16168 research outputs found
Sort by
Graphene direct growth by microwave PECVD on h-BN films deposited by reactive HIPIMS /
No full textThis research explores the synthesis of graphene using microwave plasma-enhanced chemical vapor deposition (PECVD) on a hexagonal boron nitride interlayer deposited by reactive high-power impulse magnetron sputtering. The effects of h-BN interlayer composition and thickness on the graphene structure, morphology, and electronic properties were investigated using Raman scattering spectroscopy, atomic force microscopy, and conductive atomic force microscopy. The electrical and photoelectrical characteristics of the graphene/Si(100) and graphene/h-BN/Si(100) diodes were studied. It was revealed that graphene self-doping effects, primarily originating from substrateinduced charge transfer, can be partially controlled, and that the dominant defect type of graphene can be changed by varying the boron interlayer thickness. The graphene layers synthesized on the SiO2 film were substantially smoother than those grown on the boron nitride films. Graphene grown on h-BN had a substantially greater surface current than graphene synthesized on the SiO2 layer. There was no relationship between graphene self-doping and graphene surface conductivity. Nevertheless, it should be noted that reduced graphene self-doping was achieved even though the surface roughness of the h-BN film was significantly greater than that of the SiO2 film or Si(100), and B–HandC–Hbonds were present in some h-BN films. Tunneling is the primary reverse current charge transfer mechanism, similar to graphene/Si(100) diodes, and h-BN interlayers cannot decrease the reverse dark current. Despite these circumstances, the insertion of the h-BN interlayer resulted in a significant increase in the photocurrent, short-circuit current, and open-circuit voltage compared to those of the graphene/Si(100) heterojunction. The observed effects of the h-BN interlayer on the graphene/Si (100) diode properties were explained by the competition between the effects of the different h-BN film compositions, thicknesses, and roughness on the one hand and the influence of the graphene structure and electronic properties
Multiscale and multi-temporal simulation of change of urban structures in the subarctic East Siberian metropolis of Yakutsk /
No full textIn the last twenty years, the Metropolis of Yakutsk has experienced significant changes characterised by intense urban growth, densification, and urban structure changes in a complex geographical environment: extremely low temperature during six to seven months, impacts of permafrost dynamics and relative melting, seasonal exposure to ice breakup on the suburban areas. The urban structure and land-use changes conditioned by the urban growth and environmental impacts are analysed at two geographic scales: the meso-urban level with the use of the Landsat-5 TM, Landsat-8 OLI, and Sentinel-2 MSI satellites images covering the period from 2010 to 2020; and at the regional level with the DMSP-OLS PL (1995–2013) and VIIRS-DNB (2015–2020) sensors. The recognition of the urban structures and land use transformations at both scales are based on the use of the combined machine learning data processing. The simulations of the urban structures and land use evolutions to 2030 at the meso-urban and regional scales by Markov chain cellular automata give comparable results of the future trends of the Yakutsk metropolis: reduction of vegetation, forests areas (due to forest fires) and agriculture zones; the increase of bare soil, water surfaces and new urban areas
Quality assessment of the lithuanised microsoft word programme.
No full textAs localisation becomes more prominent in the digital world, there is an increasing need to study its processes or end products. Therefore, the quality of localised products is often assessed. Research on the quality of products and/or services adapted to the Lithuanian language is scarce. This study is also relevant because it examines Microsoft Word software, which is frequently used in professional and academic environments. The study of this Lithuanised software’s quality is important for all those who want to understand how well the software has been translated into Lithuanian. The object of the study is the quality of the graphical human user interface of Microsoft Word. The project aims to conduct an expert evaluation of the Microsoft Word (version 2021) software according to the list of errors in the Software Localisation Quality Assessment E-Service User Guide. To achieve this goal, five main objectives were set: to discuss the concept of localisation and the research articles on the quality of localisation, identifying the main theoretical aspects; to determine the attitudes of Microsoft Word users towards the use of the Lithuanised version of the software and the errors it contains; to identify, by using the Software Localisation Quality Assessment E-Service User Guide, the errors and problem areas found in Microsoft Word; to compare the original and the Lithuanised versions of Microsoft Word based on the errors found and assess the quality of the localised version; to assess the applicability of the list of errors in the Software Localisation Quality Assessment E-Service User Guide to the study of the quality of the localised product. In total, 104 errors were detected in Microsoft Word, and twenty-four examples are discussed in the paper. The errors are categorised according to the use of the Software Localisation Quality Assessment E-Service User Guide. A survey of Microsoft Word users also found that users are more likely to use the original version of the software because of errors in the Lithuanised version. Overall, this study contributes to the rapidly growing field of localisation, especially in the area of Lithuanisation, and to improving the quality of software localisation. In addition, it helps to assess the quality of localised software not only for the public but also allows language/localisation experts/developers to consider errors and further improve localisation processes
Model for automated information technology security policy management.
