Design Optimization of a Switched Reluctance Machine with an Improved Segmental Rotor for Electric Vehicle Applications

Round 1

Reviewer 1 Report

 

I am not working in this area, so can only comment on the paper as presented, rather than comparing it to the state of the art in the area.

Some specific comments:

On my pdf version the title for fig1 is on a different page. 1b shows conventional concentrated windings, but in 1a it is not clear how the windings are physically formed. It appears that the end turns go half way round the stator, which is not desirable from various points of view. One of the interesting facts missing from Table 1 is how much copper is needed. It would be better if 1a and 1b were interchanged to match fig 2. In fig. 3 you can explicitly say that the previous phase is F.

In line 103 FEM with the optimization method is mentioned. What is this? This is a key part of the contribution. This is then talked about in Table 2, but golden sections and parabolic interpolation are mentioned with no reference. It is interesting that the optimization seems to apply to both rotor and stator, so a few more words on how the possible combinations were optimized would be interesting.

In section 4 the cooling system is discussed, bur it is not clear if anything is being optimized, whether any rotor losses are included, or the rotor is considered at all. It seems that the end turns are the hot spots, but we have not been told anything about these, or anything about the thermal coupling of the various components to the jacket.

 

In section 6 we get on to the interesting experimental results, except the interesting details are missing. I don't need to know about the DC machine load except that I am surprised you are just using a resistor as the load. You say nothing about how you obtain what appear to be the very good static measurements of fig 16. Fig 18 shows experimental torque ripples for different machines. How were these measured? The implication is that you actually replaced the new machine with a conventional one to obtain the comparison.

Overall the paper is readable and well presented. However, the title is a bit misleading. The key point should be that the low speed ripple has been reduced. How the new design has changed the power rating and efficiency for a given frame size is of major practical interest, but does not seem to be mentioned.

Author Response

Dear reviewer, these are the responses.

Reponses to reviewer 1

Point 1): On my pdf version the title for fig1 is on a different page.
Response: It is updated to the correct position.
Point 2): 1b shows conventional concentrated windings, but in 1a it is not clear how the windings are physically formed. It appears that the end turns go half way round the stator, which is not desirable from various points of view.
Response: The figure with winding is added in Figure 1. The end turns indeed go half way round the stator, it reduced the efficiency compare to the concentrated windings from the simulation results. The last response shows the detail of the efficiency comparison.
Point 3): One of the interesting facts missing from Table 1 is how much copper is needed.
Response: The copper winding length values for the two motors are added to Table 1.
Point 4): It would be better if 1a and 1b were interchanged to match fig 2.
Response: The figures are interchanged.
Point 5): In fig. 3 you can explicitly say that the previous phase is F.
Response: The fig. 3a is updated with ‘phase F’.
Point 6): In line 103 FEM with the optimization method is mentioned. What is this? This is a key part of the contribution. This is then talked about in Table 2, but golden sections and parabolic interpolation are mentioned with no reference.
Response: The optimization was running with MATLAB optimization toolbox. The MATLAB command fminbnd was used. It is the function to find minimum of single-variable function on fixed interval. The reference papers on the optimization method  are added as [26, 27] on the line 161 when the Microsoft Word showing the Track Changes.
Point 7): It is interesting that the optimization seems to apply to both rotor and stator, so a few more words on how the possible combinations were optimized would be interesting.
Response: The explanation is added from line 178 to 183.
Point 8): In section 4 the cooling system is discussed, but it is not clear if anything is being optimized, whether any rotor losses are included, or the rotor is considered at all. It seems that the end turns are the hot spots, but we have not been told anything about these, or anything about the thermal coupling of the various components to the jacket.
Response: There is no optimization is because the original design of the cooling system is proper for the SRM, as shown in the paper. Because the rotor of the SRM only has iron losses, but without copper losses, so the rotor losses are not included. The cooling jacket is connected to the stator yoke directly. The liquid channel is make on the jacket directly. So the coupling between the cooling jacket and the stator is direct contact. The coupling between the jacket and the cooling liquid is direct contact. 
Some explanation sentences are added from line 211 to 203.
Point 9): In section 6 we get on to the interesting experimental results, except the interesting details are missing. I don't need to know about the DC machine load except that I am surprised you are just using a resistor as the load. You say nothing about how you obtain what appear to be the very good static measurements of fig 16.
Response: The DC machine shaft is connected to the SRM shaft. The field winding of the DC machine is connected to a DC supply. The armature winding of the DC machine is connected to the resistor bank. When the SRM drive the DC machine to rotate, it generate DC voltage from the armature winding. This DC voltage is connected to the resistor bank to dissipate the energy.
The static measurement was done with a rotor block system. The rotor block system is shown in the following figure. 
 
With the system, the rotor position is fixed at different angle. The angle value is read from the encoder, the torque value is read from the torque sensor. The flux linkage data is calculated form the static torque data. The calculation equations are shown in the paper as equation (3) and (4).
Point 10): Fig 18 shows experimental torque ripples for different machines. How were these measured? The implication is that you actually replaced the new machine with a conventional one to obtain the comparison.
Response: The torque waveforms are measured with a torque sensor and sensor reader on the shaft of the motor. The torque data is collected from the reader with software.
Yes, the new machine is indeed replaced with the conventional one when I do the test. Only the motor is replaced and other parts are same for the two motors, including torque sensor, driver, controller, and load, etc.
Point 11): Overall the paper is readable and well presented. However, the title is a bit misleading. The key point should be that the low speed ripple has been reduced.
Response: The title is updated from ‘Design of a Switched Reluctance Machine with an Improved Segmental Rotor for Electric Vehicle Applications’ to ‘Design optimization of a Switched Reluctance Machine with an Improved Segmental Rotor for Electric Vehicle Applications’.
The title is decided by my supervisor. Without the permission from my supervisor, I cannot change the title.
Point 12): How the new design has changed the power rating and efficiency for a given frame size is of major practical interest, but does not seem to be mentioned.
Response: This proposed SRM is the small scaled 2.5kW SRM. It is from a full scale SRM, which is 80kW. The full scale SRM has an exact same structure as this proposed SRM, but only the rated power and size are different. 
For the full scaled 80kW SRM, the efficiency is analysed. The following figure shows the efficiency analysed of the full scaled 80kW proposed SRM. It can be seen that the proposed 80kW SRM has 2% efficiency lower than the reference SRM at the high efficiency range. 
The efficiency analysis for the scaled 2.5kW proposed SRM is added to the future works from line 302 to 303.
 

 

Author Response File: Author Response.docx

Reviewer 2 Report

This paper proposes a novel SRM design with an asymmetrical rotor for electric vehicles. A restricted experimental verification is also provided. There is a number of unclear points: 

1) The term “engine” does not apply to an electric motor. Please replace.

2) For what type of electric vehicle is it supposed to use the proposed SRM. Please add to the article the required dependency of torque in rotational speed for the motor of the vehicle in question. What constant power speed range is required?

3) There are a number of articles that also propose the SRM design with asymmetrical rotor poles, which the authors do not mention in the introduction. Please add references and discuss such articles in the introduction to better explain the novelty and advantages of your proposed rotor asymmetry.

4) Add to the paper a theoretical comparison of the torque waveform between the conventional SRM and the proposed asymmetrical rotor SRM.

5) Why is the theoretical torque ripple shown in Fig. 14 and the experimental one shown in 18 are so different? How were the waveforms of the motor torque measured in the experiment at different speeds?

6) It would be instructive to compare the main characteristics (efficiency, torque ripple, dimensions, etc.) and the maximum torque versus speed curves for the permanent magnet motor used in the application under consideration and for the proposed SRM. It would also be instructive to compare the current rating and cost of the inverter power modules used.

