The thesis focuses on a design of visualization and setting limits of important parametres of power and own-consumption transformers at the nuclear power plant Dukovany. In the first part there is a description of a present technical state of oil power transformers at this power plant, including a description of a currently installed transformers monitoring system and electro monitoring system. The second part deals with a design of a visialization of parametrs and a diagram design of active-access displays of monitored parametrs of these transformers. In the thesis there is also a description of web interface with a new visualization. The last part of the thesis contains a design of setting the limits and criteria of important monitoring parametres of these transformers.

The thesis deals with possibilities of numerical modelling of an overflow over selected types of auxiliary spillways and then the possibilities are verified in particular localities. In thesis, there is carried out the determination of capacity of selected spillways by using several methods: the calculation of the overflow equation, one-dimensional and two-dimensional numerical models and physical model. The author created parametric analysis to each single method. Parametric analysis represent various hydraulic conditions. Subsequently, there is possible to express the effect of parametric analysis. The author makes a comparison of water levels received by mentioned methods set out under the specific hydraulic conditions by calibration of each model with the physical model. Finally, author recommends values of all coefficients which have an impact on results of numerical modelling of the flow over selected types of weirs mostly using like auxiliary spillways.

MCur / Department of Advanced Nursing Science

Recurrent Neural Networks(RNNs) are powerful models that have obtained outstanding achievements in many sequence learning tasks. Despite their accomplishments, RNN models still suffer with long sequences during training. It is because error propagate backwards from output to input layers carrying gradient signals, and with long input sequence, issues like vanishing and exploding gradients can arise. This thesis reviews many current studies and existing architectures designed to circumvent the long-term dependency problems in backpropagation through time (BPTT). Mainly, we focus on the method proposed by Trinh et al. (2018) which uses semi- supervised learning method to alleviate the long-term dependency problems in BPTT. Despite the good results Trinh et al. (2018)’s model achieved, we suggest that the model can be further improved with a more systematic way of balancing auxiliary signals. In this thesis, we present our paper – RNNs with Private and Shared Representations for Semi-Supervised Learning – which is currently under review for AAAI-2019. We propose a semi-supervised RNN architecture with explicitly designed private and shared representations that regulates the gradient flow from auxiliary task to main task. / Les réseaux neuronaux récurrents (RNN) sont des modèles puissants qui ont obtenu des réalisations exceptionnelles dans de nombreuses tâches d’apprentissage séquentiel. Malgré leurs réalisations, les modèles RNN sou˙rent encore de longues séquences pendant l’entraî-nement. C’est parce que l’erreur se propage en arrière de la sortie vers les couches d’entrée transportant des signaux de gradient, et avec une longue séquence d’entrée, des problèmes comme la disparition et l’explosion des gradients peuvent survenir. Cette thèse passe en revue de nombreuses études actuelles et architectures existantes conçues pour contour-ner les problèmes de dépendance à long terme de la rétropropagation dans le temps (BPTT). Nous nous concentrons principalement sur la méthode proposée par cite Trinh2018 qui utilise une méthode d’apprentissage semi-supervisée pour atténuer les problèmes de dépendance à long terme dans BPTT. Malgré les bons résultats obtenus avec le modèle de cite Trinh2018, nous suggérons que le modèle peut être encore amélioré avec une manière plus systématique d’équilibrer les signaux auxiliaires. Dans cette thèse, nous présentons notre article - emph RNNs with Private and Shared Representations for Semi-Supervised Learning - qui est actuellement en cours de révision pour AAAI-2019. Nous propo-sons une architecture RNN semi-supervisée avec des représentations privées et partagées explicitement conçues qui régule le flux de gradient de la tâche auxiliaire à la tâche principale.

<p>The purpose of this research was to determine what components of a complex centrifugal  compression system inlet needed to be modelled to accurately predict the swirl and total pressure  distortions at the compressor face. Two computational models were developed. A full-fidelity case  where all the inlet geometry was modelled and a reduced model where a small portion of the inlet  was considered. Both the numerical cases were compared with experimental data from a research  compressor rig developed by Honeywell Aerospace. The test apparatus was designed with a  modular inlet system to develop swirl distortion patterns. The modular inlet system utilized  transposable baffles within the radial-to-axial section of the inlet and blockage plates of varying  sizes and geometries at the inlet to this section.  Discerning the dominant inlet component that dictates distortion behavior at the compressor  face would allow the reduced modelling of inlet components for compression systems and would  allow coupling with more tortuous systems. Furthermore, it would reduce the design iteration and  simulation time of the inlet systems. Several investigations utilizing a reduced model only  considering a radial-to-axial inlet are available in literature, but no comprehensive justification has  been presented as to the impact this has on the distortion behavior.   Experimental surveys of flow conditions just upstream of the inducer of the centrifugal  compressor were conducted at several operating conditions. The highest and lowest mass flow  rates of these operating points were simulated using ANSYS CFX 2020R1 for both the  computational models. Multiple inlet configurations were simulated to test the robustness of the  reduced model in comparison to the full fidelity. The numerical simulations highlighted  shortcomings of the instrumentation used to characterize the experimental flow field at the inducer,  particularly with respect to total pressure distortion. Furthermore, transient pressure data were  measured in experiment and indicated unsteady fluctuations in the inlet that would not be captured  by steady computational fluid dynamic simulations. These data matched locations of disagreement  with swirl distortion behavior at high mass flow rates. This suggested that transient vortex  movement occured at the aerodynamic interface plane in certain configurations.   The total pressure distortion metrics between the two models were remarkably comparable.  Furthermore, the simplified model accurately predicted the mixing losses associated with the  blockage plates at the inlet to the radial-to-axial inlet using a simple inlet extension. Swirl  18  distortion was dictated by the radial-to-axial inlet. The reduced model data trends were comparable  with experiment for both the baffle and blocker plate configurations. The swirl intensities for all  configurations were comparable between the two models. The reduced model swirl directivity  trends matched those of experiment. The most notable deviations between the full-fidelity model  and the reduced model were observed with swirl directivity numerics. </p>

No description available.

The subject of the dissertation is the origin of the Old Norse periphrastic future. First, it is shown that the main exponent of future in Old Norse, as opposed to other old Germanic languages, is the auxiliary verb munu, both in terms of its frequency and its remarkably neutral meaning. It is attested as such even in the oldest 9th century sources, the Skaldic poems. These characteristics of munu are in contrast with the other exponent of the periphrastic future, the verb skulu, which was less frequent and more modal. The futural character of munu and the modal character of skulu are further demonstrated by an analysis of the means of future expression used in prophecies and curses. Discussion of the grammaticalization of munu has shown that as early as the 9th century the verb had developed meanings that are derived from future (particularly the probability meaning). Hence, the futural use must have been fairly common for a considerably long period before the first attestations. Further evidence for the early use of munu-future in North Germanic comes from the reconstruction of the morphological development of the verb. It shows a transition from the class of regular weak verbs to the preterite-present class which was associated with modal/function verbs in Germanic. These findings show that the...

This thesis report is about designing a prototype and establishing audio communication between a computer and Digital Signal Processor (DSP) using two preamp circuits using both auxiliary and USB connection. The paper gives the reader an overview on how audio is transmitted from a computer, through the system and to the desired output. The reader should also get a better understanding of how an AD-converter samples the incoming signal to the Discrete plane and how an AUX or phone connector works. This information can be used for designing preamp circuits to communicate between a computer and the DSP. The DSP circuit uses an STM32 processor to control the incoming and outgoing signals with the use of ADC and DAC conversion. The DSP also uses microphones to capture surrounding sound. An addition is to make a prototype on how to use these microphones to  send the signal upstream to the computer. The microphone is then benchmarked with the use of Matlab, calculating Total Harmonic Distortion. Management and optimization of code structure and resources is done in the source files of the project. Using imperative C programing, large functions are broken down into smaller functions to ease readability and control flow. The result is a prototype circuit that can communicate audio signals with both audio jack and USB between computers input and output to the DSP. Using CubeMX in conjunction with CubeIDE to add additional ADC channels to be able to incorporate an automatic source control when the audio jack or USB is connected. / Denna avhandling handlar om att designa en prototyp och etablera ljudkommunikation mellan en dator och en digital signalprocessor (DSP) med hjälp av två förstärkarkretsar genom både aux- och USB-anslutning. Rapporten ger läsaren en översikt över hur ljud skickas från en dator, genom systemet och till önskad utgång. Läsaren bör också få en bättre förståelse för hur en AD-omvandlare fungerar för att sampla den inkommande signalen till tidsdiskreta planet och hur en AUX- eller telekabel fungerar. Målet är att använda denna information för att skapa förstärkare som kan kommunicera mellan en dator och DSP:en. DSP-kretsen använder en STM32-processor för att hantera de inkommande och utgående signalerna med användning av ADC- och DAC-omvandling. DSP:en använder också mikrofoner för att fånga omgivande ljud. Ett tillägg är att skapa en prototyp för hur man kan använda en av mikrofonerna för att skicka signalen till datorn. Mikrofonen jämförs sedan med hjälp av Matlab genom att beräkna total harmonisk distorsion. Hantering och optimering av kodstruktur och resurser görs i projektets källkodsfiler. Genom att använda imperativ C-programmering bryts stora funktioner ned i mindre funktioner för att underlätta läsbarheten och styrningen av flödet. Resultatet är en prototypkrets som kan kommunicera ljudsignaler med både ljuduttag och USB mellan datorns in- och utgång och DSP:en. Genom att använda CubeMX tillsammans med CubeIDE läggs ytterligare ADC-kanaler för att möjliggöra automatisk källkontroll när ljuduttaget eller USB-anslutningen är ansluten.

The master's thesis consists of three parts: technological, design and research. In the technological part, the improvement of the technological process for the manufacture of basic parts in modern conditions using automated production is considered. In the design part, the development of a mechanized mounting and clamping device for a vertically milling operation and a control and measuring device for a control operation are carried out. In the research part, optimization of cutting conditions for milling a flat surface of the body with milling cutters of various diameters of domestic and foreign manufacturers is given. This part contains a comparison of the milling resistance, determination of the number of workpieces and a comparison of the cost of cutting tools of two manufacturers. The necessary set of design and technological documentation has been formed. / Магистерская диссертация состоит из трех частей: технологической, конструкторской и научно-исследовательской. В технологической части рассматривается совершенствование технологического процесса на изготовление корпусной детали в современных условиях с использованием автоматизированного производства. В конструкторской части производится разработка механизированного установочно-зажимного приспособления на вертикально-фрезерную операцию и контрольно-измерительного приспособления на контрольную операцию. В научно-исследовательской части приведена оптимизация режимов резания для фрезерования плоской поверхности корпуса фрезами различного диаметра отечественного и зарубежного производителей. Данная часть содержит сравнение стойкости фрез, определение количества обрабатываемых деталей и сравнение стоимости режущих инструментов двух производителей. Сформирован необходимый комплект конструкторской и технологической документации.

This study delves into the application of Small Area Estimation (SAE) techniques to enhance the accuracy of predicting global needs for assistive spectacles. By leveraging the power of SAE, the research undertakes a comprehensive exploration, employing arange of predictive models including Linear Regression (LR), Empirical Best Linear Unbiased Prediction (EBLUP), hglm (from R package) with Conditional Autoregressive (CAR), and Generalized Linear Mixed Models (GLMM). At last phase,the global spectacle needs’ prediction includes various essential steps such as random effects simulation, coefficient extraction from GLMM estimates, and log-linear modeling. The investigation develops a multi-faceted approach, incorporating area-level modeling, spatial correlation analysis, and relative standard error, to assess their impact on predictive accuracy. The GLMM consistently displays the lowest Relative Standard Error (RSE) values, almost close to zero, indicating precise but potentially overfit results. Conversely, the hglm with CAR model presents a narrower RSE range, typically below 25%, reflecting greater accuracy; however, it is worth noting that it contains a higher number of outliers. LR illustrates a performance similar to EBLUP, with RSE values reaching around 50% in certain scenarios and displaying slight variations across different contexts. These findings underscore the trade-offs between precision and robustness across these models, especially for finer geographical levels and countries not included in the initial sample.