No full textThe objective of this Master's thesis is to develop and investigate an automated IT security policy management model suitable for use in modern organizations. The thesis analyzes existing security models and identifies their shortcomings, emphasizing the risks of human errors and the necessity of reducing manual work in the implementation of IT policies. A new model integrating Natural Language Processing (NLP) technology and the "Policy-as-Code" principle was developed to automatically interpret IT security policies documented within organizations and effectively deploy them. In the experimental part of the research, a prototype of the proposed model was created and tested in practice. Results indicated that the automated solution significantly improved the level of information security, reduced the likelihood of human errors and ensured compliance of the organization's policies with the infrastructure security settings. IT security policy management became automated, which resulted in improved transparency in security management processes and more efficient resource utilization were observed, demonstrating the practical applicability and benefit of the proposed model for enterprises aiming to ensure policy compliance and to enhance their IT security posture
Deep reinforcement learning for computer games agents’ control.
No full textTo apply artificial intelligence to a wider range of problems, it must be able to: explore its environment independently, learn the most optimal actions in specific situations, and be capable of planning several steps ahead. To solve such tasks, reinforcement learning algorithms are used and recently, neural networks–based ones. High demand for training data is one of the main issues. In this study, we conducted experiments with modifications of the popular PPO algorithm and aimed either to reduce the amount of data required or to achieve better results with the same amount of data. The experiments were carried out in three environments with different characteristics — computer games: “Ms. Pacman”, “VizDoom Health Gathering”, and “VizDoom Defend the Center”. We proposed two separate modifications: a distributed policy function and VAE-based image compression. The distributed policy function is based on ideas from ensemble machine learning methods such as random forests. During training, the branches of the distributed policy function acquire different random weights, allowing the agent to learn a more diverse overall strategy. In one of the games, the best variation achieved an average of 1,518 points per episode, while the unmodified algorithm scored 1,377, indicating ~10% improvement. All algorithm variations had training instabilities and converged to either better or worse results, but the modified algorithm showed a much higher variance in results. In the best trial, our modified algorithm scored 2,046 points, while the unmodified one scored 1,666, which corresponds to ~23% better performance. In the other modification, we explored how a VAE could be used as a potentially faster neural network training method when applied to image compression. This was intended to help us use the data more efficiently by enabling repeated training on it. However, after various integration experiments, it turned out that training the VAE and PPO simultaneously was quite unstable and did not yield better results than the original algorithm
Analysis of optimization model for interconnection of Baltic electricity bidding zones.
No full textThe master’s thesis analyses the configuration of electricity bidding zones in the Baltic States and its impact on market performance. This topic is relevant due to the increasing share of renewable energy sources, the rapid integration of electricity markets, and the need to ensure efficient price formation and supply security. The objective of the study is to evaluate how different bidding zone configurations affect market indicators, based on the assessment criteria defined by the European Union. The research compares scenarios of separate and merged bidding zones in the Baltic region, covering three target years – 2025, 2030, and 2050 – and incorporating various climatic years to assess sensitivity to meteorological conditions. The study was conducted using the PLEXOS software, with a mathematical optimization model of the electricity system developed for scenario analysis. Three key evaluation criteria were selected: electricity prices, unserved energy, and curtailed renewable energy generation. Additionally, a sensitivity analysis was carried out by modelling the isolation of the Baltic States from external bidding zones. The modelling results showed that in 2025 and 2030, interconnected zone ensured lower price volatility, more efficient use of renewable resources, and high supply reliability compared to separate zones. In contrast, the 2050 results revealed increased price differences and renewable curtailments, indicating that interconnected configuration becomes less effective under these conditions. The price homogeneity analysis, based on separate zone results, confirmed that in 2025 and 2030 the zones met the criteria for consolidation, while the 2050 outcomes suggested a deterioration of key indicators, making consolidation unjustified without additional grid development. The sensitivity analysis demonstrated a high dependence of the system on cross-border interconnections. Under isolation conditions, electricity prices, unserved energy volumes, and renewable generation curtailments increased significantly, especially in the 2050 scenario. When comparing configurations, the merged zone performed more effectively than the separate zones by achieving lower curtailment levels and more balanced system operation; however, reliable performance was not ensured in either configuration without infrastructure expansion. An optimal bidding zone configuration is closely linked to system stability and the ability to absorb extreme fluctuations in both price and supply. A merged Baltic zone contributes to more efficient resource allocation, reduces price volatility, and enables better risk management related to climatic variation and imbalances in supply and demand. This type of zone design could serve as one of the strategic solutions to ensure a competitive, sustainable, and reliable electricity system in the Baltic region
Research of corona discharge and its generated electric wind.
