Yassen Gorbounov
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Yassen Gorbounov has a total of 44 co-authors. Here are the top co-authors by number of shared publications:
- Hao Chen
China University of Mining and Technology
15 shared publications - Patrick Wheeler
University of Nottingham
5 shared publications - Xing Wang
Shenzhen Research Institute, China University of Mining and Technology, Shenzhen, China
5 shared publications - Atanas Kolev
New Bulgarian University
4 shared publications - Hemiao Liu
School of Electrical Engineering, China University of Mining and Technology, Xuzhou, China
3 shared publications - João Francisco Martins
NOVA School of Science and Technology, UNINOVA-CTS and LASI, NOVA University Lisbon, 2829-516 Caparica, Portugal
3 shared publications
Latest Publications
- 19 October 2026
- Published by Springer Nature
- p. 421-433
Unsupervised Vibration Fault Detection Using Cepstrum-Based Feature Extraction and Mahalanobis Analysis
- 19 October 2026
- Published by Springer Nature
- p. 421-433
Unsupervised Vibration Fault Detection Using Cepstrum-Based Feature Extraction and Mahalanobis Analysis
- 19 October 2026
- Published by Springer Nature
- p. 434-449
Data-Driven Predictive Maintenance via Adaptive Hilbert-Huang Feature Extraction and XGBoost Classification
- 19 October 2026
- Published by Springer Nature
- p. 434-449
Data-Driven Predictive Maintenance via Adaptive Hilbert-Huang Feature Extraction and XGBoost Classification
- 2 September 2026
- Published by Springer Nature
- p. 368-382
Enhancing Robotic Simulations in Webots Using AI-Driven Decision Making
- 2 September 2026
- Published by Springer Nature
- p. 368-382
Enhancing Robotic Simulations in Webots Using AI-Driven Decision Making
- 2 September 2026
- Published by Springer Nature
- p. 283-307
Applying Feature Extraction from Deep Learning Models for Vibration Analysis
- 2 September 2026
- Published by Springer Nature
- p. 283-307
Applying Feature Extraction from Deep Learning Models for Vibration Analysis
- Open Access
- 1 May 2026
- Published by IOP Publishing in Journal of Physics: Conference Series
- Vol. 3239 (1)
Enhanced Fault Diagnosis of Rolling Bearings Using LSTM Autoencoders and Feature-Based XGBoost Classification
Accurate diagnosis of rolling bearing faults is critical for ensuring the robustness and reliability of rotating machinery. This study presents a hybrid diagnostic framework that combines an LSTM autoencoder, trained on healthy vibration signals, with an extreme gradient boosted classifier. The features are extracted from the autoencoder’s compressed signal representation and are used as inputs to an XGBoost model trained on labeled data. These features reveal underlying signal patterns that traditional methods may overlook, leading to improved fault classification. The approach achieves high detection accuracy ( F 1 -Score and Accuracy > 0.98) while reducing computational overhead by omitting the decoder during inference. Comparative results show that it significantly outperforms ARIMA-based anomaly detection. These findings support leveraging deep signal encoding to improve predictive maintenance strategies in industrial environments.
- 1 May 2026
- Published by IOP Publishing in Journal of Physics: Conference Series
- Vol. 3239 (1)
Open AccessEnhanced Fault Diagnosis of Rolling Bearings Using LSTM Autoencoders and Feature-Based XGBoost Classification
Accurate diagnosis of rolling bearing faults is critical for ensuring the robustness and reliability of rotating machinery. This study presents a hybrid diagnostic framework that combines an LSTM autoencoder, trained on healthy vibration signals, with an extreme gradient boosted classifier. The features are extracted from the autoencoder’s compressed signal representation and are used as inputs to an XGBoost model trained on labeled data. These features reveal underlying signal patterns that traditional methods may overlook, leading to improved fault classification. The approach achieves high detection accuracy ( F 1 -Score and Accuracy > 0.98) while reducing computational overhead by omitting the decoder during inference. Comparative results show that it significantly outperforms ARIMA-based anomaly detection. These findings support leveraging deep signal encoding to improve predictive maintenance strategies in industrial environments.
- Open Access
- 24 April 2026
- Published by University of Zilina in Communications - Scientific letters of the University of Zilina
- Vol. 28 (2) , C35-C46
Design and Permanent Magnets Reduction of Linear Oscillatory Machine for Reciprocating Drives in Transport
Electric vehicles (EVs) require efficient compressors for air-conditioning and battery thermal management, typically driven by a dedicated electric motor. In this paper is dealt with the design and optimization of a stator-magnet transverse-fl...
