Search the digital library catalog

Query:	search in
	search in
	search in
	search in
Work type:	* old and bologna study programme
Language:
Search in:
Options:	Show only hits with full text
	Reset

1 - 2 / 2

Simulation-based study of structural changes in electrical time-series signals
Luka Živković, Željko Hederić, Tin Benšić, Goran Kurtović, Marinko Stojkov, 2025, original scientific article

Abstract: his paper uses statistical indicators to address the detection of changes in electrical signals typical of industrial and power systems. A dedicated MATLAB algorithm was developed to identify change points by tracking shifts in signal behaviour and statistical properties. To evaluate the method, synthetic signals were generated through simulation to reproduce the common patterns observed in these systems, allowing testing under different operating conditions and varying noise levels. The results demonstrate that the algorithm detects change points reliably across multiple scenarios, showing flexibility and robustness. This study highlights the value of simulation-based signal generation as a controlled environment for testing detection methods. It provides a foundation for future applications to more complex real-world electrical signal analysis tasks.
Keywords: break points, energy system, noise, segmentation, signals, simulation, time series
Published in DKUM: 01.10.2025; Views: 0; Downloads: 1
Link to file

Towards digital twinning of electrical motors – simulation models
Goran Kurtović, Luka Živković, Tin Benšić, Željko Hederić, Marinko Stojkov, 2025, original scientific article

Abstract: This paper presents a methodological framework for building a digital shadow of an induction motor based on standardised tests and a two‑axis (dq) simulation model. The tests were carried out according to IEEE Std 112 and IEC 60034‑2‑1. The parameters of the equivalent circuit were identified and entered into the model. Validation was performed by comparing the torque–speed and current–speed curves at 180 V and 220 V, while the nominal behaviour at 400 V was estimated using the model and voltage scaling. The model was then calibrated to reduce the discrepancy between the simulation and measurements, and the error was quantified using the root‑mean‑square error (RMSE) and mean absolute percentage error (MAPE). An automated load‑simulation setup that reproduces the torque test is also presented, enabling rapid evaluation of parameter influence. The results show a very good match in the current channel, with larger deviations in the prediction of characteristic torque points, indicating the limitations of linearised parameters and motivating nonlinear model extensions. The approach enables summarised reliable estimates at nominal voltage when direct measurements are not feasible.
Keywords: induction motor, digital shadow, standardised tests, dq model, torque test, automated load simulation, model calibration
Published in DKUM: 01.10.2025; Views: 0; Downloads: 2
Full text (624,48 KB)
This document has many files! More...

Search done in 0.02 sec.