Online Rotor and Stator Resistance Estimation Based on Artificial Neural Network Applied in Sensorless Induction Motor Drive

Round 1

Reviewer 1 Report

Thank you for your detailed responses and additional information in the paper text. I think the authors are sufficiently responded all my previous comments.

Other minor comments:

1. Why is page 3 practically blank in this version?

1. A D-Space controller board is used to implement the proposed algorithm. I think it would be interesting to inform readers what types of industrial controllers can be used instead D-Space?

Author Response

Dear Reviewer,

Firstly, we would like to thank You for all remarks. We appreciate Your comments that helped to improve our paper. The respond to 2 other minor issue we would like to inform

#1

R: “1. Why is page 3 practically blank in this version?”

A: Thank You for this suggestion. It’s our mistake. In final version it’s correct.

#2

R:  “A D-Space controller board is used to implement the proposed algorithm. I think it would be interesting to inform readers what types of industrial controllers can be used instead D-Space?

A: Thank You for this suggestion. We decided to add additional text with 2 new literature review, that are as follow:

Application of the controller is known in literature for control of electronic converters like in [39], [40].

39. Isen, E.; Bakan, A.F. Development of 10 kW Three-Phase Grid Connected Inverter. Automatika 2016, 57, 319–328, doi:10.7305/automatika.2016.10.1081.
40. Zarkov, Z. Application of dSPACE Platform in the Study of Electric Generators with RES. In Proceedings of the 2018 20th International Symposium on Electrical Apparatus and Technologies (SIELA); IEEE, 2018; pp. 1–4.

Reviewer 2 Report

In the presented work, a new method is proposed for estimating the rotor and stator resistance of induction motor using artificial neural networks with the variable learning rate determined by a function instead of constant learning rate, which represents the novelty of the proposed method leading to higher quality of IM speed control.

The paper is theoretically easy to understand, contains a concise introduction with a necessary comparison with the current state of knowledge. The references are provided applicable and sufficient. The results are verified by simulations and experimental measurements but in my opinion the main drawback is that all case studies are presented for the quasi-steady operating state of the drive (constant drive speed, constant values of estimated resistances and constant load moment), from which it is not possible to assess the quality and stability of speed estimation in dynamic states. To support the conclusions it would be good to evaluate the properties of the proposed method for various operating states.

Other comments:

In row 199 there should be no dot after the word "value".

The vertical axis in Fig.7a is not described correctly.

Author Response

Firstly, we would like to thank You for all remarks. We appreciate Your comments that helped to improve our paper. The respond to 2 other minor issue we would like to inform

#1

R: “To support the conclusions it would be good to evaluate the properties of the proposed method for various operating states.”

A: The elements concerning properties of the proposed method for various operating states are included in conclusion as a toward to future research. The added text is “Future direct of the research will be properties evaluation of the proposed method for various operating states”

#2

R: “In row 199 there should be no dot after the word "value".”

A: We totally agree. We changed in final version of manuscript. The dot was delated

#3

R: “The vertical axis in Fig.7a is not described correctly.”

A: We totally agree. We changed in final version of manuscript. Fig. 7a was changed

Reviewer 3 Report

The authors in this paper propose a relatively known and old method for estimating the rotor and stator resistance of induction motor using artificial neural networks with the variable learning rate determined by a function. The results presented in this paper show that online rotor and stator resistance estimation using the proposed method is more accurate when estimating at a constant learning rate, and also demonstrates that the online estimation of rotor and stator resistance has contributed to improving the control quality of the sensorless induction motor drive. The overall approach in this article is well known and the only valuable information is experimental results. Everything else is already seen in the literature.

Author Response

Dear Reviewer,

Thank You for Your review. We are grateful for appreciate our experimental results. Thank You for positive assessment.

Reviewer 4 Report

1. In the paper, the methods of online rotor and stator resistance estimation based on artificial neural network applied in sensorless induction motor drive are presented. The simulation and experimental studies were performed.

2. The reviewed article was probably largely based on the doctoral dissertation, cited as Reference [2]. Many mistakes in the article are the same as in Ref.[2]. There is also the influence of Ref.[14] and [35] - coauthor of the doctoral dissertation.

3. The subject of the article is not new and has already been presented in many articles. The article was written by an international team of authors composed of 2 authors from Vietnam, 3 authors from Poland and 1 author from India. It was not stated on what basis the team was selected. None of the authors of the article has documented achievements in the field related to the subject of the article. The work of any of the authors of the article is not cited in the list of References in the article.

4. Why the methods of estimation presented in the article are limited only to the IFOC control ? What about DFOC and DTC control ?

