OKOleksii Kudin

Oleksii Kudin

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article


Oleksii Kudin has a total of 18 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    Discrete models of geometric objects in parallel computing systems

    by Anastasiia Kaliuzhniak, Oleksii Kudin

    The usage of powerful computer systems defines computational experiments as novel and significant research methods enabling to solve complex. Taking into account the fact that raditional analytical methods for visualizing mathematical models are of a determinative nature, there is still a need to apply modern mathematical theories which in their turn will enable to expand the possibilities of applied mathematical research. The object of the research is the process of developing functional design of complex geometric models. The subject of the research is the usage of parallel methods for constructing the surfaces of discrete models of some geometric objects. The methods of the research are: application of the apparatus of analytical geometry, mathematical analysis, the theory of R-functions, parallel architecture and numerical methods. The purpose of the research is to solve current scientific and technical problems, in particular to increase the efficiency in the design of the programme for building discrete models, which can be implemented in the finite element analysis of complex technical systems by means of using parallel architecture. The following tasks were set to achieve the goal: analysis and review of currently known methods and approaches related to the construction of discrete models in complex computing systems. Development of the appropriate method and visualization of mathematical models based on the functional approach. Modification of the "Marching Cubes" method. Implementation in parallel architectures resorting to modern technologies and programming libraries, such as OpenMP and MPI, and conducting test experimental calculations that prove the efficiency of the proposed algorithm. Considering the importance of geometric model accuracy for the safety of complex technical systems, the application of parallel methods for building discrete models can significantly impact the reliability and safety in the development and testing of high-tech products, particularly in fields such as rocket engineering. In the process of creating models used for designing rocket systems and other advanced technologies, it is crucial to achieve high precision, computational speed, and reliability of results. The developed parallel computation methods help reduce the risk of errors and optimize the design processes, which is vital for ensuring safety in this critical area.

  • Open Access

    HOMOGENIZATION OF A TRANSTROPIC COMPOSITE WITH RANDOM PARAMETERS OF A HOLLOW FIBER

    by O. H. Spytsia, S. I. Homeniuk, O. V. Kudin, A. V. Stoliarova, О. Г. Спиця, С. І. Гоменюк

    When designing structures made of composite materials, a composite model is used in practice as a continuous homogeneous medium. Its physical and mechanical properties as a homogeneous material reflect the effective constants. Of course, the parameters of the components of the composite are random variables. Therefore, its effective constants are also random variables.In this regard, the actual task is to homogenize the composite taking into account the stochasticity of its parameters, including the determination of the characteristics of elastic constants as random variables.In the work, the object of research is a unidirectional fiber composite, the phases of which are a transversal-isotropic matrix and hollow fiber. The radius of the fiber and the cavity in it are modeled as random variables with a normal distribution law. Accordingly, the effective parameters of the composite are also random variables. The method of determining the mathematical expectation of an effective longitudinal modulus of elasticity, which can be used to find other numerical characteristics of effective parameters as functions of random variables, is proposed.To solve the homogenization problem, the composite is modeled by a continuous homogeneous transversal-isotropic material. The conditions of perfect bonding are fulfilled at the contact surface between the matrix and the fiber. As result of the research, a formula was obtained for the effective longitudinal modulus of elasticity as a function of random variables. The volume fractions of the fiber and the cavity in the fiber for the composite are random. The distribution densities of these random variables are found. The elastic constants of the fiber and matrix are considered as deterministic variables. The obtained expression of the effective longitudinal modulus of elasticity as a function of random variables allowed us to determine its characteristics – mathematical expectation and variance. The obtained distribution densities of random arguments can be used to determine other moments of this random variable.The proposed method can be used to solve the problem of homogenization of composites, the phases of which have random characteristics with other continuous distributions different from the normal one. Під час проєктування конструкцій з композитних матеріалів на практиці використовують модель композита у вигляді суцільного однорідного середовища. Фізико-механічні його властивості як однорідного матеріалу відображають ефективні сталі. Звичайно, параметри складових частин компонентів композита є випадковими величинами. Тому його ефективні сталі теж є випадковими величинами. У зв’язку з цим актуальною задачею є гомогенізація композита з урахуванням стохастичності його параметрів, у тому числі визначення характеристик пружних сталих як випадкових величин.У роботі об’єктом дослідження є односпрямований волокнистий композит, фазами якого є трансверсально-ізотропні матриця та порожнисте волокно.Радіус волокна та порожнини у ньому моделюються випадковими величинами, що мають нормальний закон розподілу. Відповідно, ефективні параметри композита також є випадковими величинами. Запропоновано методику визначення математичного сподівання ефективного поздовжнього модуля пружності, яку можна застосувати для знаходження інших числових характеристик ефективних параметрів як функцій випадкових величин.Для розв’язання задачі гомогенізації композит моделюється суцільним однорідним трансверсально-ізотропним матеріалом. На поверхні контакту між матрицею та волокном виконуються умови ідеального з’єднання. У результаті виконаного дослідження отримано формулу для ефективного поздовжнього модуля пружності як функції випадкових величин. Випадковими є об’ємні частки волокна та порожнини у волокні для композита. Знайдено щільності розподілу цих випадкових величин. Пружні сталі волокна та матриці розглядаються як детерміновані величини. Отриманий вираз ефективного поздовжнього модуля пружності у вигляді функції випадкових величин дозволив визначити його характеристики – математичне сподівання та дисперсії. Отримані щільності розподілу випадкових аргументів можна застосувати для визначення інших моментів цієї випадкової величини. Запропоновану методику можна застосувати для...