 

Author Response

Dear reviewer, these are the responses:

Point 1): The term “engine” does not apply to an electric motor. Please replace.
Response: The “engine” is updated to “machine” in the paper at line 27 when the Microsoft Word showing the Track Changes.
Point 2): For what type of electric vehicle is it supposed to use the proposed SRM. Please add to the article the required dependency of torque in rotational speed for the motor of the vehicle in question. What constant power speed range is required?
Response: The pure electric vehicle is supposed to use the proposed SRM. The rated power is 80kW for the full scale SRM. But for this paper, the proposed SRM is the scaled 2.5kW SRM. The rated torques and speeds for the 2.5kW and 80kW SRM are added.
The explanation sentences are added to the paper from line 76 to 79.
Point 3): There are a number of articles that also propose the SRM design with asymmetrical rotor poles, which the authors do not mention in the introduction. Please add references and discuss such articles in the introduction to better explain the novelty and advantages of your proposed rotor asymmetry.
Response: The references and discussion on asymmetric rotor SRM are added from line 51 to 63.
Point 4): Add to the paper a theoretical comparison of the torque waveform between the conventional SRM and the proposed asymmetrical rotor SRM.
Response: The theoretical comparison is added from line 119 to 124.
Point 5): Why is the theoretical torque ripple shown in Fig. 14 and the experimental one shown in 18 are so different? How were the waveforms of the motor torque measured in the experiment at different speeds?
Response: I think the difference between the Fig. 14 and 18 is because the simulation for Fig. 14 doesn’t have electromagnetic noise, but the 18 has numbers of electromagnetic noise from the real world test.
The motor torque is measured with a torque sensor on the shaft of the motor. A software with a reader box from the torque sensor company is used.
The following figure shows the torque sensor and the sensor reader.
Point 6): It would be instructive to compare the main characteristics (efficiency, torque ripple, dimensions, etc.) and the maximum torque versus speed curves for the permanent magnet motor used in the application under consideration and for the proposed SRM. It would also be instructive to compare the current rating and cost of the inverter power modules used.
Response: The sentences about the dimension and power comparison are added from line 282 to 284. The efficiency and cost analysed sentence is added to the future works from line 302 to 303. This proposed SRM is a 2.5kW small scaled SRM from a 80kW full scaled SRM. 
The full scaled SRM is referenced from a permanent magnet motor. The permanent magnet motor is used on an EV from the market. For the full scaled 80kW SRM, the efficiency is analysed. It is compared to a three-phase SRM. Compared to the three-phase SRM, the efficiency is 2% lower at the high efficiency area from the simulation results. The following figure shows the efficiency comparison.

Author Response File: Author Response.docx

Round 2

Reviewer 1 Report

The details on the testing methods could still be improved. I don't need to know that the energy goes into a resistor bank, but you still say very little about the way the static and dynamic tests were actually done.

Fig2 seems to have been modified without deleting the original.

 

Fig 1a does imply how the end turns are connected, but the numbers are difficult to believe. In a conventional machine the end turns are around one slot whereas in the proposed machine they are half way round the stator. With the given stack length of 120mm and the approximate diameter of 80mm. the length of a loop in the proposed machine is 400mm, assuming the impossible situation of the end turn going the shortest route through the middle of the shaft. In the conventional machine, even if the slots are full of copper the loop length should be more like 300mm or about 75% of the total copper. Routing 6 slots of copper round the circumference will take more space so the give ratio of 37.12:32 needs clearer explanation. You say you have built it so a picture of these end turns would help. It would be more interesting than 12a.

Author Response

Thank you to the reviewer to help me check my paper, and help to make my paper and research better. Here are my responses to the comments.
Point 1): The details on the testing methods could still be improved. I don't need to know that the energy goes into a resistor bank, but you still say very little about the way the static and dynamic tests were actually done.
Answer:    
a)    Static test:
The static tests were done with only current source, torque sensor and position sensor. The encoder was used as position sensor. The Figure 1 shows the diagram of the static test. With a rotor blocking system, the motor shaft was fixed at different positions. The position value was read from the encoder. The torque value was read from the torque sensor and reader box.
 
The Figure 2 shows the setup to do the static test.
 
The Figure 16a in the paper is updated to integrate the torque sensor position to explain how the torque is measured. The torque sensor is added to the figure to explain how the static data is measured. The sentence is added from line 270 to 271.
b)    Dynamic test:

The dynamic tests were done as the diagram showing in Figure 3. When the SRM is running with the driver, the dynamic torque is measured by software on the PC. The output signal from the torque sensor is an analog signal. The software on the PC converts the analog signal to a digital signal with an analog-to-digital-converter (ADC).
 
