DOI: 10.15199/48.2025.07.01
Authors: Pawel MILEWSKI, Pawel MORDAWSKI
Abstract: The article describes the implementation of a 499.5 kWp PV power plant for trolleybus traction. The focus is on the practical application of the power management system (SEMS) and the analysis of its architecture, including the converter station and the assemblies of PV panels. The modes of operation were discussed, taking into account, among other things, the possibility of realizing the power supply of own needs. The technical difficulties arising from direct cooperation with urban traction are shown. An analysis of traction voltage variability was made. By way of verification of operational data, the validity of the implementation was confirmed.
Keywords: trolleybuses, traction PV power plant, SEMS, energy storage tanks
Download PDFDOI: 10.15199/48.2025.07.02
Authors: Dorota BUGAŁA, Dominik ZIEMIÊCKI
Abstract: This paper presents the results of computer mapping of the generator of fast electrostatic EFT/Burst transients in two widely used and accepted simulation environments Matlab and LTSpice. The simulation models of a single pulse generator and a series of pulses were performed in accordance with the current PN-EN 61000-4-4 standard. The time courses of the obtained signals in the two programs are presented, which were then compared with each other using selected statistical parameters, relating the obtained responses of the programs to the normative theoretical description. It was shown that LTSpic software allows obtaining a more accurate model of the generator with the relevance of simulation parameters such as pulse duration, signal rise time, pulse duration at 50%, or shift of the maximum value of the generated voltage signal. When analyzing the voltage values over the entire pulse duration, lower error values were obtained using a generator modeled in Matlab.
Keywords: electromagnetic disturbance modeling, fast electrical transients, Matlab, LTSpice, electromagnetic compatibility
Download PDFDOI: 10.15199/48.2025.07.03
Authors: Adrian MŁOT, Mariusz KORKOSZ, Marian ŁUKANISZYN, Piotr ZASOWSKI
Abstract: This paper presents the influence of selected design parameters of a synchronous motor with magnets embedded in the rotor (IPM) on the density of torque produced. On this basis, the number of design parameters in the optimization process using a genetic algorithm was determined. The objective function responsible for reducing the cost of manufacturing engine components was investigated. Calculations were carried out using a two-dimensional finite element method implemented in the Matlab environment. Field calculations were extended to include analysis of efficiency and power losses
Keywords: AC IPM motor, optimization, genetic algorithms, electromagnetic torque
Download PDFDOI: 10.15199/48.2025.07.04
Authors: Sylvester FILIPIAK
Abstract: The article presents methods for multi-criteria optimization of medium-voltage (MV) network operation, taking into account current operational challenges faced by MV distribution networks. The methodology proposed in the article enables the use of evolutionary algorithms to coordinate the operation of local energy sources in the distribution network. The article presents analyses related to multi-criteria optimization using selected evolutionary algorithms and their proposed modifications. The analyses include the adopted optimization criteria for minimizing technical losses in the analyzed MV power distribution networks, minimizing voltage deviations, minimizing overloading of network equipment and optimizing power distributions ().
Keywords: Power grids, multi-criteria optimization, evolutionary methods, heuristic methods
Download PDFDOI: 10.15199/48.2025.07.05
Authors: Kamil SZOSTEK
Abstract: The article presents the results of a simulation to optimize the cleaning schedule of photovoltaic panels in the context of maximizing the net profit of a PV farm. The model takes into account panel degradation, fouling (in linear and logarithmic terms) and inflation affecting energy prices and cleaning costs. The effects of various parameters (β, Cc, r) on the cost-effectiveness of maintenance strategies were analyzed. The results indicate that properly selected cleaning intervals can significantly improve ROI.
Keywords: photovoltaics, simulation, optimization, energy production
Download PDFDOI: 10.15199/48.2025.07.06
Authors: Kacper BEDNARZ, Jaroslaw WOJTUŃ,Cezary ZIÓŁKOWSKI
Abstract: The most commonly studied scenario for localizing radio emitters using the SDF method is the case, in which a moving receiver allows determining the coordinates of a stationary emitter. Evaluation of the correctness of the implemented localization procedures requires practical setup of a test bed and empirical testing. The OOT module allows emulation of various scenarios under laboratory conditions using the GNU Radio environment. Such solution greatly simplifies the process of testing the developed solutions, both hardware and software. OOT module to emulate the Doppler effect in the GNU Radio environment.
