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.

Sentiment transfer involves changing the sentiment of a sentence, such as from a positive to negative sentiment, whilst maintaining the informational content. Whilst this challenge in the NLP research domain can be constructed as a translation problem, traditional sequence-to-sequence translation methods are inadequate due to the dearth of parallel corpora for sentiment transfer. Thus, sentiment transfer can be posed as an unsupervised learning problem where a model must learn to transfer from one sentiment to another in the absence of parallel sentences. Given that the sentiment of a sentence is often defined by a limited number of sentiment-specific words within the sentence, this problem can also be posed as a problem of identifying and altering sentiment-specific words as a means of transferring from one sentiment to another. In this dissertation we use a novel method of sentiment word identification from the interpretability literature called the method of rationales. This method identifies the words or phrases in a sentence that explain the ‘rationale' for a classifier's class prediction, in this case the sentiment of a sentence. This method is then compared against a baseline heuristic sentiment word identification method. We also experiment with a pretrained encoder-decoder Transformer model, known as BART, as a method for improving upon previous sentiment transfer results. This pretrained model is fine-tuned first in an unsupervised manner as a denoising autoencoder to reconstruct sentences where sentiment words have been masked out. This fine-tuned model then generates a parallel corpus which is used to further fine-tune the final stage of the model in a self-supervised manner. Results were compared against a baseline using automatic evaluations of accuracy and BLEU score as well as human evaluations of content preservation, sentiment accuracy and sentence fluency. The results of this dissertation show that both neural network and heuristic-based methods of sentiment word identification achieve similar results across models for similar levels of sentiment word removal for the Yelp dataset. However, the heuristic approach leads to improved results with the pretrained model on the Amazon dataset. We also find that using the pretrained Transformers model improves upon the results of using the baseline LSTM trained from scratch for the Yelp dataset for all automatic metrics. The pretrained BART model scores higher across all human-evaluated outputs for both datasets, which is likely due to its larger size and pretraining corpus. These results also show a similar trade-off between content preservation and sentiment transfer accuracy as in previous research, with more favourable results on the Yelp dataset relative to the baseline.

No description available.

Geomagnetically induced currents (GIC) could saturate a transformer core. Two significant effects are the abnormal stray flux in transformers and extremely large harmonic contents in excitation currents, which can lead to serious equipment damage and power system misoperation. Such incidents have occurred during the March 1989 K-9 solar magnetic disturbance. This dissertation starts with a systematically study of transformer magnetization under GIC. It reviews both dc and ac saturation patterns of five transformer core designs. Magnetic fields along various traverses for dc excitation are presented. Impedance matrix entries of a single phase transformer are compared for normal and dc operations. New observations have been formed based on the simulation results. The study helps to reveal the fundamental transformer magnetization mechanism under GIC in order to assess potential stray flux heating possibilities of geologically vulnerable transformer units. Based on the finite element analysis, an improved method of modeling transformer excitation under dc bias using equivalent magnetic circuit is developed. There are two unique points in this approach: first, information of 3D finite element magnetic flux distribution analysis is used to construct and verify the circuit model; second, the effect of the transformer tank: is included The model is capable of simulating transformer excitation currents under different levels of dc bias with good accuracy. As a consequence, the complete variations of excitation current harmonics with respect to an extended range of dc bias are revealed. The sensitivity of transformer winding impedances and core loss on the excitation characteristics are examined. The saturated transformer under no-load and various loading conditions is simulated. A laboratory test is performed on a small scale transformer and compared with the model results. Excitation harmonics generated from dc biased three phase transformer banks with different types of equivalent loads are also simulated The effect of both unbalanced dc excitations and unbalanced loads are investigated The results of this study contribute in understanding transformers as harmonic sources and the impact on power systems during a geomagnetic disturbance. / Ph. D.

In high-frequency power transformers, magnetic material properties cannot be assumed to be constant. These properties vary with frequency, temperature, and magnetic flux density. Heat generation is, in turn, a function of the magnetic permeability, magnetic flux density, and frequency. Current design methods are either empirical or based on linear, uncoupled models. To better understand the relationship between heat transfer, magnetic flux density, material properties, and core geometry in a miniature, high-frequency transformer, a finite-element program has been developed to solve the coupled thermal-magnetic equations for an axisymmetric transformer. The program accounts for nonlinear temperature and magnetic field dependent material properties, geometry, and driving frequency. The program, HT-MAG, is based on a series of derived magnetic field equations. The Ritz method is applied to the magnetic and thermal equations in the development of the program. The program alternately solves the finite element approximations to the thermal and magnetic governing equations until the magnetic properties match within a specified fraction or a maximum number of iterations are performed. In addition, the program can be linked with existing pre- and post-processors or can accept manual pre- and post-processing. Six test cases were run to test the validity of the program. The first two cases tested the uncoupled heat transfer calculations. One of these tested the thermal conduction calculations while the other tested the heat generation calculations. The next two cases tested the uncoupled magnetic equations. The first was a direct current (DC) case, while the second was an alternating current (AC) case. The final two cases tested the thermal magnetic coupling. Solutions to these cases are presented and discussed. / Master of Science

