The formulation of the Hall–Petch relationship in the early 1950s has raised immense interest in studying the influence of the grain size of solid materials on their properties. Grain refinement can be achieved through extreme deformation. In the presented study, Equal-Channel Angular Pressing (ECAP) was successfully applied to produce an ultrafine-grained microstructure in a pure commercial Cu of 99.9 wt%. Samples were processed by ECAP at 21 °C for six passes via route A. A new equation of equilibrium that allows the exact determination of the number of extrusions and other technological parameters required to achieve the desired final grain size has been developed. The presented research also deals, in a relatively detailed and comparative way, with the use of ultrasound. In this context, a very close correlation between the process functions of extrusion and the speed of longitudinal ultrasonic waves was confirmed.

This study presents a novel multidimensional computational framework for optimising the tilt angle of photovoltaic systems by utilising complex variable modelling, transfer function analysis, and artificial neural network modelling. A 35 kWp photovoltaic system was monitored continuously over one year at a fixed 25° tilt, and comprehensive statistical analyses were conducted to evaluate monthly and seasonal variations in energy output. The proposed methodology makes differential equations with constant coefficients and transfer function models to describe time-dependent energy balance, enabling robust dynamic system identification. Empirical relationships between tilt angle and produced energy yield were effectively quantified, with a negative correlation coefficient value of 0.95 observed between steady-state energy production and tilt angle. Comparative analysis demonstrated negligible differences in annual yield between the 25° and 35° tilt configurations. Validation by neural network, utilising various architectures, revealed that a single hidden layer with 12 neurons achieved optimal model prediction performance value of 99.15%. The results confirm the efficiency of combining complex variable models with AI-based validation for precise and practical PV system modelling. This integrated approach facilitates improved decision-making in PV design by providing accurate predictions of energy yield under varying operational conditions.

This article deals with the regression analysis of the ultrasonic signal amplitude when the character of the reflection surface has been changed from a planar case to a sharp corner case. The experiment was performed at a measurement distance within the interval from 100 mm to 215 mm. A nonlinear correlation between the amplitude of the ultrasound signal and the measured distance was demonstrated. By analyzing the frequency spectra, a poor nonlinear correlation between the maximum frequency component and the distance vector was found for the sharp corner case versus the planar case, which proved similar nonlinear characteristics as the signal amplitude marker. The strong linear correlation in the distance difference vectors in the amplitude analysis of the ultrasound signal confirmed the hypothesis of a direct relationship between the reflection surface geometric characteristic and the polarity of the difference. The ultrasound signal was identified as a 3rd-order dynamic system. The nonlinear correlation of the steady-state values of the modelled transfer functions versus distance likewise shows the characteristic of the polarity difference or character derivative as a quantification marker of the characteristics of the reflection surface from the geometric point of view.

In automated tomato harvesting, it is essential to distinguish between the main stem of the plant and the stem with the fruit, which was the aim of this research. We hypothesised that instance segmentation can recognise the main stem and the stem with the fruit in tomato images, even with unbalanced data, with a precision better than 0.7 and an inference speed better than 3 fps. The tomato variety Merlice was used. From the measurement, 399 images were extracted concerning three conditions. The training, testing, and validating sets were randomly divided in an 80 : 10 : 10 ratio, with 10% of the training set allocated as the validation set for cross-validation during model training. We propose to use the YOLO-v8 model for recognition. With image size 800 × 800 pixels and model YOLO-v8m-seg, precision was 0.736, recall 0.738, mAP50 0.712 and mAP50-95 0.301. Inference speed was 6.34 fps. When using the YOLO-v8l-seg model, precision achieved was 0.761, recall 0.673, mAP50 0.689 and mAP50-95 0.308. Interference speed was 3.92 fps. Using integrated GPU, the inference can reach the speeds of up to 10 fps. Experimental results demonstrate usable overall recognition performance for unbalanced samples. The findings of this work provide a technical foundation for developing tomato harvesting robots.