Figure 4 shows the setup to do the dynamic test and measurement.
c)    Torque waveform reconstruction:
Because the torque waveforms are measured on the shaft, so the torque is smoothed by the rotor of the SRM. In order to estimate the accurate torque waveforms, I did some experiment test again and reconstructed the torque waveforms according to the Fig. 16 in the reference paper https://doi.org/10.3390/machines9120348 . 
Figure 18 in the paper is updated with new current waveforms. Figure 19 and Figure 21 in the paper are added. Figure 19 presents the current waveforms for the reference SRM. Figure 21 presents the torque reconstruction results at 1500rpm speed. The detail is explained from line 309 to line 320.
Point 2): Fig2 seems to have been modified without deleting the original.
Answer: 
The original one is deleted. The original one is still visible with “All markup” showing when the “Track Changes” is on. In this way, the editor can easily know where it was revised.
Point 3): Fig 1a does imply how the end turns are connected, but the numbers are difficult to believe. In a conventional machine the end turns are around one slot whereas in the proposed machine they are half way round the stator. With the given stack length of 120mm and the approximate diameter of 80mm. the length of a loop in the proposed machine is 400mm, assuming the impossible situation of the end turn going the shortest route through the middle of the shaft. In the conventional machine, even if the slots are full of copper the loop length should be more like 300mm or about 75% of the total copper. Routing 6 slots of copper round the circumference will take more space so the give ratio of 37.12:32 needs clearer explanation. You say you have built it so a picture of these end turns would help. It would be more interesting than 12a.
Answer: 
The Figure 5 shows the calculation of the copper length:
a) For the reference SRM:
Only half of each slot is filled with winding, the number of turns per phase is 32. The stack length is 120mm. Assume that the end winding length is small and about 10mm. There are 4 poles with winding. The total length of the copper wire per phase is (120*2+10)*32*4 = 32m.
b) For the proposed SRM:
Full slot is filled with one winding, then the number of turns per phase is doubled to 64. The stack length is 120mm. The end winding length is calculated with the CAD software. As shown in the Figure 4b, the end winding length per phase is (175.198+143.327)/2*2 =318mm. Because the turning corner is not accurate, so I modified the 318mm to 340mm. In this way, the total length of the copper wire per phase is (120*2+340)*64=37.12m
      
The photos of the windings are added in the Figure 6. Figure 6 shows the windings for the reference conventional SRM and the proposed SRM.
And thanks to the reviewer to point this problem. Indeed, the proposed SRM has too long end windings. When I start to design this SRM long time ago, I tried so many solutions. But only this fully-pitched winding configuration can make the proposed SRM working with two phases simultaneously. It means that only this type of fully-pitched winding configuration can make the proposed SRM working to reduce the torque ripple. This is the main disadvantage of this proposed SRM. The main disadvantage is the lower efficiency because of the long end winding.

Author Response File: Author Response.docx

Reviewer 2 Report

Thank you for your responses. But, unfortunately, some responses to the reviewer's comments are unsatisfactory:

2) The authors have not responded to my comment: “Please add to the article the required dependency of torque in rotational speed for the motor of the vehicle in question. What constant power speed range is required?”

The only have noted that “The rated torques and speeds for the 2.5kW and 80kW SRM are added.”

However, a feature of motors for electric vehicles is that they must provide a certain range of the vehicle speed control, even without the use of the gearbox. This fact strongly influences the criteria for optimizing such motors. For example, in this application, motors with magnets on the rotor  surface are practically not used, but motors with magnets inserted in the rotor (interior permanent magnet motors, IPM), which have a substantial reluctance torque, are able to provide a wide constant power range (CPSR).

Please add to the article information about the required torque-speed curve and CPSR for the motor of the vehicle in question. See examples of such characteristics in doi.org/10.1109/TIA.2012.2227092, fig. 2; or in https://upcommons.upc.edu/handle/2117/8246, fig.1. 

3) The authors write: “The references and discussion on asymmetric rotor SRM are added from line 51 to 63”. 

Please, not only mention these studies, but also explain what is the novelty of the rotor asymmetry proposed by you in comparison with them.

4) The authors write: “The theoretical comparison is added from line 119 to 124” and “This working method of operation reduces the torque ripple.” 

I believe, the conclusions about the reduction of torque ripples when using the proposed design cannot be made based on figure 2. Does your motor model take into account torque ripple from the interaction of the stator and rotor teeth? How much torque ripple has been reduced (in N*m, in %)? Plot the torque waveforms under comparison in one axis.

5) The authors write: “I think the difference between the Fig. 14 and 18 is because the simulation for Fig. 14 doesn’t have electromagnetic noise, but the 18 has numbers of electromagnetic noise from the real world test.”

The reviewer believes that the motor torque ripple cannot be measured by the applied method, since when the shaft rotates at high speed, the instantaneous torque of the sensor does not correspond to the motor torque, since the torque ripple at the sensor are smoothed out by the elasticity of the mechanical transmission. The torque ripple of the motor itself can only be measured at a rotational speed close to zero, using a test bench without a long mechanical transmission.

For SRM, this measurement is difficult to perform due to the inherently oscillatory operation of the motor. But this issue should at least be discussed in the article how this was done, for example, in doi.org/10.3311/PPee.14012.

At least, it is possible to compare the static characteristics of the motor torque with the experiment, as shown in Fig. 10 at doi.org/10.3311/PPee.14012.

6) The authors write: “The sentences about the dimension and power comparison are added from line 282 to 284. The efficiency and cost analysed sentence is added to the future works from line 302 to 303. This proposed SRM is a 2.5kW small scaled SRM from a 80kW full scaled SRM. 

The full scaled SRM is referenced from a permanent magnet motor. The permanent magnet motor is used on an EV from the market. For the full scaled 80kW SRM, the efficiency is analysed. It is compared to a three-phase SRM. Compared to the three-phase SRM, the efficiency is 2% lower at the high efficiency area from the simulation results. The following figure shows the efficiency comparison”.

Please note that the question was not about comparing two SRM designs, but about comparing an optimized SRM with a state-of-art interior permanent magnet motor used in the target application (target electric vehicle).

Author Response

Point 1): Unfortunately, some responses to the reviewer's comments are unsatisfactory:
Answer: 
Thanks a lot for the reviewer to read my article and comment on it. The comments help me to improve the paper and improve my research. In the following paragraphs, I elaborate further on my responses.
Point 2): The authors have not responded to my comment: “Please add to the article the required dependency of torque in rotational speed for the motor of the vehicle in question. What constant power speed range is required?”
The only have noted that “The rated torques and speeds for the 2.5kW and 80kW SRM are added.”
However, a feature of motors for electric vehicles is that they must provide a certain range of the vehicle speed control, even without the use of the gearbox. This fact strongly influences the criteria for optimizing such motors. For example, in this application, motors with magnets on the rotor  surface are practically not used, but motors with magnets inserted in the rotor (interior permanent magnet motors, IPM), which have a substantial reluctance torque, are able to provide a wide constant power range (CPSR).
Please add to the article information about the required torque-speed curve and CPSR for the motor of the vehicle in question. See examples of such characteristics in doi.org/10.1109/TIA.2012.2227092,  fig. 2; or in https://upcommons.upc.edu/handle/2117/8246, fig.1.
Answer:
The figure is added as Figure 2 in the paper. One paragraph before Figure 2 from line 109 to 112 is added to explain the figure.
The Figure 2 in the paper:

