Mathematical Modeling and Simulation of Systems

A web system for the analyzing unknown sources of radioactive pollution in the air was developed. With inverse modeling of atmospheric transport using the atmospheric transport model FLEXPART and the newly developed minimization module SIMFLEX, it allows to automate the solving of the inverse problem when radioactive pollution in the air is detected. The probability density distribution of the source location, estimates of release start time, duration, and inventory are visualized on a cartographic basis. Testing of the system was performed using measurement data collected during an incident with the detection of ruthenium-106 in the atmosphere in the fall of 2017. With the developed system, the estimated source location that minimized the cost function was closer to the most probable source (Mayak, Chelyabinsk region of the Russian Federation) than in previous calculations, which were based on the use of an identical set of measurements and Eulerian long-range atmospheric transport model. Presently the system is implemented for use by the radiology experts of the Ukrainian Hydro-meteorological Center.

The cascade of water reservoirs is a complex and non-linear system with many unpredictable natural and human-dependent factors that define water flow characteristics and influence water quality. This work provides an analysis of radioactive pollution data collected in 1999 for all six reservoirs and major inflowing rivers. The regression techniques were applied to forecast water flow and future water contamination. The initial stage of research is to determine the optimal set of input parameters that balances the addition of new parameters with noise available in the dataset. The next stage is to find the best machine learning algorithm, and for this specific task, we selected an extra trees regressor. The final stage is to forecast the propagation of radioactive pollution through a cascade when only initial pollution impulse data are provided.

The article studies the problem of constructing a system mathematical model of the differential inversion signal to determine the concentrations of chemical elements in multicomponent signals when measuring the ecological state of environmental objects using electrochemical methods of inversion chronopotentiometry. Using the least squares method, a mathematical model of measurement signals was obtained that allows more accurate determination of trace concentrations of toxic elements. For this purpose, the inversion signal coordinate systems were changed, an intensity signal was constructed as a transformed differential inversion signal, the baseline model of the background discharge of electrodes was built, the spectrum of the inversion components of elements was formed, models of the component signals were introduced and the inversion time was determined for calculating concentrations by the method of adding standard ion samples. The developed mathematical model was tested, which showed that the application of the model is adequate: it does not distort the integral characteristics and the obtained values of the concentrations of chemical elements. The relative error in determining the zinc concentrations in a standard solution using the mathematical and test models is 2.7%, which is acceptable in engineering calculations. The developed mathematical model of the differential inversion signal can be used when developing analytical instruments with an automated software system for processing electrochemical data.

The article is devoted to the modeling of widespread process of blowing gas through a liquid with several interacting substances in closed or top-opened unmovable reservoir. The model can be applied to wastewater treatment of the Kamianske city, as example of medium-size city, particularly to aeration tanks, where maintaining optimal conditions for the biological degradation of pollutants is important. The article presents a fuzzy control system and mathematical model, which accounts for key variables including ammonia concentration, oxygen level, and the concentration of activated sludge, all of which are essential for effective wastewater purification. The results show overall trend of gradual reduction in outlet ammonia concentration. In addition, the observed trends show that increasing of active sludge concentration can lead to worse or the same treatment efficiency when dissolved oxygen is low. In contrast, the level of dissolved oxygen much better affects of the resulting ammonia concentration. The effectiveness of the fuzzy control approach is demonstrated during simulations that show improved regulation of ammonia and oxygen levels, as well as enhanced maintenance of activated sludge concentration. This approach offers a solution with practical significance for optimizing the performance of aeration tanks, contributing to more efficient and sustainable wastewater treatment process.

The problem of optimizing traffic on the road network is considered. Congestions within road networks continue to be relevant, and in situations where there is no systematic approach to its solution, losses due to the inefficiency of the infrastructure can be very large. Three modified modern algorithms for solving variational inequalities applied to the user-optimal traffic equilibrium search problem. For that, mathematical model of traffic equilibrium corresponding to the first principle of Wardrop is provided, with formulation in form of variational inequality. The algorithms are adaptive versions of extragradient algorithm, extrapolation from the past, and Malitsky algorithm. We do not need to know Lipschitz constant of operator to select the step size. All three algorithms are extended with step size modification rule which allows both step increasing and decreasing. Numerical experiments are conducted with the toy problem (Braess classical example) and with the well-known Sioux Falls transportation network. Developed software complex for applying algorithms based on variational inequalities allows to make empirical conclusions about behavior of algorithms during calculation process.

