Kun Feng
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Kun Feng has a total of 37 co-authors. Here are the top co-authors by number of shared publications:
- Yuan Xiao
State Key Laboratory of High-End Compressor and System Technology, Beijing University of Chemical Technology, Beijing, China
4 shared publications Dongyan Miao
State Key Laboratory of High-End Compressor and System Technology, Beijing University of Chemical Technology, Beijing, China
3 shared publicationsZhouzheng Li
State Key Laboratory of High-End Compressor and System Technology, Beijing University of Chemical Technology, Beijing, China
3 shared publicationsJinji Gao
2 shared publicationsMinghui Hu
Key Lab of Engine Health Monitoring-Control and Networking of Ministry of Education, Beijing University of Chemical Technology, Beijing, 100029, P.R. China and Beijing Key Laboratory of High-end Mechanical Equipment Health Monitoring and Self-Recovery, Beijing University of Chemical Technology, Beijing 100029, P.R. China
2 shared publications- Minghui Hu
State Key Laboratory of High-End Compressor and System Technology, Beijing University of Chemical Technology, Beijing, China
2 shared publications
Latest Publications
- 9 April 2026
- Published by Springer Nature in Journal of Mechanical Science and Technology
- Vol. 40 (4) , 3139-3154
High-precision simulation method for gas turbine gas-path performance based on LSTM–KAN network error correction
- 9 April 2026
- Published by Springer Nature in Journal of Mechanical Science and Technology
- Vol. 40 (4) , 3139-3154
High-precision simulation method for gas turbine gas-path performance based on LSTM–KAN network error correction
- 29 December 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Instrumentation and Measurement
- Vol. 75 (00189456) , 1-15
An Improved Generalized Correntropy-Driven Adaptive Cyclostationary Blind Deconvolution Method for Bearing Fault Diagnosis Under Strong Interference
Accurate extraction of second-order cyclostationary (CS2) components in bearing fault diagnosis requires effective suppression of first-order cyclostationary (CS1) interference. Under strong cyclostationary disturbances and non-Gaussian noise, conventional filtering methods often lack robustness to impulsive noise and fail to sufficiently suppress CS1 components. To address this, a generalized correntropy-driven adaptive joint filtering framework is proposed, integrating adaptive filtering and blind deconvolution. In the first stage, an adaptive filter based on generalized correntropy and an energy-guided dynamic step-size mechanism is used to remove CS1 and enhance noise robustness. A multi-group candidate fault frequency (MGCFFs) extraction strategy then identifies dominant fault frequencies, which are passed to the second-stage generalized Gaussian cyclostationary blind deconvolution (CYCBDβ) filter. The filter length is adaptively optimized using a sparsity-based performance efficiency ratio (PER) to amplify CS2 features. The proposed method enables adaptive parameter tuning and coordinated filtering, ensuring accurate extraction of weak fault features under complex, noisy conditions. Simulation, experimental, and engineering validations confirm its superiority over conventional approaches in feature enhancement and robustness.
- 29 December 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Instrumentation and Measurement
- Vol. 75 (00189456) , 1-15
An Improved Generalized Correntropy-Driven Adaptive Cyclostationary Blind Deconvolution Method for Bearing Fault Diagnosis Under Strong Interference
Accurate extraction of second-order cyclostationary (CS2) components in bearing fault diagnosis requires effective suppression of first-order cyclostationary (CS1) interference. Under strong cyclostationary disturbances and non-Gaussian noise, conventional filtering methods often lack robustness to impulsive noise and fail to sufficiently suppress CS1 components. To address this, a generalized correntropy-driven adaptive joint filtering framework is proposed, integrating adaptive filtering and blind deconvolution. In the first stage, an adaptive filter based on generalized correntropy and an energy-guided dynamic step-size mechanism is used to remove CS1 and enhance noise robustness. A multi-group candidate fault frequency (MGCFFs) extraction strategy then identifies dominant fault frequencies, which are passed to the second-stage generalized Gaussian cyclostationary blind deconvolution (CYCBDβ) filter. The filter length is adaptively optimized using a sparsity-based performance efficiency ratio (PER) to amplify CS2 features. The proposed method enables adaptive parameter tuning and coordinated filtering, ensuring accurate extraction of weak fault features under complex, noisy conditions. Simulation, experimental, and engineering validations confirm its superiority over conventional approaches in feature enhancement and robustness.