Keywords: Emitter localization, SDF method, emulation, Doppler effect.
Download PDFDOI: 10.15199/48.2025.07.07
Authors: Tuan -Ho LE, Van -Ninh NGUYEN, Tham X. NGUYEN
Abstract: The main purpose of this article is to propose a hybrid approach based on feed -forward neural network model for wind power generation forecasting. First, the ARIMA model is used to predict the time series of wind power. Second, the nonlinear autoregressive neural network model is trained to produce wind energy forecasts. Third, the obtained forecast values of wind power generation using the above two models are combined using a feed -forward neural network model. Finally, three periods from Belgian wind farms are performed to validate the effectiveness of the proposed approach. The final results show that the proposed hybrid approach can provide more effective solutions than existing models based on average absolute error, average weighted percentage error and directional statistics.
Keywords: Wind power forecasting, ARIMA, Nonlinear autoregressive neural network, Coupled-forward neural network
Download PDFDOI: 10.15199/48.2025.07.08
Authors: Marcin BEDNAREK, Tadeusz DĄBROWSKI
Abstract: A distributed control system consisting of process and operator stations is considered. The execution of the tasks of the controller takes place cyclically. Diagnosing the execution time of operations allows determining the minimum cycle time of the task - that is, leaving a certain reserve of time for transmission protection operations. For this purpose, a non-standard use of function blocks based on the use of a time output, which continuously provides the value of the elapsed time from its driving.
Keywords: function block, FBD language, diagnosis, time reserve
Download PDFDOI: 10.15199/48.2025.07.09
Authors: Zbigniew WATRAL, Adam ROSIEK, Andrzej MICHALSKI
Abstract: The article describes laboratory tests of a DMFC-type fuel cell operating under variable load conditions. The results of tests with changing the nature of the load from resistive to resistive-inductive and stepping the value of the load current are presented. The tests were conducted on an arranged test stand using a measuring card connected to a computer and controlled by a program written in the LabView environment. Support preparing for the tests were simulations in Simulink.
Keywords: fuel cell, load shifting, power system
Download PDFDOI: 10.15199/48.2025.07.10
Authors: Ryszard SROKA, Janusz GAJDA, Piotr BURNOS, Mateusz DANIOŁ
Abstract: The paper presents the results of the study of metrological properties of dynamic vehicle weighing systems (Weigh-In-Motion - WIM). The main objective of the conducted research is to determine the technical conditions that enable implementation of WIM systems operating in the mode of direct elimination of overloaded vehicles from traffic. The parameters of the test stand are presented, the accuracy of the results obtained from such systems is determined, and the basic factors affecting the accuracy of dynamic vehicle weighing are indicated.
Keywords: Dynamic vehicle weighing, WIM systems, measurement accuracy, influencing factors
Download PDFDOI: 10.15199/48.2025.07.11
Authors: Fabian GIL, Robert BERCZYŃSKI
Abstract: The article presents a study of selected graphics card models on the training time of convolutional neural networks using a set of mammographic images. It is shown that appropriately selected parameters allow effective use of older GPUs in environments with limited computational resources. The use of transfer learning and multi-GPU configurations enables additional optimization of the learning process.
Keywords: GPU computational efficiency, deep learning, transfer learning, medical image classification
Download PDFDOI: 10.15199/48.2025.07.12
Authors: Maciej KLEBBA, Adrianna KLEBBA, Michal BRODZICKI, Arkadiusz FRĄCZ
Abstract: The article presents problems that occur during power measurement when distorted signals are present. Using the example of a system with a single-voltage rectifier, the results of measurements using RMS and TRMS myristors are presented. It also proposes a prototype solution that allows a different view of power measurement signals in which energy transfer does not occur throughout the entire period of the waveform.