A comprehensive and general analysis of the electromagnetic fields, power dissipation, and energy storage within transformer windings is presented. Emphasis is placed on applications in switched-mode power conversion. One-dimensional radial variation of the field quantities is assumed. The first phase of the investigation is for sinusoidal excitation; solutions for the current density and magnetic field intensity are derived and studied in order to develop a fundamental understanding of the basic phenomena. Expressions for the power dissipation and energy storage in both single- and multi-layer windings are then derived which, upon investigation, yield a technique for minimizing the power dissipation by choosing an optimum conductor thickness. Several levels of accuracy, ranging from exact solutions to very simple and physically meaningful series approximations, are defined and examined to determine their usefulness and range of validity. The time-harmonic treatment is generalized to arbitrary periodic exoitation by means of Fourier analysis, resulting in a powerful extension of its applicability to any possible converter topology. Results for several representative waveshapes are presented from which a fundamental dependence cn the waveform bandwidth is discovered. Practical application of the theoretical analysis is considered by developing models for several couon winding types: single and multi-filar round wire, litz wire, and sheet conductors. Experimental results are presented and compared with the theoretical results for each of these cases. Finally, a design procedure is outlined for switched—mode pour transformers which is based on this work. / Ph. D.

A method of calculating the yearly owning and operating cost of a distribution transformer is developed taking into consideration the loss of life due to overload. Using the developed methods, the potential benefits of managing an overloaded distribution transformer was calculated for a transformer on the Virginia Electric and Power Company (Vepco) System. By loading the transformer according to a saturation type load growth curve considered typical for Vepco System, its life was approximated. The fixed carrying charges were then applied at a rate sufficient to recover all invested capital during the life of the transformer. The potential savings were calculated when cutting the secondary and adding a transformer of equal one size smaller and two sizes smaller than the original. The study indicated no savings would be obtained when cutting the secondary. The only savings indicated were obtained by taking down an overloaded transformer and replacing it with the next larger size. The potential savings of managing these transformers presently installed, as well as those to be installed in the next years, as well as those to be installed in the next ten years, was calculated using a critical rate of return of 6, 7, 8, 9, and 10 percent. The calculated savings were $3,251,500 at 6 Percent, $2,674,400 at 7 Percent, $2,075,400 at 8 Percent, $1,602,200 at 9 Percent, $1,257,300 at 10 Percent. / Master of Science

The theory of operation of a controlled transformer is described. A PWM controlled transformer is proposed and implemented in a forward converter to provide post regulation. Experimental results are presented to verify the new control scheme. Overall efficiency of 82%-86% is achieved in a 200khz, 500-watt, 5V-output off-line regulator. A discussion of design issues of the controlled transformer is also presented. / Master of Science

Road accidents pose a significant threat, leading to fatalities and injuries with far-reaching consequences. This study addresses two crucial challenges in road safety: analyzing traffic intersections to enhance safety by predicting potentially risky situations, and monitoring driver activity to prevent distracted driving accidents. Focusing on Virginia's intersections, we thoroughly examine traffic participant interactions to identify and mitigate conflicts, employing graph-based modeling of traffic scenarios to evaluate contributing parameters. Additionally, we leverage graph neural networks to detect and track potential crash situations from intersection videos, offering practical recommendations to enhance intersection safety. To understand the causes of risky behavior, we specifically investigate accidents resulting from distracted driving, which has become more prevalent due to advanced driver assistance systems in semi-autonomous vehicles. For monitoring driver activity inside vehicles, we propose the use of Video Transformers on challenging secondary driver activity datasets, incorporating grayscale and low-quality data to overcome limitations in capturing overall image context. Finally, we validate our predictions by studying attention modules and introducing explainability into the computer vision model. This research contributes to improving road safety by providing comprehensive analysis and recommendations for intersection safety enhancement and prevention of distracted driving accidents. / Master of Science / Road accidents are a serious problem causing numerous deaths and injuries each year. By studying driver behavior, we can uncover common causes of accidents like distracted driving, impaired driving, speeding, and not following traffic rules. New vehicle technologies aim to assist drivers, raising concerns about driver attentiveness. It is crucial for car manufacturers to develop systems that can detect and prevent accidents, especially in semi-autonomous vehicles. This study focuses on intersections in Virginia and examines driver behavior within vehicles to identify and prevent dangerous situations. We create models of different traffic scenarios using graphs/networks and utilize machine learning to identify potential accidents. Our objective is to provide practical recommendations for improving intersection safety. Existing datasets and algorithms for recognizing driver activities often fail to capture common distractions like eating, drinking, and phone use. To address this, we introduce two challenging datasets specifically designed to capture distracted driving activities. Finally, we try to understand the predictions bade by the chosen deep learning model by visualizing the inner workings.

The popularity of AI audio applications is growing, it is used in chatbots, automated voice translation, virtual assistants, and text-to-speech translation. Audio classification is crucial in today’s world with a growing need to sort and classify millions of existing audio data with increasing amounts of new data uploaded over time. In the area of classification lies the difficult and lucrative problem of music recommendation. Research in music recommendation has trended over time towards collaborative-based approaches utilizing large amounts of user data. These approaches tend to deal with the cold-start problem of insufficient data and are costly to train. We look to recent advances in music generation to develop a content-based method utilizing a joint embedding space to link text with music audio. This approach has not been previously applied to music recommendation. In this thesis, we will examine the joint embedding methods used by recent AI music generation models and introduce a music recommendation system using joint embeddings. This music recommendation system can avoid cold-start, reduce training costs for music recommendation, and serve as the foundation for a cost-efficient content-based multimedia recommendation system. The current model trained on MusicCaps recommends the correct song per tag input within the top 50%-80% of all songs about 65%-70% of the time and we expect better results after further training.