- 5 March 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Instrumentation and Measurement
- Vol. 74 (00189456) , 1-11
Integrating Casing Vibration and Reconstructed Pressure Signal for Gas Turbine Blade Fault Diagnosis
Blade faults pose significant risks to the reliability and safety of gas turbines. Blade monitoring conditions through casing vibration is feasible, yet the transmission path attenuates blade features in the signal, leading to inaccuracies in fault diagnosis. Considering the effectiveness and directness of internal pressure signals for diagnosing blade fracture faults, we propose a method that utilizes transfer functions to reconstruct pressure signals from casing vibration, despite the fact that the pressure signal cannot be measured directly during gas turbine operation. This paper first investigates the changes and correlations between pressure signals and casing vibrations under blade fracture faults through theoretical modelling. Secondly, we propose a new framework that uses transfer functions to reconstruct pressure signals from vibration signals and combines it with an autoencoder model for blade fault diagnosis. Moreover, a rotor-blade-casing test rig was constructed to validate the proposed method. Experimental results and comparative analysis show that the method of integrating vibration signals and reconstructing pressure signals can improve the accuracy and robustness of blade fault diagnosis. This research lays a theoretical foundation for more reliable real-time monitoring of gas turbine blades.
- 5 March 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Instrumentation and Measurement
- Vol. 74 (00189456) , 1-11
Integrating Casing Vibration and Reconstructed Pressure Signal for Gas Turbine Blade Fault Diagnosis
Blade faults pose significant risks to the reliability and safety of gas turbines. Blade monitoring conditions through casing vibration is feasible, yet the transmission path attenuates blade features in the signal, leading to inaccuracies in fault diagnosis. Considering the effectiveness and directness of internal pressure signals for diagnosing blade fracture faults, we propose a method that utilizes transfer functions to reconstruct pressure signals from casing vibration, despite the fact that the pressure signal cannot be measured directly during gas turbine operation. This paper first investigates the changes and correlations between pressure signals and casing vibrations under blade fracture faults through theoretical modelling. Secondly, we propose a new framework that uses transfer functions to reconstruct pressure signals from vibration signals and combines it with an autoencoder model for blade fault diagnosis. Moreover, a rotor-blade-casing test rig was constructed to validate the proposed method. Experimental results and comparative analysis show that the method of integrating vibration signals and reconstructing pressure signals can improve the accuracy and robustness of blade fault diagnosis. This research lays a theoretical foundation for more reliable real-time monitoring of gas turbine blades.
- Open Access
- 16 January 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 13, 14225-14238
A State-Supervised Model and Novel Anomaly Index for Gas Turbines Blade Fault Detection Under Multi-Operating Conditions
To meet industrial demands, gas turbines typically operate under multiple conditions, presenting unique challenges for fault diagnosis. This paper proposes a novel blade fault detection framework designed for such environments. First, a State-Supervised Variational Autoencoder (SS-VAE) model is introduced, which integrates the learning process of turbine operational states into the VAE bypass, enabling it to capture variations in vibration signal data across different operating conditions. Through the fusion decoding of operating states and vibration signals, the foundation of the SS-VAE model and its loss function is established. Secondly, a new State Mapping Error (SME) index is introduced to further address the issue of reconstruction error variability across multiple conditions. The proposed method is validated through a blade fault test rig and applied in an industrial gas turbine blade fault case. Results demonstrate that the SS-VAE model and SME index effectively capture vibration signal changes due to blade faults across multi-operating conditions, achieving accurate monitoring and diagnosis. Compared to traditional methods, the proposed approach shows higher effectiveness and robustness.
- 16 January 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 13, 14225-14238
Open AccessA State-Supervised Model and Novel Anomaly Index for Gas Turbines Blade Fault Detection Under Multi-Operating Conditions
To meet industrial demands, gas turbines typically operate under multiple conditions, presenting unique challenges for fault diagnosis. This paper proposes a novel blade fault detection framework designed for such environments. First, a State-Supervised Variational Autoencoder (SS-VAE) model is introduced, which integrates the learning process of turbine operational states into the VAE bypass, enabling it to capture variations in vibration signal data across different operating conditions. Through the fusion decoding of operating states and vibration signals, the foundation of the SS-VAE model and its loss function is established. Secondly, a new State Mapping Error (SME) index is introduced to further address the issue of reconstruction error variability across multiple conditions. The proposed method is validated through a blade fault test rig and applied in an industrial gas turbine blade fault case. Results demonstrate that the SS-VAE model and SME index effectively capture vibration signal changes due to blade faults across multi-operating conditions, achieving accurate monitoring and diagnosis. Compared to traditional methods, the proposed approach shows higher effectiveness and robustness.