The paragraph added in the paper:
“Figure 2 Presents the torque speed specification for the 80kW full scale SRM and the 2kW small scale proposed SRM. For the both SRMs, the rated speeds are 3000rpm, and the maximum speeds are 6000rpm. For the full scale SRM, the rated torque is 254Nm. The rated torque is 6.5Nm for the small scale SRM.”
Point 3): The authors write: “The references and discussion on asymmetric rotor SRM are added from line 51 to 63”.
Please, not only mention these studies, but also explain what is the novelty of the rotor asymmetry proposed by you in comparison with them.
Answer:
The sentences are added from line 63 to 68.
The added sentences:
“In general, the steps on the asymmetric rotor or stator poles increased the total reluctance between the rotor and stator poles when the rotor pole is on the aligned position. The proposed SRM in this paper doesn’t have the step structure and have identical airgap length when the rotor pole is on the aligned position. It is an advantage to decrease the reluctance and increase the output torque, compared to the asymmetric rotor or stator structure.
Point 4): The authors write: “The theoretical comparison is added from line 119 to 124” and “This working method of operation reduces the torque ripple.”
I believe, the conclusions about the reduction of torque ripples when using the proposed design cannot be made based on figure 2. Does your motor model take into account torque ripple from the interaction of the stator and rotor teeth? How much torque ripple has been reduced (in N*m, in %)? Plot the torque waveforms under comparison in one axis.
Answer: 
a) The Figure 3 in the paper is based on the FEM simulation results. The following figures show the FEM simulation with maximum current.
From the Fig. 2, it can be seen that, the reference SRM has higher torque ripple ratio, compared to the proposed SRM.
b) The motor model takes into account the torque ripple from the interaction of the stator and rotor teeth. Because this is switched reluctance motor, there is no magnet in the motor. And this motor only works with the reluctance torque. It means that the torque is only the interaction of the stator and rotor teeth.
c) The torque waveforms in Figure 20a and Figure 21b of the paper are plotted in one axis. The torque ripple is reduced by 42% at 1500rpm, as shown in Figure 20b of the figure. The torque is measured on the shaft, as shown in the Fig. 3. And indeed, this measurement at this position is not accurate, because the torque is smoothed by the rotor inertia. The torque sensor position is updated to the Figure 16a in the paper.
 
Furthermore, I did one more experiment test this week to estimate the torque waveform on the airgap, based on the method from Fig 16 at https://doi.org/10.3390/machines9120348 . The phase current with torque load and speed was measured again. The previous current waveforms in Figure 18 at this paper is with zero load. It is updated with the new test results with load. The current waveforms for the reference SRM are added as Figure 19. The current waveforms were imported to the Simulink model of the SRM system. In this way, the SRM with the measured current waveforms and 1500rpm speed is simulated. The torque waveforms on the airgap were estimated and reconstructed.
From this estimation and reconstruction results, the torque ripple reduction is 14.86%.
The detail of this estimation method is explained in the next answer.
Point 5): The authors write: “I think the difference between the Fig. 14 and 18 is because the simulation for Fig. 14 doesn’t have electromagnetic noise, but the 18 has numbers of electromagnetic noise from the real world test.”
The reviewer believes that the motor torque ripple cannot be measured by the applied method, since when the shaft rotates at high speed, the instantaneous torque of the sensor does not correspond to the motor torque, since the torque ripple at the sensor are smoothed out by the elasticity of the mechanical transmission. The torque ripple of the motor itself can only be measured at a rotational speed close to zero, using a test bench without a long mechanical transmission.
For SRM, this measurement is difficult to perform due to the inherently oscillatory operation of the motor. But this issue should at least be discussed in the article how this was done, for example, in doi.org/10.3311/PPee.14012 .
At least, it is possible to compare the static characteristics of the motor torque with the experiment, as shown in Fig. 10 at doi.org/10.3311/PPee.14012.
Answer:
a) Yes, indeed, when I measure the torque on the shaft of the SRM, the torque ripple is smoothed by the elasticity of the mechanical transmission. So, I changed the sentences in the paper on line 301 to 307 to make this point more clear. And, I tested the SRM again to get the current waveforms with speed and load, and used the method from the paper https://doi.org/10.3390/machines9120348 Fig. 16 to estimate and reconstruct the torque again. 
Both the proposed SRM and reference SRM are estimated with Simulink model. The results are added to the Figure 21 in the paper. From the estimation and reconstruction results, the torque ripple is reduced by 14.86%.
The detail is added to the paper from line 309 to 320.
b) The torque sensor signal is analogy signal. The analogy signal is converted to digital by analogy-digital-converter in high frequency. So, the torque ripple can be measured when the motor has some speed. On the other hand, the measured torque ripple is smoothed by the rotor.
c) The static characteristics of the motor were tested and presented in the Figure 17 in the paper. The agreement between the measured data and FEM data is very good.
Point 6): The authors write: “The sentences about the dimension and power comparison are added from line 282 to 284. The efficiency and cost analysed sentence is added to the future works from line 302 to 303. This proposed SRM is a 2.5kW small scaled SRM from a 80kW full scaled SRM. 
The full scaled SRM is referenced from a permanent magnet motor. The permanent magnet motor is used on an EV from the market. For the full scaled 80kW SRM, the efficiency is analysed. It is compared to a three-phase SRM. Compared to the three-phase SRM, the efficiency is 2% lower at the high efficiency area from the simulation results. The following figure shows the efficiency comparison”.
Please note that the question was not about comparing two SRM designs, but about comparing an optimized SRM with a state-of-art interior permanent magnet motor used in the target application (target electric vehicle).
Answer: 
The sentences about interior permanent magnet motor are updated to the future works from line 341 to 343.

