Computational Science and Techniques

This paper presents a comparative analysis of ballast water treatment technologies, along with a detailed evaluation of the selection of treatment equipment, calculations of hydraulic pressure loss, and a theoretical layout of the equipment. The technology analysis assesses 13 different treatment methods based on six criteria for installing such systems in the space-restricted engine room, aiming to mitigate the threat posed by untreated ballast water to marine life. The selected technologies are filtration and ultraviolet (UV) as the primary and secondary ballast water treatment technologies. These methods ensure efficient, rapid, and environmentally friendly ballast water treatment.

This study presents an analysis of floating cranes and the design of one such crane. During the analysis phase, various types of floating cranes and their structural designs were examined. Based on this analysis, the necessary calculations were performed to select a suitable crane and pontoon.

In Lithuania, solar energy is expanding rapidly. In 2020, the installed capacity of solar power plants was 170 MW, and by 2025, it exceeded 900 MW. This growth reflects not only technological progress but also an increasing interest from consumers and businesses. With rising electricity consumption and more solar modules being installed, it is becoming increasingly important to analyse how technological efficiency is evolving in the Lithuanian climate and the influence of the legal framework on this process. Properly coordinated technical and legal development can ensure the sustainable and long-term advancement of solar energy. This research paper explores the interplay between legal solutions and technological efficiency in the development of solar power plants, analyses the main challenges facing Lithuania and the EU, and identifies areas that remain underexplored. Directions for further research are provided, emphasising system optimisation and long-term performance.

Augant naudojamų daiktų interneto įrenginių bei tinklų skaičiui, kyla ir naujos saugumo rizikos, susijusios su šiais daiktų interneto tinklais. Šiame darbe yra pristatoma lyginamoji įvairių mašininio ir gilaus mokymosi modelių klasifikavimo našumo analizė – tikslumo bei atpažinimo metrikos – naudojant CIC-IoT-IDS2023 duomenų rinkinį, skirtą apmokyti mašininio mokymo modelius aptikti prieš daiktų interneto įrenginius nukreiptas atakas. Taip pat pateiktiamos gaires tolimesniems darbams kuriant kibernetinių įsilaužimų aptikimo sistemą daiktų interneto tinkluose. Siūlomos gairės apima sprendimo medžių modelių hiperparametrų tobulinimą, realaus laiko įsilaužimų aptikimo komponento kūrimą bei testavimą naudojant kraštinius (angl. edge) įrenginius.