5. It should be stated, how was the range of learning rate determined ?

6. The static neural network is applied for dynamic signal processing. Thus, the memory element was implemented and it can be considered as appropriate solution. Did the Authors notice significant problems with the algorithms for other values of sampling time?

7. The order of Keywords is not justified. The Keywords closely related to the topic of the article should be in the first place: Rotor Resistance Estimation, Stator Resistance Estimation, etc.

8. In the Section 1 Introduction the novelty of the proposed methods should be wider described.

9. The assumptions for model of induction motor are not described. It should be given that all variables and parameters of the rotor side were transformed on the stator side. Lm in Line 40 should be named magnetizing inductance, but not mutual inductance.

10. The statement on P.2, L.48, that the rotor flux angle depends on the rotor resistance is false.The slip rotational speed  of the rotor flux angle depends on the rotor resistance.  The rotor flux vector is rotating and the rotor flux angle  is the function of time. The citing of Ref.[2] in this case is not true. The Ref.[2] is dated on 2005 year, but the principles of Field Oriented Control were established many years earlier.

11. The statement on P.2, L.51-53: "Therefore, the accuracy of the estimated rotor and stator resistance will improve the accuracy of the estimated speed and the estimated rotor flux, thereby improving the quality of the sensorless drive control [4]." is based on Re.[4]. But the Ref.[4] concerns PMSMs, not IMs. 

12. The flux variables in Eq.(1) were mistakenly named as rotor leakage fluxes. These fluxes are not rotor leakage fluxes.

13. The flux variables in Eq.(2) were mistakenly named as stator leakage fluxes. These fluxes are not stator leakage fluxes.  The same name is repeated twice  on P.5, L.126.

14. The presentations of stator resistance change on P.9, L.207   and in Fig.7b   are  incompatible.  

15. It should be explained or commented why in Eq.(4) and in Eq.(14) we have operations on the alfa-components and the beta-components. Usually such components are independent.

16. The Section 5 Conclusions is too short - it includes only two sentences.

17. The references in the article are not formatted correctly.  

18. There are some grammar mistakes and some editorial mistakes in the article, that should be corrected.

The grammar and editorial mistakes are as follows:

P.1, L.18-19    The sentence with grammar mistake.      

P.1, L.19-20    The sentence with grammar mistake.

P.2, L.73            The sentence with grammar mistake.

P.2, L.74-75     The sentence is incomprehensible.  

P.2, L.82            The sentence with grammar mistake.      

P.2, L.66-67     The sentence with grammar mistake.      

P.2, L.71           the DREM technique   -  the abbreviation is not explained.

P.4, L.99            The sentence with grammar mistake.     

P.4, L.112- 113  The sentence with grammar mistake.     

P.11, L.236         The sentence with grammar mistake.      

Author Response

Firstly, we would like to thank You for all remarks. We appreciate Your comments that helped to improve our paper. The step by step respond is as follow:

#1

R: In the paper, the methods of online rotor and stator resistance estimation based on artificial neural network applied in sensorless induction motor drive are presented. The simulation and experimental studies were performed.

A: We agree that this paper presents such solutions.

#2

R: The reviewed article was probably largely based on the doctoral dissertation, cited as Reference [2]. Many mistakes in the article are the same as in Ref.[2]. There is also the influence of Ref.[14] and [35] - coauthor of the doctoral dissertation.

A: We just concern the proposition to compares results in our article.

#3

R: The subject of the article is not new and has already been presented in many articles. The article was written by an international team of authors composed of 2 authors from Vietnam, 3 authors from Poland and 1 author from India. It was not stated on what basis the team was selected. None of the authors of the article has documented achievements in the field related to the subject of the article. The work of any of the authors of the article is not cited in the list of References in the article.

A: We agree that this subject is deeply discussed in literature. However we performed new proposition and it was clearly verified in experimental part. We belief that self-citation are not encourage. Thus no citation are included. But to emphasis suitable to Energies journal, the citation from this journal are included.

#4

R:Why the methods of estimation presented in the article are limited only to the IFOC control ? What about DFOC and DTC control ?

A: During planning this article we decided to focus only on IFOC control. However we agree that expedition to DFOC and DTC control would be applied. However in our opinion it’s sufficient in this form.

#5

R: It should be stated, how was the range of learning rate determined ?

A: We agree. This information is included after equation 9. Please check line 142 in the revised manuscript.

#6

R:”The static neural network is applied for dynamic signal processing. Thus, the memory element was implemented and it can be considered as appropriate solution. Did the Authors notice significant problems with the algorithms for other values of sampling time?