No full textThe final master's thesis investigates the corona discharge and the electric wind it generates. The literature review covers the main methods for modeling the electric field, the applications and types of corona discharge, possible electrode configurations, and the principles of corona discharge formation. Additionally, the review examines the historical development of the electric wind generated by corona discharge, the dependence of the electric wind formation mechanism on polarization, and the generation of electric wind by corona discharge. A wire-to-plate electrode configuration was chosen for the study due to its wide practical application and the higher EHD wind velocity it produces compared to other configurations. The modeling uses the "COMSOL Multiphysics 6.2" software, applying the finite element method. A mathematical model of corona discharge is developed, the onset voltage for the discharge is determined, and the electric field strength is calculated. The dependencies of corona current and EHD wind velocity on applied voltage are analyzed, along with the spatial distributions of electric potential, space charge density, and electric field intensity on the electrode surface. An experimental investigation of corona discharge and generated EHD wind is also carried out using the same wire-to-plate electrode configuration. The experimental results are analyzed and compared with those obtained from the mathematical model to assess the model’s accuracy and reliability
Research of wind power plants impact on power system stability.
No full textThe purpose of the final thesis is to study the impact of wind farms on the stability of the transmission network. A system model has been prepared to simulate various scenarios. The results obtained are evaluated based on various sources and scientific articles. The DIgSILENT PowerFactory software package is used to conduct the study. Using this software, the “Transmission system” model is used as a base, which was modified to meet the requirements of the Lithuanian electricity system. During the study, the system response to a fault and its ability to return to normal operation were assessed. Frequency fluctuation, rotor angle deviation, network voltage and the system's ability to return to normal operation were assessed. It was found that integrating a large number of wind farms, especially those with low or no inertia, significantly reduces the system's ability to suppress frequency fluctuations after disturbances. Rotor angle analysis also showed an increased risk of loss of synchronization at low inertia. The study conclusions are relevant for energy system designers and network operators seeking a smooth transition to a climate-neutral energy system
Influence of enzymatic hydrolysis on composition and technological properties of black currant (Ribes nigrum) pomace /
No full textBlackcurrant (Ribes nigrum) is valued for its health-promoting compounds, many of which remain in the pomace after juice extraction. Berry pomace can be considered a valuable source of dietary fiber. However, it is typically dominated by insoluble dietary fiber (IDF), and the soluble-to-insoluble fiber ratio is often nutritionally suboptimal. The aim of this study was to evaluate the influence of enzymatic hydrolysis on the composition and technological properties of blackcurrant pomace (BCP). Three commercial enzyme preparations—Viscozyme® L, Celluclast® 1.5 L, and Pectinex® Ultra Tropical (Novozymes A/S, Denmark)—were used for enzymatic hydrolysis, which was conducted at 50 °C for 1 h. The enzymatic treatments altered BCP’s chemical composition and technological properties. Pectinex® Ultra Tropical and Viscozyme® L primarily hydrolyzed SDF, while Celluclast® 1.5 L was more effective on IDF, resulting in increased SDF content and an improved SDF/IDF ratio. Enzymatic hydrolysis reduced the oil retention capacity and impaired stabilizing properties, but it increased both the water retention capacity and the solubility index. It was found that the creaming index of the pomace deteriorated with decreased IDF content. The findings indicate that the effects of enzymatic modification on BCP’s composition and technological properties can vary significantly, supporting its potential application in the development of novel food products
Advancing fractal dimension techniques to enhance motor imagery tasks using EEG for brain–computer interface applications /
No full textThe ongoing exploration of brain–computer interfaces (BCIs) provides deeper insights into the workings of the human brain. Motor imagery (MI) tasks, such as imagining movements of the tongue, left and right hands, or feet, can be identified through the analysis of electroencephalography (EEG) signals. The development of BCI systems opens up opportunities for their application in assistive devices, neurorehabilitation, and brain stimulation and brain feedback technologies, potentially helping patients to regain the ability to eat and drink without external help, move, or even speak. In this context, the accurate recognition and deciphering of a patient’s imagined intentions is critical for the development of effective BCI systems. Therefore, to distinguish motor tasks in a manner differing from the commonly used methods in this context, we propose a fractal dimension (FD)-based approach, which effectively captures the self-similarity and complexity of EEG signals. For this purpose, all four classes provided in the BCI Competition IV 2a dataset are utilized with nine different combinations of seven FD methods: Katz, Petrosian, Higuchi, box-counting, MFDFA, DFA, and correlation dimension. The resulting features are then used to train five machine learning models: linear, Gaussian, polynomial support vector machine, regression tree, and stochastic gradient descent. As a result, the proposed method obtained top-tier results, achieving 79.2% accuracy when using the Katz vs. box-counting vs. correlation dimension FD combination (KFD vs. BCFD vs. CDFD) classified by LinearSVM, thus outperforming the state-of-the-art TWSB method (achieving 79.1% accuracy). These results demonstrate that fractal dimension features can be applied to achieve higher classification accuracy for online/offline MI-BCIs, when compared to traditional methods. The application of these findings is expected to facilitate the enhancement of motor imagery brain–computer interface systems, which is a key issue faced by neuroscientists