The aim of the article is to develop and apply a modelling system to support decision-making regarding the spread of infectious diseases. The recent course of the COVID-19 pandemic worldwide has emphasised the need for software tools to prevent, analyse, and manage the epidemic process. Although by the time the pandemic began, there were already quite a few mathematical tools for modelling infectious diseases, practical experience with their application revealed their weaknesses and limitations. One of the main shortcomings of mathematical models is the uncertainty of parameters in the case of an epidemic caused by a new, unknown infection. When there is not enough time to conduct medical experiments to determine the properties of the virus, the challenge arises to develop algorithms for calibrating models based on available observational data. This work develops a modelling system for managing the epidemic process, which allows for the prediction of the spread of the COVID-19 viral infection, taking into account the age groups of the population and their spatial heterogeneity. It also includes a parameter optimisation block that allows for flexible adjustment of the model as new information about the virus and new statistical data become available. Model parameter calibration is based on comparing the predicted results with observed data by solving a nonlinear constrained least squares problem. The modelling system was applied to simulate four waves of COVID-19 in Ukraine.

Estimating the excess mortality in Ukraine caused by the COVID-19 pandemic is an important challenge in understanding the true burden of the pandemic. This article examines the purely excess mortality that occurs against the background of the mortality from COVID-19 with a replication of its dynamics on a larger scale. So, we are talking about both the underdiagnosis of COVID-19 when determining the causes of death, which is often a challenge because the vast majority of those who died from COVID-19 had comorbidities, and other excess deaths that occurred in the context of the mortality from COVID-19, which may be related to the pressure on the health care system, etc. Using a correlation analysis approach to daily all-cause mortality, an excess mortality estimate of 250,000 deaths (95% CI: 242,000 to 258,000) in 2020–2021 against confirmed nearly 98,000 COVID-19 fatalities was obtained. This approach is based on the assumption of low variability in the ratio of excess mortality associated with COVID-19 to confirmed mortality from COVID-19, which assumes a constant value for each of the three predominant variants of SARS-CoV-2. Using a correlation analysis approach, monthly data with age-gender structure of all-cause deaths were examined. The ratio of excess mortality to confirmed mortality from COVID-19 was estimated for nine age categories of each gender. It was established that this ratio increases with age for each gender, is higher for women regardless of age, and is higher for men adjusted for age structure.

A model for regulating the viability and efficiency of nodule bacteria to increase soybean productivity has been developed. The component of this model is a block that assumes the use of a complex of chemical compounds (CCS). The complex consists of polymers and compounds of protein nature (sodium alginate, starch, gelatine), which, in addition to protective action, contribute to better fixation of rhizobia on seeds. In the field experiment conditions it was established that the greatest effect of soybean bacterization can be obtained by strain of nodule bacteria Bradyrhizobium japonicum and complex of chemical substances. The developed model has demonstrated its viability. In compared without CCS the combined use of microbial inoculum and CCS contributed the formation of balanced symbiotic system, increase the level of molecular nitrogen fixation. In a field experiment inoculation of soybean seeds with B. japonicum and CCS to increase the level of soybean yields by 12–13% compared to the inoculation without CCS. The effectiveness of the combination of CCS and B. japonicum in the processing on the day of sowing and one month before sowing of seeds was studied. It has been shown that the effectiveness of treatment with B. japonicum and CCS in advance was less efficient in terms of yield formation relatively same variant of treatment of seeds directly pre-sowing.

The article discusses the use of modern modeling methods to optimize technological processes in the production of bakery products. The main goal of the research is to develop mathematical models that will improve baking parameters and ensure high quality of finished products. The key factors influencing the quality of baked goods, in particular temperature, humidity, baking time and dough composition, are analyzed. A variety of approaches are used to model baking processes, including linear and polynomial regression models, partial least squares, decision trees, and ensemble methods. The choice of these methods is justified by their ability to efficiently process categorical and numerical data, as well as to take into account complex relationships between process parameters. The study provides a comparative analysis of models using R2 and MSE metrics, which allows us to assess the accuracy and reliability of predictions. Particular attention is paid to the issues of coding categorical variables, in particular the use of Label Encoding, and its impact on modeling results. The results of the study demonstrate that the use of the proposed models in the control process can improve the quality of baked goods, providing optimal baking conditions.

