</span></div> <div><span style="background-color:initial">​</span><br /></div>'s thesis presentation<p>Online</p><p>​Miriam Skarin: Monte Carlo simulations for Brachytherapy</p>​<br />Supervisor: Torbjörn Lundh's thesis presentation<p>Online</p><p>​Oscar Forsman and Johanna Warnqvist: Adaptive Driver Modelling for Collision Warning Systems</p><br />Supervisor: Torbjörn Lundh Based Modelling of Roller Bearings<p>Online</p><p>​Ramin Rahani presents his master thesis project. Online presentation.</p>​<span style="background-color:initial">Master program: Applied Mechanics</span><div>Examiner: Fredrik Larsson</div> <div>Supervisor: Sven Norberg</div> <div>Opponents: Pouya Sheikholeslami and Yousef Amiri</div> ​'s thesis presentation<p>Online</p><p>Erik von Brömssen: ​Computing persistent homology in parallel with a functional language</p>​<br />Supervisor: Martin Raum's thesis presentation<p>Online</p><p>​Marcus Martinsson and Per Nordeman: An automated test framework for vascular surgery simulations</p>​<br />Supervisor: Torbjörn Lundh's thesis presentation<p>Online</p><p>​Harald Westling: Deep-learning-accelerated Bayesian inference for FRAP experiments</p><br />Supervisor: Magnus Röding<br />Examiner: Aila Särkkä thesis presentation: High-Temperature Behaviour of 850 nm VCSELs<p>Online</p><p>​Welcome to a master thesis presentation with Hans Kaimre, MPWPSSupervisor: Alexander Grabowski, Anders Larsson Examiner: Johan Gustavsson Join from PC, Mac, Linux, iOS or Android: Password: kaimremsc</p><strong>​Abstract:</strong> <span style="background-color:initial">TBA</span><span style="background-color:initial">​</span> Prakash, MPBME<p>Web seminar</p><p>​ Autonomous calibration of fundus camera</p><div>​</div> <div><a href="">Join the seminar via Zoom.</a></div> <div><br /></div> <div>Examiner: Carl Olsson, Dept of Electrical Engineering</div> <div>Opponent: Melanie Torres</div> <div><br /></div> <h2 class="chalmersElement-H2">Abstract</h2> <div><br /></div> <div>Diabetic retinopathy (DR) is one of the leading causes of blindness in the world, affecting up to 80% of people who have had diabetes for 20 years or more. It is usually asymptomatic until it has progressed to advanced stages and affects the patient's vision. To prevent DR, all diabetics are advised to get their eyes screened once a year. However, only 15% follow-up on this advice, largely due to lack of access to eye specialists and screening equipment.<br />This thesis considers calibration of a given system consisting of three cameras mounted on a (6-axis) Stewart platform. Two side cameras are used for stereo imaging to locate how far the platform is from the pupil. After moving the platform to the required location, IR leds inside the middle camera are turned on to help the main fundus camera take image of the eye.<br />Calibration and alignment are necessary to take proper fundus images using 3 cameras on the device. The thesis goes into detail on how to calibrate the cameras autonomously to capture high quality images of the eye.A special calibration tool consisting of planar regions with aruco markers is designed and used for estimation of both intrinsic and extrinsic camera properties. A process is defined to move the platform to capture images of the different planes on the calibration tool. These are used to estimate intrinsic and extrinsic movement properties. Further, a pixel calibration procedure is described to find the best distance of the camera from the calibration tool to take high quality images of the eye. The calibration method accuracy was certified by making sure validation error stays below 0.2 mm.Finally, It was tested on a patient in a clinical setting to verify high quality images were captured due to an accurately obtained estimation of the patient eye location.<br /></div>'s thesis presentation<p>Online</p><p>​Johan Gustafsson: TBA</p>​<br />Supervisor: Magnus Röding