Mechanics and Advanced Technologies

Neural Networks and Artificial’s Intelligence’s Algorithms in the Tasks of Diagnostics Foreign Bodies in the Wound Channel of Patients

Oleksandr Salenko — 2026-02-02

The paper considers a device in which information about the presence of a foreign body a fragment in a wound is obtained using a special compact sensor: a mechanical elastic flexible rod, which is placed in a protective soft tube and connected to a sealed chamber, one of the walls of which is a membrane capable of converting the vibrations of the rod into sound waves, which, propagating in the chamber, are recorded by a microphone. Mounted in a convenient housing, the sensor is connected to sound frequency filters (up to 15 kHz) via an amplifying link and allows you to obtain a noise signal, based on which you can detect the contact of the rod with a foreign body during its movement in the patient's wound. The difference in the spectrum patterns suggests that such a device can be used not only to detect foreign bodies, but also to identify them.
The use of neural networks and artificial intelligence algorithms significantly improves the accuracy of foreign body detection, especially for small objects and those whose rheological properties are similar to the patient's tissue. Thus, with the appropriate adjustment of the device, the accuracy of foreign body diagnosis is currently up to 65–95 % for bodies larger than 3.5–5.0 mm; higher accuracy is observed when detecting bodies with pronounced elastic properties.
It has been shown that the use of this device allows obtaining information about the presence of a foreign body (due to the appearance of separate clearly defined signal peaks), while the absence of foreign bodies gives a picture of noise emission distribution close to the Gaussian distribution; about the type of foreign body. The shape of the pattern allows one to assess the type of foreign inclusion (more elastic bodies have more pronounced frequency spikes); to estimate the approximate size of the body: larger bodies are in contact with the probe for a longer time as it moves.

Fabrication and Property Comparison Aluminium TiB2 Composites Manufactured Through Additive Manufacturing and Conventional Manufacturing: a Comparative Review

Rajaneesh Marigoudar — 2026-02-02

Additive manufacturing is most emerging techniques of fabricating components with desired dimensions to the near net shape. Metal additive manufacturing is the challenging area and lot of research under progress. In the present review paper, an attempt is made to compare the properties of aluminium reinforced TiB₂ metal matrix composites fabricated through conventional manufacturing technique and also LASER assisted additive manufacturing technique. Bulk composite materials casting (stir casting) is followed for conventional method and layer wise material printing is done in additive manufacturing. Upon fabrication, the specimens are checked for particle distribution using SEM images. It is observed from micrographs that, distribution of reinforcing particles i.e. TiB₂ is fair enough in the composite fabricated through additive manufacturing technique. This influences possessing the uniform properties for further testing. The microhardness of the two different specimens are checked and it is found that, composites manufactured though conventional manufacturing showed higher hardness than composites manufactured though additive manufacturing. Higher composite ductility is observed in additive manufactured specimen.

Real-Time GPU-Accelerated Topology Optimization of a Compact Bracket: A Simulation-Only Workflow

Aswin Karkadakattil — 2026-03-04

Background: Topology optimization (TO) is widely adopted for lightweight structural design; however, its integration into early-stage engineering workflows is often constrained by computational expense and long solution times associated with conventional CPU-based solvers. The emergence of GPU-accelerated simulation environments offers the possibility of transforming topology optimization into a more interactive and accessible design tool. Objective: This Technical Note evaluates a practical GPU-accelerated workflow for topology optimization and examines its suitability for conceptual lightweight structural design using commercially available software. Methods: A compact triangular bracket was selected as a representative case study and analysed in ANSYS Discovery Live. A static concentrated load of 100 N was applied at one mounting interface, while the remaining interfaces were constrained using cylindrical supports to represent mechanically consistent boundary conditions. No dynamic or transient loading effects were considered. The optimization problem was formulated as compliance minimization subject to a 50 % global volume constraint. Material behaviour of AlSi10Mg was modelled as linear elastic and isotropic to ensure compatibility with the real-time GPU solver. Mesh sensitivity analysis and supplementary simulation-based validation checks were performed to assess structural consistency within a conceptual design framework. Results: The optimized configuration achieved approximately 50 % reduction in material volume while maintaining stresses and deformations within conservative limits under the prescribed static loading condition. Material redistribution followed principal load paths, and mesh refinement studies indicated stable topology convergence. The GPU-based solver enabled continuous visualization of stress evolution and structural response throughout the optimization process. Conclusions: The results demonstrate that GPU-accelerated topology optimization can provide mechanically interpretable and computationally efficient support for early-stage structural exploration. While limited to a simulation-only scope, the proposed workflow illustrates how interactive GPU-based tools can enhance structural insight and accelerate preliminary design decision-making without requiring high-performance computing infrastructure.

Normative and Methodological Support for the Life Cycle of Science-Intensive Products: Integrating Systems Engineering And Project Management

Svitlana Kryvova — 2026-02-02

Ukrainian high-tech manufacturing enterprises face a persistent gap between the national regulatory–methodological framework for life-cycle (LC) management and the requirements of global markets and certification. The problem stems from limited coverage of late LC phases, terminological inconsistency, and fragmented adoption of digital tools (PLM, MBSE, Digital Twins), which complicates alignment with international frameworks (ISO/IEC/IEEE 15288, ISO 2150x series, ECSS, AAP) and PMI/INCOSE practices. The aim is to develop an implementation-ready model, harmonized with international standards, that integrates Systems Engineering (SE) with Project/Programme/Portfolio Management (PPPM) while enabling a robust digital thread and transparent decision-making. The methodology comprises a comparative analysis of international and national normative documents, process modelling of the LC with Decision Gates, construction of mapping tables, and synthesis of a cross-project plan set. The results deliver a reference architecture for SE–PPPM integration; three application profiles (L/M/H); organizational roles and institutions; a digital infrastructure; a 12–24-month implementation roadmap; and a metrics system. The conclusions indicate that the proposed model bridges methodological gaps, increases LC governance and traceability, shortens time-to-market, and facilitates compliance with EASA/FAA/EN 9100/AQAP. The practical contribution is a set of typical recommendations and harmonization tools for enterprises in Ukraine’s aerospace and defense manufacturing sectors.