Operation mode modeling of the power supply system for an electric centrifugal pump plant of an oil well cluster under conditions of power supply voltage unsinusoidality provided a comparative evaluation of electric power quality indicators: the total coefficient of harmonic components of voltage K and current K (coefficients of sinusoidality curve voltage distortion and current). Active power losses in the grid with and without downhole reactive power compensator are evaluated. The modelling data agrees closely with the experimental data and characteristics of the used devices declared by the manufacturers. Installation of the reactive power compensator does not cause noticeable distortions of the supply voltage in the system created by the control station. The value of the total coefficient of voltage harmonic components of K compensator does not exceed the acceptable levels and makes less than 5%. Installation of downhole compensating device (uncontrolled) in conditions of supply voltage unsinusoidality cuts down active power losses, thus reducing power consumption by downhole equipment.

The study presents a combined machine learning model for predicting plant traits. The model combines a pre-trained convolutional neural network (CNN) with auxiliary variables represented as comma- separated values (CSV). The aim of the study is to improve the accuracy of plant trait prediction using both visual information and contextual data.

The model is trained on a dataset consisting of 9,147 plant images and the corresponding 167 auxiliary variables. The images are preprocessed using three pre-trained convolutional neural network architectures (InceptionV3, ResNet, VGG19). The supplementary data are combined using a concatenate layer after directing the outputs of the above CNNs. The model is optimized using the Adam algorithm and evaluated with appropriate metrics.

The results show the superiority of the combined model over the baseline CNN in predicting plant traits, confirming the effectiveness of using supplementary data to improve performance. The study demonstrates the potential of hybrid machine learning models to analyze plant data from users around the world.

The purpose of the paper is to design and implement an efficient and affordable automatic surveillance system, which can operate in real time and integrate with existing surveillance systems. The method presented includes analyzing existing solutions on the market, selecting and training a profound learning model for objects detection, developing user interfaces, containerizing of the application and testing the system in a real-world environment. The result is a system capable of detecting objects of interest in real time using neural networks and notifying the user of the detected items. This system is designed for public places like airports, railway stations, schools, and other institutions needing enhanced security

The article describes a comparative analysis using the following quality assessment criteria: system load, RAM consumption, time spent on training, RMSE (Root Mean Square Error), MAE (Mean Absolute Error), FCP (First Contentful Paint) and MSE (Mean Square Error). Using mathematical decision-making methods, the most practical algorithm was selected, a system based on DLRM (Deep Learning Recommendation Model) architecture, which showed the best results in terms of accuracy and fl exibility in processing large amounts of data, despite having high resource intensiveness. The implementation of the selected algorithm included the development of the algorithm itself, database design, creation of a graphical user interface and development.

In the 70s–80s of the last century, scientists of Kazakhstan working in the institutes of the Academy of Sciences conducted a study of potential applications of the residue number system in conjunction with a ternary number system. Their aim was to develop specialized devices for data processing. As part of that study, they sought to create an efficient way to represent complex numbers and quaternions in specialized formats. The use of this combination made it possible to reduce information redundancy when displaying modular vectors in digital registers that use a ternary element base.

In modern world, cryptography has become an essential part of the information security. With the constant development of technologies and increasing volumes of transmitted data, the necessity of their reliable protection becomes extremely urgent. This article is devoted to a comparative analysis of modern symmetric encryption algorithms. The comparison criteria are security, complexity, performance, applicability, and standardization.

The integration of modern information systems and digital technologies in medical institutions helps to reduce costs, improve the quality of medical services, expand their accessibility, and save time for doctors. We give special attention to the rehabilitation of patients with neurological disorders during outpatient treatment, requiring careful monitoring and adjustment of therapeutic and restorative procedures. Existing solutions mainly rely on specialized equipment that is either not supported by the Russian Federation or does not provide the necessary functionality for medical personnel. The aim of this study is to identify practices that can improve the process of diagnosis and monitoring of neurological patients by developing methods and algorithms for supporting neurological patients that comply with Russian standards (GOST) and regulatory legal acts.

The article reviews current neurologist diagnostic methods and tools for automated patient monitoring during outpatient treatment. The analysis produced a system description and the development of functional and non-functional requirements, including safety, reliability, user-friendliness, and compatibility with other medical systems. The NeuroDom system, for example, enables remote patient monitoring and support through a web application that integrates with wearable and mobile applications.

The results of the study show that information systems ensure more accurate diagnosis, increase the efficiency of patient monitoring during rehabilitation, and enable adjustments to the prescribed treatment, ultimately leading to an improved quality of life for neurological patients and enhancing the efficiency and accessibility of medical care during outpatient treatment.

The ratios for calculating the signal level scattered by a plastic object during subsurface radar sensing are obtained. Radar images of plastic discs buried in the sand are shown. Researchers performed the analysis of the connection between the radar contrast of the plastic disk image and the system’s energy potential. The article presents a procedure for measuring the orientation of a subsurface pipe using radar. This article describes a subsurface sounding radar design using resonant signal compression.

Abstract. The article provides an overview of mathematical models for the placement of process equipment. These models are used to solve problems of designing and upgrading production shops. Cases of placement of interconnected equipment are considered. A mathematical model for the placement of production lines and process equipment on them is constructed. The lines are connected by a vertical viaduct through which communications are carried out. The model designates prohibited zones on the lines where no other elements can be placed. The results of numerical experiments using the constructed nonlinear programming model and the CPLEX package are presented. The study considers two series of test examples: one with a fixed viaduct location and another determining its optimal placement. A better value of the objective function was obtained when searching for the location of the viaduct, but the calculation time increased. The model can be used in the design and assessment of the process lines placement with equipment in industrial sectors.

The article considers the creation of a mathematical process component for transfer medium heating in a model hot water boiler KVGM. To solve this problem, an analytical approach is used to build the model, which minimizes the use of empirical coefficients in the calculation scheme. Researchers created an original analytical model of the transfer medium heating in the furnace and convection chamber, accounting for heat radiation according to the Stefan-Boltzmann law. The created mathematical model is expected to be used as a necessary component of a computer simulator.

The article considers the scope of application of control algorithms for a multicomponent dynamic system with uncertain parameters varying depending on a non-stable range flow of input data. The proposed methodology is aimed at simplifying the process of analyzing and selecting data ranges with different levels of relevance used to form a control action. A mathematical model describing the connection between data stocks and flows is presented. This allows identifying the key parameters of the system and focus on their further detailing. The assumption of data transition between classes is based on automatic data update, which reflects the real conditions of the system operation. Initially, data classes do not have a strict hierarchy, but their volume changes depending on time intervals, with the most significant changes occurring near the end of the period under study. These changes are due to internal data movements between classes and external data flows. The expected data flow is calculated based on the initial stock and transition probability, which is consistent with the assumption that flows are proportional to stock. The total number of new data is equal to the number of discarded data.

As a result of comparison of standard algorithmic and coding technologies with the proposed method and subsequent evaluation, the system performance improvement as well as predictability of critical states that can occur at any moment in the presence of uncertain parameters are revealed. A software model for reading and interpreting input signals, identifying and correcting artefacts, processing input data, and testing a multi-component dynamic system with range input data has been implemented