Analysis and Adaptive Separation of IGBT Switching Noise in PD Monitoring of Flexible HVDC Valves: An Evolutionary Perspective

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

1. Describe the software program used for simulation of spectral characteristics.
2. Explain if Figure 5 has been generated with only 5-6 measurement points. If this is true, it is advised to increase the number of measurement points.
3. It's challenging to compare Figure 5, which shows both turn-on and turn-off, to Figure 6. Furthermore, the frequency domains in these two pictures differ.
4. Figures 7a) and b) should be displayed at the same band frequency. The same extends for Figures 8, 9, and 10.
5. Describe the differences between Figures 8, 9, 10, normal, and Figure 5.
6. Most figures can be improved.
7. Update the reference section according to the current article domain. Seven of the 16 references are older than fifteen years.
8. This scientific originality should be explained, especially in the CONCLUSIONS section, requiring additional improvement.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

Dear Authors,

I have carefully reviewed your research paper titled "Analysis and Adaptive Separation of IGBT Switching Noise in PD Monitoring of Flexible HVDC Valves: An Evolutionary Perspective". Below are the specific revision suggestions for your reference.

This study focuses on revealing the generation mechanism, spectral characteristics, and evolutionary laws of high-frequency switching noise in insulated-gate bipolar transistors (IGBTs) during device aging, while clarifying the intrinsic nature of overlapping interference between this noise and the ultra-high-frequency (UHF) band used for partial discharge (PD) detection. The core research question directly addresses the key technical bottleneck in online PD monitoring of converter valves in flexible DC transmission systems. The topic exhibits clear relevance and certain originality. The main contribution of this research lies in breaking through the limitations of traditional signal separation methods, which rely on fixed frequencies or static feature thresholds and thus lack robustness in scenarios involving aging-induced spectral drift. By constructing feature parameters based on the fundamental differences in energy accumulation processes, this study provides a new approach for signal separation in dynamic interference environments.

Regarding the existing issues in the paper, the following suggestions are proposed:

1.The current simulation model adopts a 1200V/100A IGBT module as a scaled-down platform. Although it can reveal basic physical mechanisms, there are differences in packaging structure and parasitic parameter distribution compared to the high-voltage press-pack IGBTs actually used in ultra-high-voltage flexible DC converter valves. It is recommended to supplement simulation verification of high-voltage IGBT modules or establish the correlation between the spectral characteristics of low-voltage modules and high-voltage devices through multi-scale modeling methods to enhance the engineering applicability of the model.

2.The simulation of bond wire aging and gate oxide layer degradation is relatively simplified, failing to consider the nonlinear dynamic changes of parasitic parameters during the aging process. It is suggested to introduce a parameter model that is more consistent with the actual aging physical process to improve the consistency between simulation and experiment.

3.The experimental design does not clearly specify the computational complexity and real-time performance indicators of the adaptive optimization process. Real-time performance is crucial for engineering field applications. It is recommended to supplement the time complexity analysis of the algorithm and optimize the parameter search strategy to improve real-time processing capabilities.

4.The algorithm design does not clearly specify the computational complexity and real-time performance indicators of the adaptive optimization process. Real-time performance is crucial for engineering field applications. It is recommended to supplement the time complexity analysis of the algorithm and optimize the parameter search strategy to improve real-time processing capabilities.

5.The conclusion does not clearly explain how to combine spectral evolution characteristics with aging diagnosis thresholds, nor has field testing in actual converter valve scenarios been conducted, leading to insufficient engineering applicability of the conclusions. It is recommended to supplement the method for determining aging diagnosis thresholds or preliminary results of on-site pilot applications, so that the conclusions can better respond to practical engineering needs.

