16168 research outputs found
Sort by
Jodu praturtintų spindulinės atsitiktinės apšvitos indikatorių kūrimas.
No full textAccidental exposure to ionizing radiation, in particular γ-rays, can result from various sources: equipment malfunction, leakage from radiation sources or even hostile action, involving radioactive materials. While occupational exposure is well-regulated, the detection and assessment of unintentional exposure remains a challenge due to the limitations of existing monitoring technologies, which are often costly, complex or lack sensitivity. In this work poly(vinyl alcohol)-iodide (PVA-I) gel colored indicators, sensitive within the 0.2 – 10 Gy range, are proposed as a simple, cost-effective, and portable solution to address the need of immediate monitoring devices that could be distributed to the broad public in the case of nuclear accidents or radiological pollution, which otherwise may cause irreversible harm to any living organism. Two types of film and three types of solution indicators were fabricated. The sensitivity of F2 (0.45 w% KIO3) films, according to the ~340 nm absorbance maximum of triiodide ions, was only 0.0429 Gy-1. According to the ~490 nm absorbance maximum of PVA-triiodide complex, S2 (2 w% KI) indicator samples had a sensitivity of 0.0058 Gy-1 and S3 (5 w% KI) indicator samples had a sensitivity of 0.0135 Gy-1, which was comparable to a similar composition indicator sensitivity (0.010 Gy-1), described in literature. However, additional experiments must be done to ensure a better sensitivity to radiation and more resistance to ambient conditions, such as light, air and temperature
Giliuoju mokymusi grįsta kritinių organų segmentacija.
No full textIn recent years, deep learning (DL) methods have been increasingly applied to improve the automation and accuracy of organ-at-risk (OAR) segmentation in radiotherapy planning. Among these, convolutional neural networks (CNNs), particularly U-Net and its extensions, have shown significant advantages over manual contouring by offering greater consistency, reduced inter-observer variability, and faster execution in medical image segmentation tasks. This study systematically evaluated the performance of three 3D DL-based segmentation models: U-Net, Residual Encoder U-Net (ResEncU-Net), and SwinUNETR, on three multi-organ CT datasets: AMOS (Abdominal Multi-Organ Segmentation), BTCV (Beyond the Cranial Vault), KC (dataset provided by The Hospital of Lithuanian University of Health Sciences Kauno Klinikos), with the goal of assessing DL segmentation model suitability for clinical use. Quantitative evaluation based on Dice Similarity Coefficient (DSC), Surface DSC (sDSC), and 95th percentile Hausdorff Distance (HD95) revealed that ResEncU-Net delivered significantly higher segmentation accuracy across datasets, achieving a DSC of up to 0.916, sDSC values exceeding 0.88. U-Net demonstrated strong baseline performance, particularly in more homogeneous datasets such as KC, where its DSC (0.913) was close to that of ResEncU-Net (0.916). However, its segmentation accuracy declined slightly in datasets with greater variability. SwinUNETR, despite its Transformer-based architecture, showed the weakest performance, with largest mean HD95 values (up to 45.95 mm) and inconsistent sDSC scores, especially for small or low-contrast structures. These findings contrast with previously published results that reported strong SwinUNETR performance in large-scale studies, suggesting that its effectiveness is highly dependent on pretraining and dataset size – factors rarely feasible in clinical settings. In conclusion, the findings demonstrated that open-source segmentation models can be effectively integrated into clinical radiotherapy workflows. With expert validation by radiation oncologists, these models can significantly reduce manual contouring time, maintain high segmentation quality, and support more efficient and reproducible treatment planning. This highlights the practical potential of open-source DL-based tools for routine clinical use in radiation oncology, especially when it is tailored to institutional data and workflows
Application of mathematical methods for researching educational processes and their outcomes.