- 24 April 2026
- Published by University of Zilina in Communications - Scientific letters of the University of Zilina
- Vol. 28 (2) , C35-C46
Open AccessDesign and Permanent Magnets Reduction of Linear Oscillatory Machine for Reciprocating Drives in Transport
Electric vehicles (EVs) require efficient compressors for air-conditioning and battery thermal management, typically driven by a dedicated electric motor. In this paper is dealt with the design and optimization of a stator-magnet transverse-fl...
- Open Access
- 1 December 2025
- Published by Wydawnictwo Naukowe Gabriel Borowski (WNGB) in Ecological Engineering & Environmental Technology
- Vol. 26 (12) , 226-237
Enhancing biomethane production with a modulated electric field in anaerobic digestion-microbial electrolysis cell systems
The advantages of integrating bioelectrochemical systems into anaerobic wastewater treatment reactors have been established over the last decade. In addition to enhancing biogas yield, its integration also improves biogas quality, eliminating the need for additional purification before its...
- 1 December 2025
- Published by Wydawnictwo Naukowe Gabriel Borowski (WNGB) in Ecological Engineering & Environmental Technology
- Vol. 26 (12) , 226-237
Open AccessEnhancing biomethane production with a modulated electric field in anaerobic digestion-microbial electrolysis cell systems
The advantages of integrating bioelectrochemical systems into anaerobic wastewater treatment reactors have been established over the last decade. In addition to enhancing biogas yield, its integration also improves biogas quality, eliminating the need for additional purification before its...
- 12 September 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Transportation Electrification
- Vol. 11 (6) , 14113-14130
A Novel Adaptive Force Ripple Suppressing Method for Double-Sided Switched Reluctance Linear Motor
This article proposes a method to suppress electromagnetic force ripple in switched reluctance linear motors (SRLMs). While the force distribution function (FDF) method is effective, conventional approaches employing fixed functions limit adaptability to varying conditions, and existing adaptive algorithms often exhibit suboptimal online adjustment. To overcome these limitations, this study introduces an adaptive reference trajectory that dynamically adjusts based on mover speed and load. In addition, turn-on/off positions are adaptively modified according to a commutation point selected by the force-per-ampere rate; the selection of the commutation point reduces the current root-mean-square (RMS) and significantly enhances the tracking performance of the system. By transforming the Sigmoid function, the reference trajectory is defined by only two parameters, enabling easy online integration of the adaptive function. Furthermore, a multistep continuous control set model predictive controller (CCS-MPC) with self-correction is adopted as the current controller to improve tracking performance. Simulation and experimental results demonstrate superior regulation performance compared to genetic algorithm (GA) iterative methods and the conventional FDF approach. Compared with the GA algorithm, this method reduces the electromagnetic force ripple by an average of 11.3% and the average current RMS by 2.7%.
- 12 September 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Transportation Electrification
- Vol. 11 (6) , 14113-14130
A Novel Adaptive Force Ripple Suppressing Method for Double-Sided Switched Reluctance Linear Motor
This article proposes a method to suppress electromagnetic force ripple in switched reluctance linear motors (SRLMs). While the force distribution function (FDF) method is effective, conventional approaches employing fixed functions limit adaptability to varying conditions, and existing adaptive algorithms often exhibit suboptimal online adjustment. To overcome these limitations, this study introduces an adaptive reference trajectory that dynamically adjusts based on mover speed and load. In addition, turn-on/off positions are adaptively modified according to a commutation point selected by the force-per-ampere rate; the selection of the commutation point reduces the current root-mean-square (RMS) and significantly enhances the tracking performance of the system. By transforming the Sigmoid function, the reference trajectory is defined by only two parameters, enabling easy online integration of the adaptive function. Furthermore, a multistep continuous control set model predictive controller (CCS-MPC) with self-correction is adopted as the current controller to improve tracking performance. Simulation and experimental results demonstrate superior regulation performance compared to genetic algorithm (GA) iterative methods and the conventional FDF approach. Compared with the GA algorithm, this method reduces the electromagnetic force ripple by an average of 11.3% and the average current RMS by 2.7%.