Keywords: metrology, power, distorted signals, PF
Download PDFDOI: 10.15199/48.2025.07.13
Authors: Michal MAZUR, Michal WIŚNIOS
Abstract: The article presents a study of combustion products using spectrophotometric analysis to identify material burned. The spectrophotometric analysis of burned substances in this study involved measuring the attenuation properties of smoke from the combustion of different materials, such as different types of wood and plastics. The authors of the article presented preliminary results of tests performed on the designed test stand.
Keywords: electronics, fire alarm systems, spectrophotometry, fire
Download PDFDOI: 10.15199/48.2025.07.14
Authors: Włodzimierz PRZYBOROWSKI, Piotr PAZIEWSKI
Abstract: Determination of circuit parameters of synchronous machines - primarily reactance, is one of the more difficult problems of experimental study of these machines. This is due to the strong dependence of reactance on the state of magnetization and the dependence on the rotation angle. There are several methods for determining these parameters due to the physical complexity of these quantities and the need for multilateral verification. One of the methods of determining the main reactances of the machine is the method of small slip. The peculiarity of this method is to supply the armature circuit and to determine the correspondence of the direction of rotation of the magnetic field with the direction of rotation of the magnetizer - rotor. The paper presents the results of experimental determination of these parameters using computer aided measurements. It was possible to extend the measurements to the interval of velocity changes beyond the values of synchronous speed. The digital recording of the measurements greatly improved the research and facilitated the analysis of the results. It should be noted that the indicated reactances are the basis of the mathematical model of the synchronous machine for description in the states of current and motor operation, as well as in various kinetic states.
Keywords: synchronous machine, reactances, small slip method
Download PDFDOI: 10.15199/48.2025.07.15
Authors: Bogdan DZIADAK, Jacek STARZYŃSKI
Abstract: This paper presents the design, implementation and laboratory validation of a sensor for weak electric field detection, based on a field-effect transistor (FET). The proposed solution uses the high input impedance of MOSFET and JFET structures to convert electric field strength into a measurable voltage signal. Theoretical analysis describes the relationship between induced charge, electrode geometry and transistor response. A prototype sensor was built and tested using a TL071 amplifier with a JFET input. Experimental results confirmed a linear relationship between output voltage and field strength in the range from 1 kHz to 300 kHz, with a minimum sensitivity of 8 Vpp/m. The effect of electrode surface area on sensitivity was also investigated. The results indicate that the proposed architecture performs well for low-field measurements in conditions requiring high sensitivity and noise immunity, such as EMC diagnostics and biomedical applications
Keywords: Electric field sensor, JFET transistor, weak electric field measurements
Download PDFDOI: 10.15199/48.2025.07.16
Authors: Aleksandra SKUZA, Stanislaw ZIEMIANEK, Marek SUPRONIUK, Bogdan PERKA
Abstract: The paper presents the possibility of correct identification of a short circuit in a high-voltage network with distributed generation. The protections most prone to malfunction are those operating on the basis of the subimpedance criterion. Traditionally, the problem of determining the value of the impedance is solved by correction coefficients or by using overlapping distance protections located on opposite sides of the line. However, the problem is remote sources, including distributed renewable energy sources, especially those operating intermittently (wind farms or photovoltaic). These sources inject power into the line through power electronic devices, which are sensitive to voltage drops. For this reason, it is possible for them to be disconnected during short circuits, which can result in a decrease in current and a failure of correct operation of the protection. The paper will present methods used to improve the efficiency of protection in a system with distributed generation.
Keywords: Short-circuits, RES, Electricity protection
Download PDFDOI: 10.15199/48.2025.07.17
Authors: Adrian KNYZIAK, Weronika NIEDZIAŁEK
Abstract: A mobile primary standard of absorbed dose in water for radiotherapy has been developed at the General Office of Measures (GUM). The graphite calorimeter design, measurement system and signal analysis are presented. The system was tested under clinical conditions, eliminating correction factors: beam quality, recombination and polarization. This allowed a twofold reduction in uncertainty under clinical conditions. The calorimeter along with the measurement system has been verified in BIPM.IR-K6 key comparisons with positive results.
Keywords: calorimeter, absorbed dose, measurement system
Download PDFDOI: 10.15199/48.2025.07.18
Authors: Filip MUSIAŁEK, Jacek WOJTAS
Abstract: This paper presents the application of an artificial neural network for the detection of methane using laser Wavelength Modulation Absorption Spectroscopy (WMS). Tests in the experimental setup have shown that the proposed method can reduce the gas concentration determination time to 70 ms. The correctness of the measurement, expressed by the mean relative error (MAPE), was 1.7% while the precision of the measurement, expressed by the standard deviation, was 0.94 ppm. The lower detection limit of the meter was estimated at 70 ppb based on the Allan deviation plot for an averaging time of 150 s.
Keywords: Artificial neural networks, wavelength-modulated laser absorption spectroscopy
Download PDFDOI: 10.15199/48.2025.07.19
Authors: Maurycy MAZIUK, Laura JASIŃSKA, Jarosław DOMARADZKI
Abstract: The increasing contribution of photovoltaics to global power generation requires reliability and security of long-term investments, which is provided by properly developed technology for early and precise detection of defects in solar cells and panels. The paper presents a comparison of selected methods of defect analysis in photovoltaic cells based on proven imaging methods. A novelty presented in the paper is the analysis of electroluminescence images in selected spectral channels.
Keywords: defect imaging, photovoltaics, electroluminescence imaging
Download PDFDOI: 10.15199/48.2025.07.20
Authors: Łukasz MAKOWSKI, Daniel WĘGRZYNOWSKI, Lidia KOCON
Abstract: This paper evaluates the feasibility of implementing linear regression algorithms on microcontrollers in measurement systems. The study included OLS, RANSAC, Huber, and Theil-Sen methods, and then extended to the Hampel estimator. Analyzed were memory consumption, execution time and robustness to outliers. Experiments on the STM32L476RG microcontroller confirmed correct operation under practical conditions and highlighted the trade-offs between accuracy, resource usage and real-time requirements.
Keywords: Regression analysis, Computational efficiency, Embedded systems
Download PDFDOI: 10.15199/48.2025.07.21
Authors: Radia TABET-DERRAZ, Abdellatif BERKAT, Noureddine BOUKLI-HACENE
Abstract: In this article, we present an inexpensive miniaturized circular patch antenna that can be integrated into devices used with 5G technology. Miniaturization was achieved by introducing slots in the patch. The bandwidth of the miniaturized antenna was increased by using a partial ground plane with a circular slot. The prototype has an impedance bandwidth of -10 dB from 3.32 GHz to 3.62 GHz and a peak gain of 2.6 dBi. The dimensions of the antenna are approximately 31.6 × 35.9 mm², making it more compact and less expensive than other complex antennas with similar performance.
Keywords: Round miniaturized patch antenna, affordable antenna, bandwidth enhancement, 5G technology
Download PDFDOI: 10.15199/48.2025.07.22
Authors: Bouziane GHOUAL, Benyssaad YSSAAD, Youssouf MEDDAHI, Bouziane MELIANI
Abstract: This paper presents an innovative control strategy for drone trajectory tracking that combines the robustness of backstepping control with the adaptive optimization capabilities offered by the Garr-Ruf method. While backstepping controllers are known for their intrinsic stability, their effectiveness diminishes in the face of external disturbances such as wind changes. To remedy this, we propose a Garr-Ruf-optimized backstepping approach that dynamically adjusts control parameters to counter such disturbances, significantly improving tracking precision and system robustness. The proposed solution has been tested on a detailed drone model that takes into account both geometric and dynamic aspects, providing realistic evaluation conditions. A comparative analysis with a conventional backstepping controller under identical conditions demonstrates the superiority of the Garr-Ruf optimization in achieving accurate trajectory tracking even in harsh environments. In addition, comparisons with the GWO-optimized backstepping controller show that Garra-Rufa provides better adaptability and higher performance, confirming its potential as a more effective optimization strategy in drone control.