No full textImproving educational research and evidence-based education policymaking requires mathematical tools that can take into account the complex interactions between social systems and educational process variables. However, there is a lack of rigorous, theoretically and mathematically sound analysis of educational process outcomes – student achievement. Therefore, the aim of this research is to create a mathematical model capable of quantitatively explaining the influence of educational processes on students' mathematics achievement. The research objectives are: to analyze scientific literature exploring factors influencing high mathematics achievement; to systematize mathematical methods used in educational research for achievement modeling, highlighting their advantages and disadvantages; to quantitatively evaluate missing data imputation methods based on their influence on the reliability and validity of statistical inferences when modeling mathematics achievement; to create a mathematical model revealing the relationship between educational processes and mathematics achievement; to interpret the model parameters and possibilities for methodological expansion. The research employs methods of scientific literature analysis and secondary analysis of the PISA 2022 Lithuanian student achievement data. Due to the design features of the PISA study, a crucial part of the data analysis is missing data imputation. Among the missing data imputation methods, multiple imputation by chained equations (MICE) best reflected the original data distributions. Following mathematical modeling, a model was created describing the relationship between mathematics achievement and statistically significant influencing factors. The model uses variables selected according to Gagné's Differentiated Model of Giftedness and Talent. According to the developed explanatory mathematical model, it was determined that student-level factors, primarily economic, social, and cultural status, and mathematics self-efficacy, are the main factors influencing mathematics achievement, while school-level effects also proved to be significant, especially for the aforementioned factors. The main limitation of the empirical research model is the context of data collection during the COVID-19 pandemic; therefore, further research could test the model by applying it, for example, to PISA 2025 data or by improving its methodology using different mathematical methods in its development process
Impact of dental orthopedic and/or dental technician errors on dental implant overloading.
No full textThis research project aims to determine the impact of dental orthopedic and(or) dental technician errors on dental implant overloading. Proper assessment of errors can help to make decisions that extend the life of implants and improve the quality of life of patients. The study addresses the factors that influence the mechanical performance of dental implant systems under functional loads, including implant positional accuracy and assembly features. A literature review is carried out which discusses the basic functions of teeth, the causes and possible consequences of tooth loss, as well as the process of osseointegration and its importance in implantology, traditional and digital oral impressions, denture bonding and denture fabrication. The causes and consequences of overloading of dental implants, dental orthopedic and dental technician errors in implantology are examined. Geometric modelling and finite element analysis were carried out using models of the implant system and the three-tooth bridge in „SolidWorks“ software environment. The geometry of the components of the implant system was reconstructed on the basis of physical measurements, as the files provided by the manufacturer were not suitable for finite element analysis. Five different dental bridge configurations were created to analyze the effect of different implant insertion angles along the jaw on the stress distribution. The materials for the components of the implant system are assigned according to the manufacturer's data: titanium for the implants and abutments, zirconium oxide for the three-tooth bridge, for which the properties of the material were obtained from data provided by dentists. The boundary conditions and constraints were chosen to simulate real clinical conditions: the implants were assigned a fixed constraint, simulating strong osseointegration with the surrounding bone, the contact between the bridge and abutments was selected as „bonded“, simulating an adhesive layer and the bridge was subjected to loading conditions simulating masticatory forces, with the assignment of locations, values and angles of tilt chosen in accordance with the recommendations of dentist. The obtained results show the importance of precise implantation and prosthetics, as even small deviations can lead to disproportionate stresses, increasing the risk of structural failure. The submitted conclusions are that errors by the dental orthopedic surgeon and/or dental technician can lead to an inappropriate distribution of loads, causing premature wear of components, loosening of fixing screws or even fracture of components
Keitimo proceso tyrimas automobilių įmonėje naudojant SMED metodą.
No full textThe LEAN methodology is widely used in various manufacturing enterprises. One of the most competitive industries is automotive. Electronics production is advanced, however, the changeover process between different types of products is time-consuming. 5 LEAN methods were analyzed and described. The 5S method is a practice used to ensure a tidy and clean manufacturing environment. Spaghetti diagrams analyze excessive movements performed during production processes. Poka Yoke method is used to reduce the possibility of human errors. Kanban raw material scheduling system used to ensure just-in-time production. SMED method LEAN management tool used to reduce duration of the changeover. Changeover is time time-consuming process where excessive movement can cause downtimes which increase production costs. The goals of this research are to compare production processes applied in changeover, choose relevant and suitable production management methods, compare the results of the changeover process before and after the application of the SMED method, and perform cost-effectiveness analysis. Various scientific sources that apply SMED were analyzed. The SMED method has a major impact on different production enterprises affecting cost saving, waste reduction, and improved productivity. The methodology was created according to examples applied in manufacturing companies. The changeover process was analyzed in an automotive electronics company that produces parts for major players in the market. Changeovers were analyzed in two production Surface Mount Technology lines using SMED methodology. Both SMT production lines were compared, and it was concluded that the same solutions were to improve the changeover process. The solutions were proposed by applying 5S and spaghetti diagram methods. Applied SMED method and solutions reduced changeover time resulting in increased productivity and cost savings. SMED method reduces downtimes and improves changeover efficiency and flexibility
Analysis and implementation of a hybrid algorithm for financial transaction reconciliation.
