The paper considers two variants of an ion source with a resistive heating, designed to generate a test ion beam and adjust the transportation channel of the IRINA mass separator complex when the PIK nuclear reactor is switched off. The operating conditions of the ion source inside the vacuum channel of the reactor impose strict, unique requirements on its materials and design, which makes direct borrowing of known solutions impossible. The article presents the results of a numerical study of the ion source, as well as experimental data obtained during its testing.
Understanding the interplay between microstructure and transport properties is essential for the development of high-performance thermoelectric materials. Here, we investigate the structure–transport–performance relationships in directionally solidified GaSb–CrSb and InSb–CrSb eutectic composites. X-ray diffraction confirms the formation of phase-pure two-phase eutectic structures comprising a cubic semiconductor matrix and hexagonal NiAs-type CrSb inclusions. The aligned CrSb rods generate a highly anisotropic architecture that strongly influences charge and heat transport. Electrical conductivity is enhanced along the rod direction due to the formation of preferential conduction pathways, whereas thermal conductivity is suppressed perpendicular to the rods as a result of intensified phonon scattering at semiconductor–metal interfaces. InSb–CrSb exhibits higher electrical conductivity than GaSb–CrSb, while both systems demonstrate pronounced transport anisotropy governed by the eutectic microstructure. The results reveal the critical role of interface-controlled transport in decoupling electrical and thermal conduction and highlight directionally solidified antimony
The results of a study on the effect of the particle size distribution of a tungsten filler on the properties of poly(methyl methacrylate)-based composites are presented. A bimodal mixture of particles with average sizes of 235 nm and 5.9 μm (in a 1:2 ratio) and a monomodal powder of 5.9 μm were used as fillers. Samples with a tungsten volume content from 0 to 50 % were fabricated by hot pressing in a vacuum chamber. The density and the longitudinal acoustic wave velocity were measured, and the acoustic impedance was calculated. It was established that the use of the bimodal filler provides a higher composite density (8.60 g/cm³) than that obtained with the monomodal filler (6.10 g/cm³) at tungsten filling fractions of 50 vol%. The acoustic impedance for both series was found to be maximal at 30 vol% tungsten and amounted to 11.49 MRayl and 8.28 MRayl for the bimodal and monomodal fillers, respectively. The optimal range of tungsten content, which ensures the maximum acoustic impedance while maintaining processability, is 25–35 vol%. The results obtained can be used in the development of composites with improved acoustic and radiation-shielding characteristics.
The research evaluates the capacity of machine learning and deep learning algorithms to classify cancer progression using dataset information from support and rehabilitation centers for cancer patients. The dataset contains five possible classifications including benign and improvement and malignant as well as progress and stable conditions. The research evaluated XGBoost and Random Forest with Support Vector Machine (SVM) and LightGBM (LGBM) through an analysis of accuracy and precision and recall and F1-score. LGBM demonstrated the best performance in classification by achieving above 90% accuracy along with F1-score 0.97 for improvement class. The Bidirectional Encoder Representations from Transformers (BERT) model achieved high predictive accuracy in classifying benign (0.93 accuracy) and malignant (0.94 accuracy) cases while processing text. Both machine learning and BERT models showed difficulties in separating the progress and stable categories indicating that model optimization remains necessary. The main achievement of this research lies in developing a hand-picked dataset which serves to boost cancer progression identification abilities. The produced findings demonstrate how advanced classification models can assist clinical diagnoses through promising developments for automatic patient diagnosis and monitoring.
Three-dimensional holographic processing of sonar signals in shallow waters is described. In comparison with two–dimensional holographic processing, which is currently being developed, a third independent variable is introduced - the spatial coordinate, which characterizes the distributed reception of sound source noise emission. The processing significantly expands the possibilities of two-dimensional signal processing, as it allows you to determine the parameters of the noise source and evaluate the dispersion characteristics of the water area in the absence of knowledge about the transfer function of the medium. A comparative analysis of the effectiveness of two-dimensional and three-dimensional holographic signal processing is performed. The results of numerical simulation are presented.