A: During our research we hasn’t noticed any significant problems with the algorithms for other values of sampling time

#7

R: The order of Keywords is not justified. The Keywords closely related to the topic of the article should be in the first place: Rotor Resistance Estimation, Stator Resistance Estimation, etc.

A” Thank You for this remark. We reorganized keywords. Now its: “Rotor Resistance Estimation; Stator Resistance Estimation; Sensorless Control; Artificial Neural Network (ANN); Indirect Field Oriented Control (IFOC).”

#8

R: “the Section 1 Introduction the novelty of the proposed methods should be wider described.”

A: Thank You for this remark. We decided to change the lines 107-109 “In this article, authors proposed a novel method to estimate rotor resistance and stator resistance using artificial neural networks with the learning rate as a function. That is an extension to previous researches to improving the control quality of the sensorless induction motor drive.”

#9

R: “The assumptions for model of induction motor are not described. It should be given that all variables and parameters of the rotor side were transformed on the stator side. Lm in Line 40 should be named magnetizing inductance, but not mutual inductance.”

Thank You for this remark. We changed name to “magnetizing inductance,”

#10

R: The statement on P.2, L.48, that the rotor flux angle depends on the rotor resistance is false.The slip rotational speed  of the rotor flux angle depends on the rotor resistance.  The rotor flux vector is rotating and the rotor flux angle  is the function of time. The citing of Ref.[2] in this case is not true. The Ref.[2] is dated on 2005 year, but the principles of Field Oriented Control were established many years earlier.

A: Thank You for this remark. According to Your proposition we decided to include old references as follows:

2. Sathikumar, S.; Vithayathil, J. Digital Simulation of Field-Oriented Control of Induction Motor. IEEE Trans. Ind. Electron. 1984, IE-31, 141–148, doi:10.1109/TIE.1984.350058.
3. Gabriel, R.; Leonhard, W.; Nordby, C.J. Field-Oriented Control of a Standard AC Motor Using Microprocessors. IEEE Trans. Ind. Appl. 1980, IA-16, 186–192, doi:10.1109/TIA.1980.4503770.
4. Takahashi, I.; Noguchi, T. A New Quick-Response and High-Efficiency Control Strategy of an Induction Motor. IEEE Trans. Ind. Appl. 1986, IA-22, 820–827, doi:10.1109/TIA.1986.4504799.
5. Murata, T.; Tsuchiya, T.; Takeda, I. Vector control for induction machine on the application of optimal control theory. IEEE Trans. Ind. Electron. 1990, 37, 283–290, doi:10.1109/41.103414.

#11

R: The statement on P.2, L.51-53: "Therefore, the accuracy of the estimated rotor and stator resistance will improve the accuracy of the estimated speed and the estimated rotor flux, thereby improving the quality of the sensorless drive control [4]." is based on Re.[4]. But the Ref.[4] concerns PMSMs, not IMs. 

A: “ Thank You for this remark. We changed reference to “Pal, A.; Das, S. Development of energy efficient scheme for speed sensorless induction motor drive. Int. Trans. Electr. Energy Syst. 2020, 30, doi:10.1002/2050-7038.12448.”

#12 and #13

R: The flux variables in Eq.(1) were mistakenly named as rotor leakage fluxes. These fluxes are not rotor leakage fluxes. The flux variables in Eq.(2) were mistakenly named as stator leakage fluxes. These fluxes are not stator leakage fluxes.  The same name is repeated twice  on P.5, L.126

A: According our best knowledge, the names are correct.

#14

R: “The presentations of stator resistance change on P.9, L.207   and in Fig.7b   are  incompatible.  

A: “The range in text about stator resistance in both figure and text are from 1.99 to 2.99 so according to our best knowledge is compatible””

#15

R: It should be explained or commented why in Eq.(4) and in Eq.(14) we have operations on the alfa-components and the beta-components. Usually such components are independent.

A: Thank You for the remark. We added text after equation 4 line 134

#16 The Section 5 Conclusions is too short - it includes only two sentences.

Thank You for this remark. We extended conclusion part. The added text is in lines 301-302. The text is:

“Future direct of the research will be properties evaluation of the proposed method for various operating states”

#17

R:The references in the article are not formatted correctly.  

A: The references were formatted using MDPI temple for Mendeley software. So we belief that now it’s correct.

# 18

R: There are some grammar mistakes and some editorial mistakes in the article, that should be corrected.

Thank You for this remark. Article was proofreader.

#19

We included all specific remarks in text. All changes are indicated in track change mode.