No full textIn the daily practice of corporate accounting departments, precisely and rapidly matching purchase and sales invoices stored in enterprise resource planning (ERP) systems with bank payment records remains a significant challenge. The heterogeneity of bank statement formats - driven by varied legal and technical requirements across jurisdictions - prevents full automation via rule‐based methods, since each client demands custom rule sets that are time-consuming to create and maintain. This thesis presents the design, implementation, and evaluation of a hybrid bank transaction reconciliation algorithm integrated into "Microsoft Dynamics Business Central“ system’s "Bankfeed" module. The core contribution is an algorithm‐selection meta‐model that, leveraging historical matching data, dynamically chooses the optimal reconciliation method – whether rule‐based or machine-learning-based, for each incoming transaction. During the analysis part, we identified the machine-learning model variant that most accurately predicts the best matching approach, measured its precision and coverage, and introduced weighting schemes to balance accuracy against computational cost. Finally, we report on a field trial assessing practical performance and statistically validate the improvements in matching accuracy and operational efficiency compared to the legacy rule-based system. This hybrid approach hopes to significantly reduce manual configuration effort and administration cost, while boosting the day-to-day effectiveness of accounting staff in ERP environments
Air pollution indicators study.
No full textThe master’s thesis investigates and analyzes air pollution indicators across the territory of Lithuania. The topic of air pollution is relevant to all residents, as it affects the health of every individual. Although this study focuses specifically on Lithuania, its methods are universally applicable to any region of the world. In addition to air pollution metrics, this project also reviews health-related data indicators. The study examines the World Health Organization’s recommended air pollution indices and their threshold values, and takes into account the European Union’s directives on air quality standards. The aim of this research is to conduct a statistical analysis of air pollution indicators within Lithuania. The objectives are to analyze open-source air quality and health data, to perform time-series analyses to identify pollution trends, to test for associations between air pollution and health outcomes, to determine relationships between potential pollution sources and air quality. The thesis employs methods for compensating time-series data and detecting outliers. It also uses statistical tests for normality, correlation analysis, and time-series decomposition techniques to uncover seasonality, cyclicality, and deterministic trends. Additionally, methods are applied to adjust pollutant concentration data for wind direction and speed. Two new approaches are introduced: one for filling missing values in a specific type of multivariate time series (a sum-interpolation with outlier detection and imputation method), and another model that links air pollution indicators to stationary pollution sources, accounting for meteorological conditions, to assess their impact on air quality metrics. The study develops a method for imputing missing time-series values and evaluates the relationship between PM2.5 concentration and increased mortality. It also derives coefficients quantifying the influence of stationary sources on air pollution levels. As a result of this master’s thesis, open air quality and health data sources have been analyzed, their strengths and limitations discussed, and the most useful datasets selected. Both univariate and multivariate time-series analyses were conducted, missing data were imputed, and outliers handled. Pollution trends were identified, and existing links between air pollution and health indicators explored. The findings are compared with similar studies, and relationships between pollution sources and air quality metrics are established
Research of business models for renewable energy communities.