- 8 August 2024
- Published by Springer Nature in Applied Intelligence
- Vol. 54 (20) , 10148-10160
A fusion autoencoder model and piecewise anomaly index for aero-engine fault diagnosis
- 8 August 2024
- Published by Springer Nature in Applied Intelligence
- Vol. 54 (20) , 10148-10160
A fusion autoencoder model and piecewise anomaly index for aero-engine fault diagnosis
- Open Access
- 31 July 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Industrial Informatics
- Vol. 20 (11) , 13223-13231
An Improved Synchrosqueezing Transform: A Promising Tool for Reconstructing the Blade Passing Frequency Waveform of Gas Turbine
The blade passing frequency in the vibration signal of the casing is an important indicator for monitoring the blade status. To address the issue of how to extract the vibration response of the blades from the complex component of the casing signal, a vibration waveform reconstruction method based on improved synchrosqueezing transform was proposed. First, the time-frequency (TF) spectrum slice is used to adaptively estimate the bandwidth of the instantaneous frequency to determine the effective local maximum search range, which improves the TF concentration of results and effectively avoids the distortion of reconstructed waveform amplitude. Second, the target ridge can be extracted using multiple ridge detection tools. Finally, the signal reconstruction ability of the method is proved. Numerical validation and engineering cases have demonstrated that this method can extract the vibration response of the blades in the vibration signal of the casing.
- 31 July 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Industrial Informatics
- Vol. 20 (11) , 13223-13231
Open AccessAn Improved Synchrosqueezing Transform: A Promising Tool for Reconstructing the Blade Passing Frequency Waveform of Gas Turbine
The blade passing frequency in the vibration signal of the casing is an important indicator for monitoring the blade status. To address the issue of how to extract the vibration response of the blades from the complex component of the casing signal, a vibration waveform reconstruction method based on improved synchrosqueezing transform was proposed. First, the time-frequency (TF) spectrum slice is used to adaptively estimate the bandwidth of the instantaneous frequency to determine the effective local maximum search range, which improves the TF concentration of results and effectively avoids the distortion of reconstructed waveform amplitude. Second, the target ridge can be extracted using multiple ridge detection tools. Finally, the signal reconstruction ability of the method is proved. Numerical validation and engineering cases have demonstrated that this method can extract the vibration response of the blades in the vibration signal of the casing.
- 30 May 2024
- Published by Springer Nature
- p. 349-364
Autocorrelated Envelopes Assisted Adaptive MED for Gearbox Rolling Bearing Fault Diagnosis
- 30 May 2024
- Published by Springer Nature
- p. 349-364
Autocorrelated Envelopes Assisted Adaptive MED for Gearbox Rolling Bearing Fault Diagnosis
- Open Access
Gas Turbine Anomaly Detection under Time-Varying Operation Conditions Based on Spectra Alignment and Self-Adaptive Normalization
Gas turbine vibration data may exhibit considerable differences under time-varying conditions, which poses challenges for neural network anomaly detection. We first propose a framework for a gas turbine vibration frequency spectra process under time-varying operation conditions, assisting neural networks’ ability to capture weak information. The framework involves scaling spectra for aligning all frequency components related to rotational speed and normalizing frequency amplitude in a self-adaptive way. Degressive beta variational autoencoder is employed for learning spectra characteristics and anomaly detection, while a multi-category anomaly index is proposed to accommodate various operating conditions. Finally, a dataset of blade Foreign Object Damage (FOD) fault occurring under time-varying operating conditions was used to validate the framework and anomaly detection. The results demonstrate that the proposed method can effectively reduce the spectra differences under time-varying conditions, and also detect FOD fault during operation, which are challenging to identify using conventional methods.
Open AccessGas Turbine Anomaly Detection under Time-Varying Operation Conditions Based on Spectra Alignment and Self-Adaptive Normalization
Gas turbine vibration data may exhibit considerable differences under time-varying conditions, which poses challenges for neural network anomaly detection. We first propose a framework for a gas turbine vibration frequency spectra process under time-varying operation conditions, assisting neural networks’ ability to capture weak information. The framework involves scaling spectra for aligning all frequency components related to rotational speed and normalizing frequency amplitude in a self-adaptive way. Degressive beta variational autoencoder is employed for learning spectra characteristics and anomaly detection, while a multi-category anomaly index is proposed to accommodate various operating conditions. Finally, a dataset of blade Foreign Object Damage (FOD) fault occurring under time-varying operating conditions was used to validate the framework and anomaly detection. The results demonstrate that the proposed method can effectively reduce the spectra differences under time-varying conditions, and also detect FOD fault during operation, which are challenging to identify using conventional methods.
- Open Access
- 26 May 2023
- Published by Association for Computing Machinery (ACM)
- p. 170-179
Decoupling Transformer with Convolutional Fusion for Mechanical Composite Fault Diagnosis
- 26 May 2023
- Published by Association for Computing Machinery (ACM)
- p. 170-179
Open AccessDecoupling Transformer with Convolutional Fusion for Mechanical Composite Fault Diagnosis