This study focuses on the analysis of the transverse vibratory motion of two homogeneous and isotropic beams, clamped at their ends. The lower beam rests on a Winkler-Pasternak foundation and is coupled to the upper beam by two elastic systems, modeled as double spring-mass. The problem is analyzed within the framework of the Euler-Bernoulli beam theory. The main objective of this study is to determine the natural frequencies of free vibration and the vibration modes associated with this complex structural configuration. The results obtained for the elastically coupled system are validated by comparison with those of a simplified configuration featuring a single double spring-mass elastic coupling, as reported in the literature. The adopted methodology is based on subdividing of the structure into beam segments, followed by the application of boundary and compatibility conditions. The system thus formulated is then solved using the iterative Newton-Raphson method, which finds the natural vibration frequencies of the vibrations and, consequently, the associated vibration modes. This approach provides an accurate analysis of the vibration phenomena within the dynamic modeling of the structure.

The paper investigates the problem of constructing a system of models for an object with many input and output variables for the case of modeling the dependence of mechanical properties of a metal casting on the chemical composition of the raw materials. A technique for constructing system models for multidimensional objects from given experimental data is describedThe initial data set contains numerical values of five input variables (carbon, silicon, manganese, chromium, and phosphorus) and four output variables (tensile strength, relative elongation, impact strength, and hardness) characterizing the iron casting. Based on these data, the system of models is obtained describing the dependence of the output vector of casting properties on the input vector of its chemical composition and allowing to model these properties instead of conducting expensive experiments. To construct the system model, a technology of building models of multidimensional objects is applied based on the combinatorial GMDH algorithm allowing automatically derive linear or nonlinear models from a set of experimental data. As a result of modeling, a nonlinear system model of optimal complexity is obtained that allows explaining the dependence of the casting properties on chemical composition and the interdependence of these properties between them. This system model can be used for choosing the relevant chemical composition of raw materials for the specified physical and mechanical properties of the iron casting.

The goal is to analyze the existing methods of moving a geometric model and to improve a macroparametric approach for moving a geometric model between specific CAD packages. The subject of the article’s research is the technology of object movement using macro tools.The method is to automate the macro-parametric approach. To transmit parametric information, including the construction tree, a set of standard commands used as a neutral format is defined. A macro file is exchanged that records the sequence of simulation commands or the user’s simulation history. The history of commands that define a high-level dynamic interface is written to a macro file that is used to exchange the static model. The set of commands issued by the CAD developer during the design task is recorded as a simulation history, which implicitly includes the CAD developer’s intent. The results. When analyzing the available methods of moving the geometric model, the degree of accuracy of object transfer was analyzed. The software for implementing the macro parametric approach was developed. The effectiveness of the development can be evaluated by measuring the accuracy of reproducing the geometric model in the source system. That is, achieving maximum correspondence between the construction tree in the input CAD package and the construction tree in the output CAD package. The scientific novelty of the results obtained is as follows: flexible, affordable, and widely scalable software has been developed for transferring the construction tree between different CAD packages. The practical significance. The results of the analysis will make it possible to transfer the geometric model more quickly and accurately.

The article analyzes the results of calculating forces and displacements in a planar frame under static loading, taking into account the physical nonlinearity of reinforced concrete. A comparison of deformation theories for composite materials is presented, along with a linear calculation and an investigation of the results of a nonlinear calculation performed in the SCAD Office software package (version 23.1.1.1). Additionally, a frame calculation was carried out in accordance with the recommendations of DBN V.2.6–98:2009 [1], which considers the nonlinear behavior of reinforced concrete using reduction factors. Significant discrepancies were identified between the results obtained from physical nonlinearity theories and those from normative calculations. These deviations emphasize the need for further verification studies to increase the accuracy of calculations of the stress-strain state in reinforced concrete structures, conducting experiments with the combination of different types of reinforcing frames and reinforcement material. The obtained results will indicate the effect of the modulus of elasticity of reinforcing materials on the change in the stress-form condition of the outer layers of concrete.