  • Open Access

    Developing of neural network computing methods for solving inverse elasticity problems

    by Anastasiia Kaliuzhniak, Oleksii Kudin, Yuriy Belokon, Dmytro Kruglyak

    This paper examines the use of neural network methods to solve inverse problems in the mechanics of elastic materials. The aim is to design physics-informed neural networks that can predict the parameters of structural components, and the physical properties of materials based on a specified displacement distribution. A key feature of the specified neural networks is the integration of differential equations and boundary conditions into the loss function calculation. This approach ensures that the error in approximating unknown functions has a direct impact on optimizing the network's weights. As a result, the resulting neural network approximations of unknown functions comply with the differential equations and boundary conditions. To test the capabilities of the designed neural networks, inverse problems involving the bending of plates and beams have been solved, focusing on determining one or two unknown parameters. Comparison of predicted and exact values demonstrates high accuracy of the constructed neural network models, with a relative prediction error of less than 3 % across all cases. Unlike analytical methods for solving inverse problems, the primary advantage of physics-informed neural networks is their flexibility when addressing both linear and nonlinear problems. For instance, the same network architecture can be employed to solve various boundary-value problems without modification. Compared to classical numerical methods, the parallelization capability of neural networks is inherently supported by modern software libraries. Therefore, the application of physics-informed neural networks for solving inverse elasticity problems of plates and beams is effective, as evidenced by the achieved relative errors and the computational robustness of the method. In practice, the proposed solution can be used for relevant calculations during the design of structural elements. The developed software code can also be integrated into automated design systems or computer algebra systems

  • Open Access

    Designing an object-oriented architecture for the finite element simulation of structural elements

    by Oleksii Hniezdovskyi, Oleksii Kudin, Yuriy Belokon, Dmytro Kruglyak, Sergii Ilin

    This paper reports the development of an architecture and software implementation of the library of classes for the finite-element analysis of problems in the theory of elasticity with an open-source code. The practical necessity of such systems is due to the fact that in modern equipment there are new types of materials whose structural elements' calculation has certain features. As a result, it is necessary to update the relevant scientific software or even devise a new one. A flexible software architecture is designed to reduce the time and complexity of such updates. Existing implementations of the method of finite elements with open source have been analyzed: it was revealed that there are no systems aimed at the most flexible and user-friendly architecture. The system of abstract classes proposed in the current work corresponds to known SOLID principles of object-oriented design and makes it possible to scale the already developed analysis program for new tasks in an easy and understandable way. To test the quality of the developed system from the point of view of software engineering, the maintainability index and cyclomatic complexity code metrics were used. The values of these metrics for the modules of the PyFEM system core vary in the following ranges: from 1 to 18 for the maintainability index, and from 22 to 100 for cyclomatic complexity. PyFEM testing was performed on the task of determining the stressed-strained state of the turbine rotor blade. Due to the ease of implementation, it was possible to build a set of effective and intuitive classes that make it possible to solve numerically the static and dynamic problems in the theory of elasticity. The developed class library can be used in the development of both universal and specialized software designed to analyze multiphysics problems.


  • Open Access

    The homogenization of multi­modular composites at their longitudinal deformation

    by Sergii Grebenyuk, Tetiana Smoliankova, Mykhailo Klymenko, Oleksii Kudin

    The homogenization of multi­modular composites at their longitudinal deformation

  • Developing a Deep Learning Sound Classification System for a Smart Farming

    by Oleksii Kudin, Anastasiia Kryvokhata, Vitaliy Ivanovich Gorbenko

    The Internet of Things (IoT) and Machine Learning (ML) are a promising technologies for automation in the different domains e.g. Smart Farming. Such systems could be used for determining noise emission, animal detection and classification, monitoring states of bees in beehive and etc. As rule the characteristics of that systems are some controversial: compactness (deploying on the single-board computers) and real time classification and prediction. Therefore, the software for IoT and ML applications should be sufficiently efficient and not demanding on large computing resources. This work are focused on machine hearing system allows to classify natural sounds in the Smart Farming software application. The original acoustic signal is converted to the frames of a certain length. Receiving a compact representation of the acoustic characteristics of a signal is the aim of the feature extraction stage. This stage exploits special coefficients such as Zero-crossing rate, Spectrum shape, Short-Time Fourier Transform and Mel-frequency cepstral coefficients. Audio classification traditionally involves such machine learning methods as K-means, support vector machine (SVM), decision trees etc. Deep neural networks can be used on both raw acoustic signal and features extracted from it. The accuracy estimation stage deploys quality assessment methods. The data preprocessing stage includes calculating of mel frequency cepstral coefficient (MFCC) for the giving sound files. This approach allows to unify and simplify the sound files presentation in the memory. Further MFCC arrays are feeding to the convolutional neural net. It is important at this stage to configure the network optimally for the most compact storage. This is due to the need to use platforms such as Raspberry Pi to deploying neural network. For neural network software implementation Python library Keras is used. For data preprocessing Python library Librosa is used. The hyperparameters of the neural network have been defined in computational experiments and the optimal combination is two Conv2D layers and three Dense layers. The accuracy of model predictions for each class has been examined by using a confusion matrix and the satisfactory classification accuracy has been defined. It has been established the convolutional neural network ensemble built to solve the problem of acoustic data classification is quite effective. The accuracy of the model on the test data is 95%. It should be noted that the neural network structure with only two packets of convolution-activation-sub-sampling layers was sufficient to solve this problem. The system can be used in Smart Farming applications to filter out unnecessary sounds.


  • Open Access

    Development of a method for triangulation of inhomogeneous regions represented by functions

    by Serhii Choporov, Serhii Homeniuk, Sergii Grebenyuk, Oleksii Kudin

    Development of a method for triangulation of inhomogeneous regions represented by functions