- 14 August 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Power Electronics
- Vol. 41 (2) , 2913-2926
A Wireless Power Transfer System Based on Hybrid Self-Switching of Coupled Inductors
This article proposes a wireless power transfer system based on hybrid self-switching of coupled inductor that not only realizes constant current (CC)/constant voltage (CV) output but also effectively copes with a variety of abnormal working conditions in the charging process by switching on the secondary side. First, the LCC-LCC/S self-switching composite topology is presented, and the CC/CV switching is achieved by utilizing a coupled inductor with a center tap based on the LCC compensation topology. The working principle of the CC/CV topology and the parameter calculation flow are analyzed in detail. Subsequently, the impedance, output characteristics, and switching characteristics of the CC/CV topology are simulated and analyzed. In addition, the topology’s stability under abnormal working conditions, such as the missing adjacency, load short-circuit, and load open-circuit, is analyzed. Finally, an experimental platform is constructed with the objective of verifying the feasibility and effectiveness of the proposed scheme. The experimental results demonstrate that the proposed charging system exhibits excellent CC and CV output characteristics without the need for additional inductor capacitance. The proposed system has a maximum output efficiency of 89.6%, a maximum current of 5.3 A, and a maximum voltage of 23 V, which fully meets the requirements of CC and CV wireless charging.
- 14 August 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Power Electronics
- Vol. 41 (2) , 2913-2926
A Wireless Power Transfer System Based on Hybrid Self-Switching of Coupled Inductors
This article proposes a wireless power transfer system based on hybrid self-switching of coupled inductor that not only realizes constant current (CC)/constant voltage (CV) output but also effectively copes with a variety of abnormal working conditions in the charging process by switching on the secondary side. First, the LCC-LCC/S self-switching composite topology is presented, and the CC/CV switching is achieved by utilizing a coupled inductor with a center tap based on the LCC compensation topology. The working principle of the CC/CV topology and the parameter calculation flow are analyzed in detail. Subsequently, the impedance, output characteristics, and switching characteristics of the CC/CV topology are simulated and analyzed. In addition, the topology’s stability under abnormal working conditions, such as the missing adjacency, load short-circuit, and load open-circuit, is analyzed. Finally, an experimental platform is constructed with the objective of verifying the feasibility and effectiveness of the proposed scheme. The experimental results demonstrate that the proposed charging system exhibits excellent CC and CV output characteristics without the need for additional inductor capacitance. The proposed system has a maximum output efficiency of 89.6%, a maximum current of 5.3 A, and a maximum voltage of 23 V, which fully meets the requirements of CC and CV wireless charging.
- Open Access
- 1 April 2025
- Published by IOP Publishing in Journal of Physics: Conference Series
- Vol. 3002 (1) , 012006
Application of Machine Learning Methods to Predict and Analyze the Condition of Rolling Bearings
Rising demands for industrial machinery quality and reliability lead to automating the analysis and prediction of failures, shifting from emergency repairs to planned maintenance. The article proposes methods for the analysis of possible rolling bearing failures. Freely accessible damage-type data sources adapted for autoregressive modeling were used. Classification techniques such as logistic regression, XGBoost, and artificial neural networks are compared. The analysis is performed in the time and frequency domains, and an attempt is made to apply the Cepstrum analysis in the field of mechanical vibrations. Some anomaly detection methods are applied, including ARIMA and autoencoder. The results are consolidated in a purpose-built web-based application. In this way, a comprehensive diagnostic system is proposed that uses modern machine learning techniques and methods for detecting and classifying anomalies in rolling bearing failures. The final goal of the proposed research is to create an integrated system with high speed and increased reliability for decision-making, thereby prolonging the life of mechatronic systems and improving the production indicators, which would have a strong economic effect.
- 1 April 2025
- Published by IOP Publishing in Journal of Physics: Conference Series
- Vol. 3002 (1) , 012006
Open AccessApplication of Machine Learning Methods to Predict and Analyze the Condition of Rolling Bearings
Rising demands for industrial machinery quality and reliability lead to automating the analysis and prediction of failures, shifting from emergency repairs to planned maintenance. The article proposes methods for the analysis of possible rolling bearing failures. Freely accessible damage-type data sources adapted for autoregressive modeling were used. Classification techniques such as logistic regression, XGBoost, and artificial neural networks are compared. The analysis is performed in the time and frequency domains, and an attempt is made to apply the Cepstrum analysis in the field of mechanical vibrations. Some anomaly detection methods are applied, including ARIMA and autoencoder. The results are consolidated in a purpose-built web-based application. In this way, a comprehensive diagnostic system is proposed that uses modern machine learning techniques and methods for detecting and classifying anomalies in rolling bearing failures. The final goal of the proposed research is to create an integrated system with high speed and increased reliability for decision-making, thereby prolonging the life of mechatronic systems and improving the production indicators, which would have a strong economic effect.