Author Response File: Author Response.docx

Round 3

Reviewer 1 Report

 

The answers to the reviewer do not seem to appear in the text. The first paragraph of section 6 still includes uninteresting details about the DC load machine, but not the static test. Fig 2 does not show a “setup” but a “specification” graph. At face value this sounds like a desired design curve and the measured results would be compared to this.

Fig 3 still seems to show the same information twice, although perhaps this is still a traack changes effect.

Fig 5 does not seem to show anything about copper length and fig4b doesn't show this either. The explanation, which again is not in the paper text, is even more confusing. The original machine has slots which are only half full, but the modified design has full slots with longer end turns. This requires only 16% more copper? If you had the two wound (and unwound) stators you could simply have weighed them to find directly the comparative quantities of copper.

Fig 6 does not show any photos of windings. Do I have the right version 3?

The notes to the reviewer appear to agree with my comments on long end windings and efficiency, but again this does not seem to be in the revised paper text. A useful paper should be honest and give full disclosure.

Author Response

Responses to the reviewer 1

Thank you for the reviewer to check my paper in the detail.
Point 1) The answers to the reviewer do not seem to appear in the text. The first paragraph of section 6 still includes uninteresting details about the DC load machine, but not the static test.
Answer:
The sentence about the DC load resistor bank in the first paragraph of Section 6 is deleted.
The sentence about DC machine is kept. Because I think it’s better to present what the load machine is for the readers.
The static test results are presented in the Figure 17 in the paper. I think the agreement between the measured data and the FEM data is very good.
Point 2) Fig 2 does not show a “setup” but a “specification” graph. At face value this sounds like a desired design curve and the measured results would be compared to this.
Answer:
Dear reviewer, I think there is a misunderstanding about my responses for the second round revision. When I upload the responses on the website, I need to submit two contents: one is the pure text version of my responses, it can only be filled to the input box on the website; the other one is the .doc file version of my responses, it is uploaded to the website.
In the .doc file, I added some figures to answer the comments. But I’m not able to add it to the filling box.
I think if you’re able to and if you download the .doc file from the website, you can see the Figures. The ‘Fig.2’ means the Figure 2 in the previous responses .doc file. It’s my mistake that I didn’t mention this in my responses text before.
The file is at this position:
 
The measured results are added to the “specification” figure, the figure is moved to Section 6 as Figure 18. The explanation is added from line 289 to 290. The added sentence is that “In the Figure 18, the tested points for this paper are marked.”
Point 3) Fig 3 still seems to show the same information twice, although perhaps this is still a track changes effect.
Answer:
I think yes, if you switch the ‘Display for Review’ bottom at the ‘Tracking’ tab from ‘All Markup’ to ‘Simple Markup’, the repeating figures will disappear I think.
Point 4) Fig 5 does not seem to show anything about copper length and fig4b doesn't show this either. The explanation, which again is not in the paper text, is even more confusing. The original machine has slots which are only half full, but the modified design has full slots with longer end turns. This requires only 16% more copper? If you had the two wound (and unwound) stators you could simply have weighed them to find directly the comparative quantities of copper.
Answer:
‘Fig 5’ means the Figure 5 in the .doc file of the previous responses.
I added the windings of the two SRMs to the Figure 2 of the paper.
The 16% more copper is only for each phase. The reference SRM has three-phase and the proposed SRM has six-phase. There is a big difference of copper length for the two SRMs. And this is the main disadvantage of the proposed SRM. I updated it to the paper from line 119 to 124.
Because the conventional SRM is a commercial one, which is bought by me. The proposed SRM is designed by me in detail. But it is manufactured by a company. When I receive the proposed SRM, it’s an assembly motor like the Figure 16a. I’m not able to disassemble it to remove the stator from the case. The disassembly part in Figure 2a was sent to me by the company. So I’m not able to weigh the stators.
Point 5) Fig 6 does not show any photos of windings. Do I have the right version 3?
Answer:
‘Fig 6’ means the Figure 6 in the .doc file of the previous responses.
I added the windings of the two SRMs to the Figure 2 of the paper.
The explanation is added from the line 119 to 124.
Point 6) The notes to the reviewer appear to agree with my comments on long end windings and efficiency, but again this does not seem to be in the revised paper text. A useful paper should be honest and give full disclosure.
Answer:
The sentences are added to the paper from line 119 to 124.
It is also added to the conclusion from line 351 to 353.