Keywords: Garra_Rufa, nonlinear drone control, drone trajectory tracking, backstepping control
Download PDFDOI: 10.15199/48.2025.07.23
Authors: Hadjer BENFATMA, Houari KHOUIDMI, Bessedik BOUBAKER
Abstract: This paper presents an improved maximum power point tracking (MPPT) strategy for autonomous photovoltaic (PV) systems through the integration of a feedforward neural network (FNN), a particle swarm optimization (PSO) algorithm and a fractional-order integral-proportional (FOPI) controller. The FNN is trained to predict optimal operating voltage based on real environmental data, such as insolation and temperature, while the PSO algorithm dynamically optimizes the parameters of the FOPI controller to ensure accurate and robust tracking of the maximum power point. The system was modeled and simulated in the MATLAB/Simulink environment, and its performance was evaluated under both steady-state and variable solar conditions. Compared to the conventional Perturb and Observe (P&O) algorithm, the proposed hybrid approach features higher tracking speed, lower power oscillations and better energy harvesting efficiency. The results highlight the potential of the proposed method for smart energy management in autonomous PV systems.
Keywords: Maximum Power Point Tracking (MPPT), Feedforward Neural Network (FNN), Particle Swarm Optimization (PSO), Fractional Order Integral-Proportional (FOPI) Regulator, Autonomous Photovoltaic (PV) System, Metaheuristic Optimization.
Download PDFDOI: 10.15199/48.2025.07.24
Authors: Mohammed Elamine LAHLACI, Mohamed MILOUDI2, Houcine MILOUDI, Abdelkader GOURBI, Mohammed Hamza BERMAKI; Mohamed MANKOUR, Abdelkader MOSTEFA
Abstract: Power electronics, commonly referred to as power electronics, is becoming a fundamental component of electrical engineering, with nearly 15% of all electricity generated, being converted. Recent advances in semiconductor power component technology have led to the development of smaller, lighter, cheaper and more efficient power converters. In modern industrial applications - especially in pumps, fans and compressors - they are widely used in drives of direct current (DC) motors because of their ability to precisely meet operational requirements. However, electromagnetic interference (EMI) in such systems can adversely affect their operation and even cause failures. This paper presents a comparative simulation study of two dc/ dc converter topologies used in dc motor drive control: a traditional series rotor and a synchronous-switching rotor. The EMI performed is analyzed for both common current (CM) and dc current (DC), using circuit modeling in LTspice. In the serial rotor, IRD530 MOSFETs and MUR 460 diodes were used, while in the conz synchronized rotor, two IRF530 MOSFETs were used for bidirectional switching. The results of the simulation show that the rotor with synchronization exhibits significantly lower EMI levels compared to the serial configuration. The improvement discussed is mainly related to reduced dv/dt and di/dt transients. The present findings underscore the effectiveness of synchronous switching in enhancing electromagnetic compatibility (EMC) in DC motor drive systems.
Keywords: Electromagnetic interference (EMC), CM common current, DM differential current, direct current (DC) motor
Download PDFDOI: 10.15199/48.2025.07.25
Authors: Razika SOUADEK, Nassir GUELLIL
Abstract: Classification of urban sounds has recently become an area of research due to its applications in environmental protection. This study presents a classification model using a convolutional neural network (CNN) combined with features based on the short-time Fourier transform (STFT), designed to recognize different urban sounds. The model was trained on three datasets: UrbanSound, ESC-50 and UrbanSound8K, divided into classes representing typical urban sounds such as traffic, construction sounds and human activity. Key stages of the model include feature extraction using STFT, data augmentation to increase robustness, and optimization of the CNN architecture. Evaluation results indicate high performance, achieving about 95% accuracy on validation data. This study contributes to the development of effective urban sound classification systems that can support smart city initiatives and urban planning.