This paper is devoted to the modeling a collaboration portal architecture and its implementation for the field of environmental initiatives. Given the modern trends in cloud services development, special attention is paid to the modeling capabilities of Microsoft 365 ecosystem. A conceptual, formalized and program models of the collaboration portal for effective support of coordinated work the participants of Erasmus + project from different regions and organizations is proposed. The key structural components of portal that provide flexibility, scalability and security in the process of information exchange and coordination withing project are identified. The performance of the involved modeling techniques is demonstrated both on the realized examples of knowledge sharing, collective decision making, project management and on evaluating the portal effectiveness with using of internal resources of Microsoft 365. In addition, the paper identifies areas for further development of the portal's functionality, focusing on the integration with the model-driven applications that will contribute to the friendliness and sustainable cooperation.

The problem of fast transportation of a large number of passengers in large cities remains relevant at all times. Searching for the optimal combination of urban transport routes for a large city is a time-consuming task that requires an automated method of solution. The main idea of the work is to use the genetic algorithm for multi-route search, because it significantly saves memory and other computer resources in comparison with the algorithm of full selection. Thus, the genetic algorithm allows for search of multi-route combination on a much larger graph. A method of genetic algorithm adaptation for the optimal routes planning is described in the paper. Moreover, in the course of work there was an alternative algorithm for the optimal selection of a transport routes combination for transport network with a small roads number. This algorithm obtains the absolute optimal solution by evaluating all possible route options. In the work, this algorithm was used to evaluate the performance of the genetic algorithm and adjust its parameters.

Agent-based modelling (ABM) has become a powerful tool for modelling and understanding complex systems in various disciplines, including economics, biology, sociology and others. To create, analyse, and experiment with agent-based models, there are universal platforms – AOM environments – that provide researchers even without a mathematical background with the opportunity to work with agent-based models. This article introduces the concept and basic principles of AOM, provides an overview of the most commonly used agent-based modelling environments, and describes the application of AOM to the model of content distribution in social networks of cyberspace, which are the most widely used environment for information influence on society today. The features, capabilities and advantages of these environments are investigated, and an agent-based model is built in the NetLogo simulation environment. The key principles of AOM were presented, which directly demonstrate the advantages of its application for modelling and analysing complex systems such as social networks.

The article considers the reliability as one of the key characteristics of the software quality assurance. The analysis of the principles of taking into account secondary faults of software tools is carried out. The main provisions of the works are considered in which the parameters of imperfect debugging are investigated. The principles of the theory of dynamics of software systems are considered in view of account of secondary faults. The application of the theory of time series and simulation modeling for estimating the number of secondary faults of software is considered. The examples of different approaches to the modifications of the risk functions of software reliability evaluation models for the possibility of taking into account secondary faults are given on the example of the Jelinski-Moranda model. The possibility of combining modified models for assessing the reliability of software tools is analyzed in order to obtain an analytical expression for identifying the number of secondary faults. The application of exponential approximation and second-degree polynomial approximation of fault trend for estimating the number of secondary faults is considered. The analysis of advantages and disadvantages of the principles of taking into account secondary faults is carried out.

The construction projects’ data collected using IoT devices is characterized by its high velocity as being streamed continuously and its varying accuracy and reliability due to possible sensor malfunctions, transmission errors and other factors, so AI technologies-based system must be able to preprocess the collected data. The use of the wrapper technique to filter out irrelevant and redundant information has two associated problems: a high computation costs and a risk to receive an overfitting learning model. In this paper the focus is put on a design of a mathematical model to be used with a search strategy in a wrapper technique to help to reduce a computation cost and to prevent a classification machine learning model to overfit. To meet the goal in the research were used the methods: data analysis and visualization; mathematical modelling, structuring by algorithm; experimental tests. The results of the experiments are analyzed and in the conclusions is recommended to use this study findings to improve the classification machine learning strategy.