 

Please see the attachment.

Author Response File: Author Response.docx

Reviewer 2 Report

I think the authors have responded satisfactory to the reviewer comments. 

Author Response

Responses to the reviewer 2

Thanks a lot for the reviewer’s work to help improving my paper.

Note:
According to one comment from the reviewer 1 for the third round revision, the Figure 2 in the paper is updated and moved to the position of Figure 18. The explanation paragraph is also moved to near the Figure 18. One sentence is added to explain the updated part: “In the Figure 18, the tested points for this paper are marked.”
The following figure shows the update of the figure.
      
Before update    After update

 

Please see the attachment.

Author Response File: Author Response.docx

Round 4

Reviewer 1 Report

You now show the new machine in fig 2 and say that it uses more copper. The statement that 16% more copper per phase is used is still misleading. If we have six vs three phases then that is twice as much copper plus 16%. I would expect to see a comparison of something like average rated torque to show that the slots, which you say have more total copper, do give more output in some way. Your main results are about ripple reduction but the title is more general. Optimization of a machine sounds to me like better efficiency etc.

Author Response

Point 1) You now show the new machine in fig 2 and say that it uses more copper. The statement that 16% more copper per phase is used is still misleading. If we have six vs three phases then that is twice as much copper plus 16%. I would expect to see a comparison of something like average rated torque to show that the slots, which you say have more total copper, do give more output in some way. 
Answer: Dear reviewer, thanks for the review. I prepared the measured inductance and maximum static torque comparison. Very sorry that I don’t have enough time to prepare the rated torque data.
Because of the more copper, the proposed SRM has higher inductance. The following figures show the inductance comparison. The following figure is added to Figure 3 of the paper. The explanation is added from line 132 to 135.
      
In the Table 1, I moved the last two lines to the position above the line with “Copper winding length per phase (m)”. I think it makes the comparison of the two SRMs more clear.
The moved lines are like the following table:
Number of phases    6    3
Number of phases working simultaneously    2    1
Copper winding length per phase (m)    37.12    32
The static torque with maximum current is compared with FEM simulation results. The simulated currents for both SRMs are 35A, which are the maximum currents. For the proposed SRM, the average static torque is 23.92Nm. For the conventional SRM, the average static torque is 6.60Nm. This difference is mainly due to the inductance increase. They are displayed in the following figures.
The figures are added to Figure 10 and Figure 11 in the paper. The text is added from line 227 to 230.
It is also added to the last line of the Table 1.
Maximum average static torque (Nm)    23.9    6.6
When I design to reduce the torque ripple, my first consideration of the constrain is to keep the motor volume, slot area and slot filling factor same for the two SRMs. But, the design results show that, the proposed SRM obtains higher static average torque. It’s more than the original target to reduce the torque ripple.
 

Point 2) Your main results are about ripple reduction but the title is more general. Optimization of a machine sounds to me like better efficiency etc.
Answer: Dear reviewer, for my opinion, I think it’s better to keep the title. When I design the SRM, the design constrains for the two SRMs include the same outer diameter, same stack length, same slot area, same slot filling factor, same airgap length and same material. The achievements of the design include the torque ripple reduction, and the average static torque increasing. Besides, the control methods are adjusted for two SRMs. So, I think it’s OK to say that it is the design optimization of the SRM.

 

Please see the attachment.

Author Response File: Author Response.docx