Keywords: Classification of urban sounds, CNN, STFT
Download PDFDOI: 10.15199/48.2025.07.26
Authors: Aissa BELHADJ, Ali KIES, Fatima Zohra MOSTEFA, Zoulikha MEKKAKIA MAAZA
Abstract: Roadside units (RSUs) are an essential component of vehicle internet infrastructure, given their effective role in reducing latency, maintaining reliable communications and enhancing data security. RSUs are expensive, requiring an effective deployment model that meets requirements while reducing costs. The location of RSUs can affect network performance in terms of time delay and transmission efficiency. This paper compares different approaches used in RSU deployment, each with different deployment rates. Simulation results showed that the connection probability ratio used in the ORSD approach was better. Based on these results, we proposed an RSU deployment scheme based on the connectivity probability at intersections in order to increase the network's connectivity speed and ensure that as many vehicles as possible are served at the same time. Since the purpose of the deployment was multi-target, we applied the algorithm, called BAOA, to an area in the city of Oran, Algeria. Simulation results showed that this approach achieves good vehicle coverage and effectively reduces delays compared to other approaches that use a road section in a candidate position and include traffic demand as a determining factor for RSU deployment.
Keywords: Roadside units (RSUs), deployment, connectivity probability, network performance
Download PDFDOI: 10.15199/48.2025.07.27
Authors: Ilyas MIMOUN, Youcef MIHOUB, Aicha ASRI, Said HASSAINE
Abstract: Partial shading conditions (PSC) complicate the power-voltage (PV) curve of photovoltaic (PV) systems, characterized by a single global maximum power point (GMPP) and multiple local maxima. Traditional MPPT algorithms are often unable to locate the GMPP under such conditions. This study proposes the Coot Optimization Algorithm (COA) as an innovative approach to tracking GMPP in photovoltaic systems. The suggested COA algorithm solves drawbacks of conventional MPPT algorithms under PSC, as well as drawbacks of particle swarm optimization (PSO), including reduced tracking performance and increased tracking time. A comparative analysis using MATLAB R2023b/SIMULINK evaluated the performance of COA-MPPT in a PV system during complex PSCs, and the results were compared with those obtained using PSOMPPT. The simulation results show that the COA technique consistently achieves efficiencies in excess of 95% under a variety of conditions, which underscores its effectiveness in enhancing the performance of PV systems. Moreover, the results indicate that COA outperforms the PSO technique in both performance and tracking speed.
Keywords: Photovoltaic (PV) systems, partial shading conditions (PSC), global maximum power point (GMPP), Coot optimization algorithm (COA)
Download PDFDOI: 10.15199/48.2025.07.28
Authors: Abdelkader BELHADRI, Ibtissem BENCHENNANE
Abstract: This paper investigates software level optimization to accelerate the convergence of deep learning training by dynamically adapting the hyper-parameter of deep learning speed. First, the basic motivations for reducing training time were presented. Then, an extensive review of learning rate scheduling methods was made. Empirical evaluation of the CIFAR-10 dataset, using the ResNet-18 architecture, shows that the strategies discussed significantly improve training efficiency without the need to compromise accuracy or increase the amount of energy consumed.
Keywords: Deep learning, learning speed, power consumption, accuracy
Download PDFDOI: 10.15199/48.2025.07.29
Authors: Fayçal MEHEDI, Ismail BOUYAKOUB, Abdelkader YOUSFI, Zakaria REGUIEG, Youcef BOT
Abstract: This paper presents an adaptive fuzzy logic proportional integral controller (AFLPI) for robust speed control of a five-phase permanent magnet synchronous motor (FP-PMSM) fed by a multilevel inverter (MLI). The main goal is to improve the dynamic response, reduce steady-state error and increase the robustness of to parameter changes compared to traditional PI-based field-oriented control (FOC). Simulations confirm the superiority of the proposed FOC-AFLPI-MLI system in terms of tracking and disturbance rejection performance, demonstrating the effectiveness of the AFLPI algorithm in multiphase PMSM drives.
Keywords: Multiphase permanent magnet synchronous motor, adaptive fuzzy PI, field-oriented control, robustness
Download PDF