This paper deals with implementation of artificial neural network in the maximum power point tracking (MPPT) controller algorithm for modern household where electric vehicle (EV) was purchased. The proposed MPPT algorithm was designed to achie...

This study introduces an innovative, non-contact method for classifying the hardness of austenitic stainless steels (grade AISI 304) based on their intrinsic magnetic fields. Utilizing a 3 × 3 matrix sensor system, this research captures weak magnetic fields to produce precise 2D magnetic field maps of the samples. A key advancement is the application of a modified GoogleNet convolutional neural network, optimized with the stochastic gradient descent with momentum algorithm, which achieves exceptional classification accuracy, ranging from 95% to 100%, and median accuracies of 97.5% to 99%. This method stands out by revealing a novel correlation between annealing temperature and magnetic field strength, particularly a pronounced decline in magnetic properties at temperatures near 1000 °C. This observation underscores the sensitivity of magnetic profiles to heat treatments, offering a groundbreaking approach to material characterization. By enabling reliable, efficient, and fully automated hardness evaluation based on magnetic signatures, this work has the potential to transform materials engineering and manufacturing, setting a new benchmark for non-destructive material analysis techniques.

This article presents a neural network-based control method for maintaining the required output voltage of a synchronous buck converter. The goal was to replace a traditional PID controller with a neural network that calculates the duty cycle based on real-time data. Several versions of the neural network were tested. The final version, which included the input voltage, reference, and output current as inputs and compensated for dead time, was successfully validated on real hardware. It was able to respond to changes in load and input voltage within a limited operating range.

Quality inspection of electronic boards during the manufacturing process is a crucial step, especially in the case of specific and expensive power electronic modules. Soldering splash occurrence decreases the reliability and electric properties of final products. This paper aims to compare different YOLOv8 models (small, medium, and large) with the combination of basic image preprocessing techniques to achieve the best possible performance of the designed algorithm. As preprocessing methods, contrast-limited adaptive histogram equalization (CLAHE) and image color channel manipulation are used. The results show that a suitable combination of the YOLOv8 model and preprocessing methods leads to an increase in the recall parameter. In our inspection task, recall can be considered the most important metric. The results are supported by a standard two-way ANOVA test.

This study discusses the automated visual inspection of electronic boards used in the mass production of power electronics equipment. Soldering splashes, produced during the necessary manufacturing stages, can affect the hybrid power semicondu...