6.There are relatively few references on the application of time-frequency domain cumulative energy functions in signal separation, which fails to fully reflect the positioning of this method among similar feature extraction technologies. It is suggested to supplement classic methods and literature on the latest progress in related fields to strengthen the innovative comparison of the feature parameters in this study. Additionally, there is a lack of targeted literature citations on the switching noise characteristics of high-voltage press-pack IGBTs. Most existing literatures focus on ordinary IGBT modules, which have insufficient relevance to the research application scenarios. It is recommended to supplement the latest research results related to high-voltage IGBTs.

7.It is recommended to add quantitative comparison charts of separation effects between traditional methods (equivalent time-frequency analysis, wavelet principal component analysis) and the proposed method (e.g., bar charts of separation accuracy and misclassification rate) instead of only displaying feature distributions, so as to make the performance advantages more intuitive.

We suggest that you revise and improve the paper according to the above comments, supplement relevant data and analyses, and enhance the academic quality and readability of the paper.

Comments on the Quality of English Language

Overall, the English writing of this paper basically meets the core requirements for academic communication in international journals. The core technical concepts, research framework and case results are clearly conveyed, without any major obstacles to understanding. However, the selection of individual words is not precise enough, and there are redundant modifiers. It is recommended that you clearly define abbreviations when they first appear and ensure consistent usage throughout the paper. prioritize the adoption of precise academic vocabulary in the field and eliminate meaningless redundant expressions to make the presentation more concise and professional.

 

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

This new version of the paper satisfies my suggestions and comments, and therefore I consider that it can be accepted in this present form.

Author Response

We would like to express our sincere gratitude to the reviewer for re-evaluating our manuscript. We are delighted to learn that the revised version has satisfied the reviewer's concerns and is recommended for acceptance. We truly appreciate the time and constructive comments provided during the review process, which have helped us improve the quality of our paper.

Reviewer 2 Report

Comments and Suggestions for Authors

Dear Authors,

I have carefully reviewed the research paper titled "Analysis and Adaptive Separation of IGBT Switching Noise in PD Monitoring of Flexible HVDC Valves: An Evolutionary Perspective".This study focuses on revealing the generation mechanism, spectral characteristics, and evolutionary laws of high-frequency switching noise in insulated-gate bipolar transistors (IGBTs), clarifying the intrinsic nature of overlapping interference between this noise and the ultra-high-frequency (UHF) band used for partial discharge (PD) detection. The core research question directly addresses the key technical bottleneck in online PD monitoring of converter valves in flexible DC transmission systems. The topic exhibits clear relevance and certain originality.

The main contribution of this research lies in addressing the limitations of traditional signal separation methods, which rely on fixed frequencies or static feature thresholds and thus lack robustness in scenarios involving aging-induced spectral drift. By constructing feature parameters based on the fundamental differences in energy accumulation processes, this study breaks through the constraints of conventional approaches and provides a new idea for signal separation in dynamic interference environments.

The authors have conducted systematic revisions to address all the previously raised issues. Below are supplementary comments based on the revision progress:

1.The authors have supplemented the simulation verification of high-voltage press-pack IGBTs (PPIs). However, it is necessary to further clarify the basis for the scaling relationship in multi-scale modeling. It is recommended to add the mapping formula or verification data for the spectral characteristics between low-voltage modules and high-voltage PPIs to demonstrate the scientific validity of extending the conclusions from low-voltage simulations to high-voltage scenarios.

2.The authors have supplemented the time complexity analysis. It is further recommended to add the real-time performance under extreme operating conditions (e.g., signal saturation, superimposed strong electromagnetic interference) to verify the algorithm's robustness in complex on-site engineering environments.

3.The authors have supplemented the quantitative comparison charts of separation effects. It is suggested that these charts should reflect data reliability. Additionally, a "method flow chart" should be added to intuitively demonstrate the complete process from signal acquisition, time-frequency cumulative energy calculation, feature extraction, and adaptive optimization to clustering separation, thereby further enhancing the demonstration of data reliability.

After the authors make the supplementary revisions as suggested above, the paper may be considered for acceptance and publication.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf