Trends and Challenges in Intelligent Condition Monitoring of Electrical Machines Using Machine Learning

Round 1

Reviewer 1 Report

-English should be corrected
-please add colorful picture of measurements (optionally);;; + arrows what is what
-please add block diagram of the proposed research step by step ;;; what is the result of paper?;;;
-please add photo/photos of application of the proposed research ;;;; 
-please add sentences about future analysis;;;
-Figures should have better quality;;;; 
-Fonts of figures should be bigger;;;
-please add arrows to photos what is what;;; 
-formulas and fonts should be formatted;;;;
-references should be 2018-2021 Web of Science about 50% or more ;; 30 at least
-Please compare with other methods, justify. Advantages or Disadvantages different methods
for example about thermal imaging

1) Fault diagnosis of electric impact drills using thermal imaging, Measurement, Volume 171, 2021, 
https://doi.org/10.1016/j.measurement.2020.108815

 

2) Sun S, Przystupa K, Wei M, Yu H, Ye Z, Kochan O. Fast bearing fault
diagnosis of rolling element using Levy Moth-Flame optimization
algorithm and Naive Bayes. Eksploatacja i Niezawodnosc – Maintenance
and Reliability 2020; 22 (4): 730–740, http://dx.doi.org/10.17531/ein.2020.4.17

-Conclusion: point out what are you done;;;;
-is there possibility to use the proposed methods for other problems?

Author Response

Dear Reviewer,

Thank you for your valuable comments, which helped us to improve the paper. Please find attached answers to your comments in .doc file.

Sincerely,

Karolina Kudelina
Junior Researcher
Department of Electrical Power Engineering and Mechatronics
Tallinn University of Technology

Author Response File: Author Response.docx

Reviewer 2 Report

The paper provides a useful and review of the machine learning algorithm in intelligent condition monitoring of electrical machines. The paper’s topic is important and up to date, since electrical machines’ reliable operation is a crucial issue of the industry. The paper fills a gap since it has not published a similar review paper in this field. The paper is well organized and structured. It is easy to read. The figures are very informative and helpful to understand machine learning algorithms. 
Unfortunately. the paper has a very critical missing point. At the beginning of the paper, the authors summarise the most common condition monitoring techniques of electrical machines. However, the authors try to show a comprehensive review; the diagnosis of machines’ insulation is missing. Since from the point of view of the reliability of high and medium voltage apparatus, the most critical part is the insulation of electrical machines, transformers, and other equipment [1, 2]. Moreover, insulation is also a key question during the design process of electrical machines, and it has a significant effect on the manufacturing cost [3]. The insulation condition can be detected by measuring chemical, mechanical and electrical parameters of the insulating materials [4]. In electrical machines, the insulation diagnostic techniques can be classified into two groups: dielectric and partial discharge measurements. 
The dielectric measurements, including the insulation resistance [5], polarisation-depolarisation current measurement [6], hi-pot tests [7], tan delta [8], and return voltage [9], are capable of detecting the general degradation of the insulations. 
The partial discharge measurement techniques, besides the general condition, also provide information about the local faults; moreover, they enable online monitoring [10, 11, 12]. 
The paper can be acceptable considering the suggestions, but in this form, major revision is suggested.

1]  G. C. Montanari, “Envisaging links between fundamental research in electrical insulation and electrical asset management,” IEEE Electrical Insulation Magazine, vol. 24, no. 6, pp. 7–21, 2008. 
[2]  G. Stone, E. Boulter, I. Culbert, and H. Dhirani, Electrical Insulation for Rotating Machines: Design, Evaluation, Aging, Testing, and Repair. IEEE Press Series on Power Engineering, Wiley, 2014. 
[3]  T. Orosz, “Evolution and modern approaches of the power transformer cost optimization methods,” Periodica Polytechnica Electrical Engineering and Computer Science, vol. 63, no. 1, pp. 37–50, 2019. 
[4]  Z. A. Tamus, “Complex diagnostics of insulating materials in industrial electrostatics,” Journal of Electrostatics, vol. 67, no. 2, pp. 154–157, 2009. 11th International Conference on Electrostatics. 
[5]  H. Torkaman and F. Karimi, “Measurement variations of insulation resistance/polarization index during utilizing time in HV electrical machines - a survey,” Measurement, vol. 59, pp. 21–29, 2015. 
[6]  E. David, R. Soltani, and L. Lamarre, “PDC measurements to assess machine insulation,” IEEE Transactions on Dielectrics and Electrical Insulation, vol. 17, no. 5, pp. 1461–1469, 2010. 
[7]  G. C. Stone, “Recent important changes in IEEE motor and generator winding insulation diagnostic testing standards,” IEEE Transactions on Industry Applications, vol. 41, no. 1, pp. 91–100, 2005. 
[8]  S. A. Bhumiwat, “Field experience in insulation diagnosis of industrial high voltage motors using dielectric response technique,” in 2009 IEEE Electrical Insulation Conference, pp. 454–457, 2009. 
[9]  Z. Á. Tamus, “Combination of voltage response method with non-contact electrostatic voltage measurement to determine the dielectric response of insulating materials,” Journal of Physics: Conference Series, vol. 1322, p. 012042, Oct 2019. 
[10]  G. C. Stone, H. G. Sedding, and C. Chan, “Experience with online partial-discharge measurement in high-voltage inverter-fed motors,” IEEE Transactions on Industry Applications, vol. 54, no. 1, pp. 866–872, 2018. 
[11]  R. Ghosh, P. Seri, R. E. Hebner, and G. C. Montanari, “Noise rejection and detection of partial discharges under repetitive impulse supply voltage,” IEEE Transactions on Industrial Electronics, vol. 67, no. 5, pp. 4144–4151, 2020. 
[12]  S. B. Lee, G. C. Stone, J. Antonino-Daviu, K. N. Gyftakis, E. G. Strangas, P. Maussion, and C. A. Platero, “Condition monitoring of industrial electric machines: State of the art and future challenges,” IEEE Industrial Electronics Magazine, vol. 14, no. 4, pp. 158–167, 2020. 

Author Response

Dear Reviewer,

Thank you for your valuable comments, which helped us to improve the paper. Please find attached answers to your comments in .doc file.

Sincerely,

Karolina Kudelina
Junior Researcher
Department of Electrical Power Engineering and Mechatronics
Tallinn University of Technology

Author Response File: Author Response.docx

Reviewer 3 Report

The article may be a good introduction to the book about Machine Learning. There are good explanations about basic terms. For the “Review Article” readers expect that there will be a summary of actual knowledge (state of the art). Information what researchers do what progress in recent years. Unfortunately, the article lacks that kind of information.

Often the references are given in groups’ for example [1-5] or [16]-[18], [34-37]. It can be a good way to multiply the amounts of references, but this notation doesn’t provide information about what was presented in these references.

The title suggests that the article will be about trends and challenges in intelligent condition monitoring. In the whole article, there is no information about trends. This word was used only twice: in the title and once in the conclusion “Due to the trend of mounting sensors on …”

In Table 1 there are “+” that shows fault signals of different parts of electrical machines. The review in the case of bearing faules was based on only one article [11] – by the authors of this article. In reality, we have more fault symptoms of bearings (like temperature etc.).

The figures may look better. The neuron construction shown in Figure 14 looks oddly.

Author Response

Dear Reviewer,

Thank you for your valuable comments, which helped us to improve the paper. Please find attached answers to your comments in .doc file.

Sincerely,

Karolina Kudelina
Junior Researcher
Department of Electrical Power Engineering and Mechatronics
Tallinn University of Technology

Author Response File: Author Response.docx

Reviewer 4 Report

The paper aims to describe Machine learning technique to monitor electrical machines

The authors describe in a short but very understandable way the main approaches in machine maintenance followed by the main techniques of Machine Learning whose concepts are well known and easily available in literature.

Nevertheless, the paper do not give any example on their practical application in the fault analysis of electrical machines (as instead reported in the paper title). Therefore, because the paper reports only a list of machine learning approaches, it remains in a too general level without perform any deep a specific analysis on the topic reported in the title.

Some minor issues
Figure 1
The caption should describe in detail what is depicted in the figure.
line 50 why "energy systems"? is this an example? The authors should express general concepts and, only in a next step, examples 
can be useful to improve the comprehension.

line 93-96 It seems unbelievable that it is so difficult to obtain lists of faults from the companies.
I don't believe that it is required to break a machine to obtain the feature of a fault.

line 108 supervised?? or unsupervised...

Author Response

Dear Reviewer,

Thank you for your valuable comments, which helped us to improve the paper. Please find attached answers to your comments in .doc file.

Sincerely,

Karolina Kudelina
Junior Researcher
Department of Electrical Power Engineering and Mechatronics
Tallinn University of Technology

Author Response File: Author Response.docx

Round 2

Reviewer 1 Report

-Please compare with other methods, justify. Advantages or Disadvantages of thermal, acoustic, vibration fault diagnosis;;; 
for example 

1) Fault diagnosis of electric impact drills using thermal imaging, Measurement, Volume 171, 2021, 
https://doi.org/10.1016/j.measurement.2020.108815

2) 
Recognition of acoustic signals of induction motor using Fft, Smofs-10 and LSVM, Eksploatacja i Niezawodnosc – Maintenance and Reliability,
2015, 17 (4), pp. 569-574. https://doi.org/10.17531/ein.2015.4.12

-please add more photo of application of fault diagnosis. The paper should be interesting.

Author Response

Dear Reviewer,

Thank you once again for your valuable comments, which helped us to improve the paper. Please find attached .doc file with answers to your review.

Sincerely,

Karolina Kudelina

Junior Researcher
Department of Electrical Power Engineering and Mechatronics
Tallinn University of Technology

Author Response File: Author Response.docx

Reviewer 2 Report

Thank you for considering my suggestions. The paper is significantly improved. I understand your paper is intended to provide a review of AI methods in the field of electrical machine condition monitoring. My review's message does not run counter to the main purpose of the authors. Please consider several AI methods are used in the insulation diagnosis of electrical machines. Some papers:

Barrios S, Buldain D, Comech MP, Gilbert I, Orue I. Partial Discharge Classification Using Deep Learning Methods—Survey of Recent Progress. Energies. 2019; 12(13):2485. https://doi.org/10.3390/en12132485

Florkowski M. Classification of Partial Discharge Images Using Deep Convolutional Neural Networks. Energies. 2020; 13(20):5496. https://doi.org/10.3390/en13205496

Please check at least these papers. 

Author Response

Dear Reviewer,

Thank you once again for your valuable comments, which helped us to improve the paper. Please find attached .doc file with answers to your review.

Sincerely,

Karolina Kudelina

Junior Researcher
Department of Electrical Power Engineering and Mechatronics
Tallinn University of Technology

Author Response File: Author Response.docx

Reviewer 3 Report

OK.

Author Response

Dear Reviewer,

Thank you once again for your valuable comments, which helped us to improve the paper.

Sincerely,

Karolina Kudelina

Junior Researcher
Department of Electrical Power Engineering and Mechatronics
Tallinn University of Technology

Reviewer 4 Report

Despite the changes made to the paper, in my opinion this remain at a too general level with respect the aim reported in the title.
The examples of specific applications reported "at the end of each section" are only lists of some applications but very few details only are given for each of them and, absolutely not sufficient to understand in which way the specific ML approach has been used on the particular problem. As a consequence, the paper results only a general description of machine learning approaches without any link to the electric world.

A review should be a paper useful to allow the reader to find a possible solution of his/her problem on the specific issue. Without any specific information about the practical implementation of the ML approach on the electrical machine (this could be given also in a not detailed form) the paper results not useful for this goal.

Author Response

Dear Reviewer,

Thank you once again for your valuable comments, which helped us to improve the paper. Please find attached .doc file with answers to your review.

Sincerely,

Karolina Kudelina

Junior Researcher
Department of Electrical Power Engineering and Mechatronics
Tallinn University of Technology

Author Response File: Author Response.docx

Round 3

Reviewer 4 Report

The new version of the paper is now suitable for publication

Minor issues 
line  37 "TEAM stresses: thermal, electric, ambient, and mechanical stresses" --> "TEAM (Thermal, Electric, Ambient And Mechanical) stresses"
lines 38-53 Too short sentences without links among them. Please rephrase in a more organized and readable way
line 91     the acronyms (ML) should be defined at their first occurrence
Figure 14  I am writing and reading papers on NN since more than 30 years. Taking into account that NN are basically optimization algorithms, to date, in my opinion, the insertion of a picture of a biological neuron to state that their behavior is similar to it no more required and it becomes to be trivial. Repeat, this is only a my personal opinion but I suggest to remove such a figure

Author Response

Dear Reviewer,

We would like to thank you once again for your concerns and comments, which helped us to improve the paper. All the corrections are made and included into manuscript now.

Sincerely,

Karolina Kudelina

Junior Researcher
Department of Electrical Power Engineering and Mechatronics
Tallinn University of Technology

Author Response File: Author Response.docx