The article focuses on the importance of cybersecurity in the energy sector, emphasizing the risks associated with the integration of technologies into energy infrastructures. The authors point out the increasing number of cyber-attacks and vulnerabilities faced by infrastructure networks, especially in the context of the growing number of connected systems and the proliferation of the Internet of Things. Current trends and challenges faced by organizations in the energy sector are highlighted, including risks related to remote access, wireless technologies, and cloud computing. The primary emphasis is on the analysis of threats and vulnerabilities in intelligent networks, risk assessment, and the development of strategies to enhance cybersecurity. The authors discuss various types of cyber threats, including viruses, espionage, DDoS attacks, phishing, system failures, and others. They emphasize the importance of developing effective mechanisms for detecting and responding to attacks, as well as the need to create monitoring and event analysis systems for timely responses to potential threats. The article also presents various methodologies for cybersecurity analysis, including system architecture research, threat modeling, vulnerability and attack identification, calculating the probabilities of exploiting vulnerabilities, modeling malicious behavior, and risk assessment. Advanced practices and technologies in the field of cybersecurity are also discussed in the article, including data encryption, biometric technologies, artificial intelligence, and machine learning for predicting and detecting cyber-attacks.

One of the important tasks of the municipality is maintaining and developing the city's road and transport network. The transport network must adhere to a number of important requirements, including the safety of all road users. Direct supervision of road traffic safety is carried out by employees of the National Police Department. However, decisions regarding optimizing the road and transport network are made by the relevant municipal departments.To make adequate management decisions to ensure safety when planning the development and optimization of the existing road network, up-to-date statistical information on the road traffic safety status is required. Such information has spatial references or the possibility of classification according to different categories, or a combination thereof. The task of comprehensive presentation of information on traffic accidents was solved with the help of spatial analysis and modeling during the development and implementation of the traffic geo-information system of the city of Chernihiv within the municipal authorities. The geoinformation system is built on the QGIS open platform, which facilitates further modernization and functional expansion of the software modules.Currently, the traffic accident analysis module enables the accumulation of statistical information about traffic accidents within the city limits. It classifies them according to various criteria with appropriate cartographic visualization within any specified time interval for analysis. The module also displays the locations of injuries and fatalities resulting from traffic accidents. In accordance with the accepted national methodology, it identifies areas of traffic accident concentration, which contributes to the adoption of organizational measures to reduce accidents.Further development of the traffic accident analysis module involves the development of a methodological framework for analyzing traffic accidents and automating data entry procedures which will be achieved by utilizing the “Safe City” visual surveillance system, which is already implemented and in use in the city of Chernihiv.

The article presents an approach to the strategic analysis of information security of transport and logistics centers (TLCs). The proposed method is based on expert assessments of the factors that determine the level of TLC IS risk. Strategic analysis is the basis of a decision support system for ensuring an appropriate level of TLC security. It is based on the tools of fuzzy set theory: Linguistic assessments of the factors influencing the information security risk using triangular membership functions; Fuzzy Analytic Hierarchy Process (AHP) method to determine the fuzzy importance coefficients of the factors and subfactors of the proposed model; Center of Area (CoA) method to defuzzify fuzzy triangular values; Fuzzy Simple Additive Weighting (SAW) method to rank the TLC IS risk factors and select an effective strategic solution.For the strategic analysis of the TLC IS status, an extended and modified informative matrix was used to assets the status of the Wilson system. The developed model for assessing the level of TLC IS risk will ensure the implementation of effective strategies to ensure it. The model is implemented in the Python programming language. Further research is seen in the implementation of the proposed model in the decision support system for the management TLC IS.

This study presents a novel probabilistic framework for analyzing path length distributions in Merkle Patricia Tries, the foundational data structure of Ethereum’s state management system. We develop a mathematical model characterizing the probability distribution of path lengths for randomly generated Ethereum addresses and validate it through extensive computational simulations. Our findings reveal the logarithmic growth of average path lengths with respect to the number of addresses, confirming a critical property for Ethereum’s scalability. The study demonstrates remarkable accuracy in predicting average path lengths, with discrepancies between theoretical and empirical results remaining below 0.01 across tested scales from 1000 to 300 million addresses, encompassing Ethereum’s current state size. We identify and confirm the right-skewed nature of path length distributions, providing essential insights into worst-case scenarios and informing optimization strategies. Statistical analysis, including chi-square goodness-of-fit tests, strongly supports the model’s accuracy. This research bridges a significant gap between theoretical computer science and practical blockchain engineering, offering immediate applications for Ethereum client optimization, stateless client development, and informed protocol upgrade decisions.

The research method presented in this article is a presentation of a method for processing a program model that is described using predicates. Such describing of software models transforms in form of non-deterministic finite automatons, where each state of the automaton describes a series of actions in the form of symbolic data of indefinite length, which at the analysis step must be processed and transformed into a series of equivalent statements of the target programming language. This article describes the application and description of the stack algorithm for the syntactic processing of each state of the model. Analysis of the description according to this algorithm is more natural for the analysis of ordinary languages. Unlike known top-down and bottom-up parsing techniques, the stacks algorithm determines and parses pairs of lexemes such as a data lexeme and an action lexeme (such it`s used subject and predicate in ordinary languages).

In the fast-paced field of information technology, continuous evaluation and improvement of professional competencies are crucial for maintaining competitiveness. This paper presents an innovative conceptual model designed to objectively assess the knowledge and skills of IT professionals, with a specific focus on UX/UI designers. The system incorporates advanced machine learning algorithms, such as Long Short-Term Memory (LSTM) networks, the Text-to-Text Transfer Transformer (T5) architecture, and the GPT-4 model, to automate the process of generating test questions and evaluating competency levels. Additionally, the concept system leverages fuzzy logic methods to enhance the assessment of qualitative and subjective competencies, particularly in UX/UI design tasks. By treating evaluation criteria as linguistic variables with corresponding membership functions, fuzzy logic enables the transition from qualitative to quantitative assessments. This approach allows for more flexible and precise evaluation of competencies, addressing the inherent uncertainty in creative and design-related tasks. A key feature of the system is the development of a competency matrix that includes design skills, research abilities, and soft skills, enabling a comprehensive assessment of a specialist’s qualifications.The study details the architecture of the assessment model. The methodology supports adaptive learning by continuously refining the competency evaluation process based on feedback from test results. Additionally, the system’s potential for integration into educational and professional development programs is examined, with promising implications across various IT sectors.The conceptual model as a whole improves the learning process and general professional development of IT professionals. Further research will focus on further improving the algorithms and expanding the application of the system to additional IT specializations.

This paper demonstrates the effectiveness of deep learning algorithms compared to traditional mathematical and machine learning algorithms in forecasting stock price volatility. Specifically, a Long Short-Term Memory (LSTM) network, representing deep learning approaches, is evaluated against ARIMA and GARCH mathematical models, as well as Support Vector Machines and Random Forest machine learning algorithms. The research aims to identify the strengths and limitations of each model in terms of prediction accuracy and efficiency by analyzing daily stock prices and trading volumes of selected companies. The findings suggest that the LSTM model offers significant advantages in capturing the complex and dynamic nature of financial markets, providing more reliable forecasting tools for investors and policymakers. The study also highlights the trade-off between accuracy and computational requirements, concluding that the GARCH model may be a suitable option when accuracy is not the primary factor, and computation speed is of utmost importance.

There is a considerable amount of research currently focusing on examining different types of degradation processes that occur within technical systems. In most cases, these processes are analyzed separately, and their combined effect on the system's overall residual life is reduced to selecting the most intense process, often referred to as the “weakest link”. This approach significantly simplifies all aspects of every degradation process, which, in turn, can lead to overestimated forecast results. The article presents a method for estimating an object's residual life under the influence of several concurrent degradation processes, each with different impact proportions and occurring at different temperatures. The method involves determining the average rate of the generalized degradation process and its coefficient of variation using a probabilistic-physical model with the Diffusion Monotonic failure distribution. The DM-distribution is formalized based on a diffusion-type Markov process with a fixed rate that approximates statistical failure data in technical systems. Considering the concurrent influence of several degradation processes and different temperature conditions, the contribution proportions of each process to the generalized degradation have been determined and applied. This method provides a more accurate assessment of an object's overall residual life compared to methods that consider only the dominant processes. Accurate estimation of residual life values will help minimize operational expenses by adjusting the frequency of technical inspections and providing a proper assessment of the examined objects’ service life.

This paper introduces NeuroPhysNet, a novel hybrid neural network architecture designed for enhanced prediction and control of cyber-physical systems (CPS). The increasing complexity and dynamic nature of CPS present significant challenges in modeling and control, often stretching the limits of traditional approaches. NeuroPhysNet addresses these challenges by seamlessly integrating deep learning techniques with physics-based modeling, offering a robust framework that leverages both data-driven insights and domain-specific knowledge. The architecture comprises three main components: a Deep Neural Network (DNN) module for capturing complex non-linear relationships, a Physical Model Integration (PMI) module for incorporating domain-specific equations, and an Adaptive Fusion (AF) module for dynamically combining outputs based on current system states. We evaluate NeuroPhysNet across multiple CPS domains, including smart grid energy management, autonomous vehicle control, and industrial process control. Results demonstrate significant improvements in prediction accuracy, physical consistency, and generalization capabilities compared to state-of-the-art baseline models. NeuroPhysNet achieves up to 15% improvement in mean squared error and a 20% increase in physical consistency scores across tested domains. This research contributes to the growing field of physics-informed machine learning, offering a versatile approach for enhancing the performance and reliability of complex cyber-physical systems.

This paper discusses a method for estimating of the level of pulsations at the output of discrete power converters applicable to on-board electric drive systems with high control accuracy. It is proved that the estimation of the level of pulsations at the output of discrete power converters based on the fundamental harmonic of “smooth” trigonometric Fourier series functions is not accurate and presents a new method of pulsation analysis based on the fundamental harmonic is compared with the peak-to-peak analysis of harmonics based on the analysis of the discrete transient process at the output of the converter. On the basis of the discrete Laplace transform, a methodology for estimating peak-to-peak ripples for uncontrolled and controlled power converters has been developed, taking into account the form of their output voltage and the process of pulse-width modulation. The results presented in the paper are of great practical importance and allow for a more accurate estimation of the output pulsation level in order to improve the quality of control in on-board electric drive systems.

With the growth of popularity of remote education, the need for fully remote laboratories for technical specialties arises. This need is especially pronounced in the field of hardware engineering, where it's important for students to gain practical knowledge and skills by interacting with real-world physical equipment like power electronics circuits, FPGA boards, microcontrollers, SDRs, and other devices. Some core features and subsystems like authentication and authorization of almost every remote laboratory system are implemented out of the box in specialized Remote Laboratory Management System frameworks like LabDicoveryEngine. This article describes a back-end architecture for a power electronics laboratory built using the LabDiscoveryEngine remote learning management system. The proposed architecture utilizes Node.js technology for the laboratory server implementation, enabling efficient bidirectional communication between client applications and hardware components through custom application protocols. The solution includes detailed multi-layer architecture design, hardware component integration, and protocol specifications for client-server and hardware interactions. This approach aims to improve the quality of remote education while providing a flexible foundation that can be adapted for other types of remote laboratories.

The paper focuses on studying and designing of one type of dynamical systems with nonlinear restoring force. Our research revolves around the well-known Duffing equation, which describes body motions in two coupled potential wells. The Duffing equation considers that these motions are influenced by the initial speed and position of the moving body, as well as externally driven harmonic oscillations. In our paper, we propose to view the dynamical system defined by the Duffing equation as consisting of two subsystems. One subsystem defines body position and speed, while the other defines the restoring force. By applying the symmetry approach to these subsystems, we can design a generator of restoring force that is driven by the position of the moved body. Our studies demonstrate that using a controlled restoring force allows us to influence the overall system dynamics. When the gain, which determines the relationship between the body and the generator, has a positive value, the body motions are damped, whereas contrary motions become less stable with a negative gain. We explore system motions in normal and canonical spaces to showcase the potential for various implementations of the studied dynamical system. Additionally, we demonstrate the implementation of the systems in normal and canonical spaces by transforming their descriptions into a discrete-time domain and utilizing the Arduino Due demo board.