No full textThe aim of the final Master's thesis is to study the business models of renewable energy communities. The literature part of the thesis analyses the legal regulation and barriers to the development of renewable energy communities and, after assessing the barriers to development and the legal aspects, examines the business models. After an analysis of several literature sources, three business models are selected for subsequent use in the study - the peer to peer business model, the multifamily solar model and the RES community with a remote solar plant and EV charging stations. In the methodological part, a technical methodology is developed for the evaluation of the solar PV plants, and a methodology for the evaluation of the economic indicators, which tells whether the indicators are economically acceptable, and sensitivity analyses are used to finally justify the economic calculations. The exploratory part of the thesis analyses the main economic indicators of the communities' business models, such as: NPV, IRR and LCOE. After the economic calculations, an analysis is carried out on which partnership is the most cost-effective. The study also assesses the payback period of the business models of the RES communities and the sensitivity analysis of the economic indicators to changes in electricity prices, the amount of the subsidy or the interest rate on bank loans
Investigation of impact of renewable energy development on profitability of offshore wind farms.
No full textThe subject of this study is a 700 MW offshore wind farm in the Baltic Sea, which represents an important step towards transitioning to a sustainable and low-carbon energy system. This transition is essential for addressing climate change issues, enhancing energy security, and meeting the growing demand for cleaner energy sources. Offshore wind turbines, characterized by high wind speeds and the ability to produce energy on a large scale, are becoming increasingly important in this process. However, their integration into the energy system is not without challenges, such as the price cannibalization effect, which reduces electricity market prices during periods of high generation and may undermine the economic viability of such investments. The aim of the study is to evaluate the impact of renewable energy resource development on the profitability of offshore wind farms, taking into account electricity market prices and the costs of wind farm installation. The literature review covers trends in renewable energy resource development in Lithuania and globally, challenges in offshore wind farm development, and the impact of the price cannibalization effect on the profitability of offshore wind farms. The methodology involves calculating park costs and creating different scenarios to assess the impact of electricity prices on park profitability. Various scenarios of renewable energy development are evaluated for their impact on the profitability of offshore wind turbines. Pessimistic and realistic scenarios show how price cannibalization reduces wind park revenue when CPF indicators are less than 1. In optimistic scenarios, CPF is greater than 1, indicating profitability without the impact of cannibalization. This topic is relevant because, with the rapid development of renewable energy in recent years, there is a growing need for the expansion of offshore wind farms in Lithuania. To attract investors, profitability must be ensured. Comparative and sensitivity analyses of different scenarios provide a comprehensive understanding of the conditions under which an offshore wind park can achieve optimal financial results, contributing to strategic planning and investment decision-making
Antrinio vėžio rizikos sąsajos su kritinių organų apšvita vertinimas ir analizė nagrinėjant galvos ir kaklo vėžio atvejus.
No full textHead and neck cancer (HNC) presents complex treatment challenges due to the proximity of tumours to critical organs, increasing the risk of radiation-induced complications. This study showed how demographic, anatomical, and clinical factors are influenced by radiation dose delivery and the associated risk of secondary cancers. The analysis indicated that HNC incidence was highest among men aged 60–69, with the oral cavity and oropharynx being the most affected sites. Statistical analysis showed that depending on the patient's age, the irradiation doses of organs such as the oesophagus and parotid glands differ, while gender had the least impact on the change in irradiation dose of critical organs. The assessment of Excess Absolute Risk (EAR) revealed a consistent trend that patients with more advanced disease stages particularly those classified as Stage IV (T4_N1_M0) faced significantly higher long-term risks. To assess the accuracy of treatment planning system (TPS) dose estimations, out-of-field doses were measured using a cylindrical ionization chamber (PTW 30013) positioned within the SHANE anthropomorphic phantom. These measured values were compared with TPS-calculated doses using two algorithms: the Anisotropic Analytical Algorithm (AAA) and Acuros XB. Both algorithms exhibited consistent underestimation of out-of-field doses, with the most significant discrepancy 57.1 % observed in Channel 4 (spinal cord region) at 0 cm using the Acuros XB algorithm. Moreover, at distances beyond 10 cm from the treatment field edge, underestimations remained notable, particularly in Channels 3 and 4. These findings highlight the limitations of TPS algorithms in accurately predicting peripheral dose, reinforcing the importance of incorporating phantom-based dose measurements into routine radiotherapy planning to ensure reliable dose assessment for organs at risk (OARs), ultimately contributing to improved long-term patient outcomes