An analytical representation of operation principles of boost-type quasi-Z-source (QZS) dc-dc converter is obtained by determining of commutation intervals (where semiconductor switches keep their fixed states) inside a single operation period. An analytical model of converter is obtained by composing of nodal admittance matrix (NAM) by using of Kirchhoff’s voltage law and written-up in Laplace-domain by using of operational calculus method applied for each commutation interval. Finding the inverse nodal admittance matrix with subsequent inverse Laplace transform in symbolic form allows obtaining a set of time-domain analytical equations for applied voltages and circulating currents inside converter at each commutation interval in a steady-state mode. A set of borderline conditions, allows determining a switching moment from one commutation interval to another, and an interval switching sequence depending on load impedance and control signal parameters, is formulated. Two above milestones together allow obtaining of converter operation laws as a set of piecewise functions for voltages and currents at the whole operation period. Analytic form allows evaluating the influence of parasitic parameter onto converter operation, analytically justifying the connection topology and impedance character of various snubber and gain-enhanced networks, and, moreover, opens up opportunities for analytically constructing the optimal control loops and control correction networks for stability improvement and operation range extension.

To carry out the search mission, an additional narrow-angle video camera is installed on board the unmanned aerial vehicle (UAV). This video camera can be re-aimed at objects detected in the image frame received from the main (navigation) video camera. The movement of the additional video camera is carried out by two coordinated electric drives using direct current motors or brushless direct current (BLDC) motors.Ensuring the required accuracy and speed of two-coordinate positioning of an additional video camera is possible through a well-designed electric drive control system. The article presents step-by-step results of simulation modeling, starting with the requirements for a two-coordinate positioning system for an additional UAV video camera. Simulink was used as a modeling tool. The simulation is based on advanced models of three-loop DC motor control systems as well as a brushless DC motor. For both possible drives, the transient processes of the winding current, speed and angle of rotation of the platform with an additional video camera of the UAV were obtained.Analysis of the results of simulation modeling allows us to determine the influence of parameters, clarify the permissible ranges of their change and reduce the number of errors when choosing technical solutions and individual components during further physical modeling of the system.

The paper examines possible directions for improving power quality assurance tools by achieving their multifunctionality, controllability, and efficiency, with the aim of enhancing their performance while considering the circuit and operational specifics of integrated ship electrical power systems with semiconductor propulsion complexes. To address the challenges posed by increasing harmonic distortions and reactive power in such systems, one- and two-stage versions of system-controlled filter-compensating devices are proposed and thoroughly investigated. These devices are specifically designed to enable full reactive power compensation while significantly reducing voltage and current harmonic distortions, thus ensuring improved power quality and system reliability in ship electrical power systems with semiconductor propulsion complexes. The proposed modeling approach provides a comprehensive framework for analyzing and comparing the impact of controlled filter-compensating devices on the integral harmonic distortion coefficient of voltage across systems with various types of input rectifiers incorporated within the frequency converters of semiconductor propulsion complexes. Additionally, the study takes into account possible deviations in the ship network frequency, which can influence system performance under real operating conditions. The results demonstrate the effectiveness of the proposed tools in mitigating power quality issues, thereby contributing to the overall stability and efficiency of modern ship electrical power systems.

The synthesis of statistical fractal models is justified for Spectral Data Processing. Many spectral sensor channels allow the use of methods for identifying distributions of spectral intensities and Rényi entropy, justifying the applicability of statistical fractal methods (multifractals) for modeling and subsequent interpretation of data, particularly plant responses to environmental changes. The derivation of multifractal models based on the Renyi information entropy is presented. The use of two multifractal measures of variability is substantiated. Multifractal models of hyperspectral characteristics of plants growing over oil deposits are presented. Examples of parameter computation of multifractal models based on real plant spectrograms are considered. Based on these results, measures of Rényi spectrum variability and multifractal spectra are determined, forming an indicative sign of oil deposit boundaries; prospective areas for oil development are forecasted. The discussed methods of computer multifractal modeling significantly expand the range of tasks addressable by established methods of spectral analysis based on spectral libraries. It has been observed that when studying complex self-organizing systems with fractal structures and power-law distributions, the application of Gibbs-Shannon entropy does not align well with observed phenomena. Describing systems with such properties is more appropriate using methods based on Rényi entropy.