Dynamic Response in Multiphase Electric Drives: Control Performance and Influencing Factors

Round 1

Reviewer 1 Report

Basically, the idea of the paper is interesting 

 -->« for the first time a comparative analysis of the dynamic performance of conventional IRFOC and FCS-MPC in multiphase drives »

-->« Its main objective is to provide consistent conclusions about speed, current and voltage behavior in transients scenarios. »

Currently the paper  is not completely  convincing.

 The performances  of  the 3 different  « control structure » « 2 conventional(S) IRFOC » and « FCS-MPC »  are depending of course of the structure but also on the tuning of the parameters of the controllers of the structure. As consequence, it is difficult to give trustworthy conclusions on experimental resultats without giving at least in annexes a detail of the tuning of the parameters. At example for the tuning of PI controllers the time response is a compromise with the robustness concerning torque load variations.  

 1>As consequence we consider that the transient of response for a new reference of speeds must be completed by a transient to a variation a load torque. If the FCS-MPC has also a better response in comparison with IRFOC, the following sentence in the conclusion part « Therefore without the utilisation of over-modulation techniques, FCS-MPC will be considered the best option for a transient situation in current terms. » is more acceptable. Even if it can be considered that the IRFOC can be improved, the paper gives an example where the tracking of the speed during a new speed reference  but also under a load torque variation is better for FCS-MPC.

Without this complementary work the authors must notify  that the conclusion is partial.

 Of course the comparison of the robustness of the methods  in a transient response after a modification of the parameters of the machine would  be also a criterion for a more global comparison. 

2>With the use of SiC and GaN the frequency PWM is intended to increase a lot  (more than 50 kHz) without increase of power component losses. In this case the sampling frequency could  be quite lower for the IRFOC in comparison with PWM frequency since the time constants are still the same. Can  the authors add a short analysis of the impact of this modification for the comparison between IRFOC and FCS-MPC?

 3>For a six-phase machine, performance of xy control is  depending also  of the time constant in xy plane. The table I must be completed by the cyclic inductances in  a-b and in x-y subspaces

 4>About 4.2.1 paragraph. As the neutral are not connected it should be easy to inject third harmonic voltage (or zero-sequence)  for CB-PWM. For a three-phase machine the result for dc-link voltage  will be in this case quite close to this one of SV-PWM but   perhaps it is different for an assymetrical 6-phase machine

4a> It is necessary to precise in the paper   that the chosen CB-PWM is quite the basic one.

4b>As it is not a three-phase but an asymmetrical 6-phase can the author precise if the injection of third harmonic has the same efficiency for the increase of DC-link use  as for a three-phase machine in comparison with the SV-PWM chosen by the authors?

4c>in [26] it is given 0.644 and not 2/3 for maximum amplitude. Can the authors  precise the formula of projection in alpha-beta  which gives 2/3  .

 5> The table 4 is surprising : iq slope of LFCS is lower tahn than SV-PWM ? Why both  5% and 10% settling times ?

 6>It is not sure that the results are still correct for a synchronous machine with non sinusoidal  electromotrive force. So the conclusion should add that the results have been tested for an induction machine (and could be probably extended to a sinusoidal synchronous machine).

 7>The comparison is done only on transient performances but it could be of interest in the paragraph 4.1.1 to precise if for a same average-switching frequency the commutation losses of the VSI  are also the same or different  for FCS-MPC in comparison with with SV-PWM.

 

8>For a scientific paper, the first sentence « The convulse worldwide situation caused by the Ukrainian war requires reformulating the management of the available ecological resources to ensure the viability of our planet. »  appears as too associated to the  present. The « climate change and energetic transition » is less cyclical even if there is probably a link  between war and « climate change and energetic transition ».

Author Response

Dear Reviwer,

Firstly, thank you for your efforts to enhnace the quality of our work.

Please find attached in this message in a single file the reply to reviewers and the new manuscript version.

My best regards,

Ignacio

Author Response File: Author Response.pdf

Reviewer 2 Report

In the submitted manuscript, the authors conducted a comparative analysis of the most common control schemes, namely the indirect field-oriented control (carrier-based/space-vector pulse width modulation) and the model predictive control, for a six-phase electric drive. The authors identified and described in detail the selected criteria: reaction time and response capacity. The conclusions of the authors are supported by the results of experimental tests.

It was my pleasure to read the article. The manuscript is well structured and provided with quality figures. The description of methods and results is detailed and suitable for a wide range of readers. Only some minor English changes are required with focus on stylistics, e.g. authors use phrases with the word "thanks" too often, perhaps in some places it is more appropriate to replace it with more strict phrase constructions.

The introduction could be enhanced by a discussion of multi-phase motor optimization techniques (e.g. https://doi.org/10.5370/JEET.2015.10.3.1007, https://doi.org/10.3390/wevj12010024,  https://doi.org/10.3390/en15092994), which may explain the choice of the electric drive from the experemental test bench.

Author Response

Dear Reviwer,

Firstly, thank you for your efforts to enhnace the quality of our work.

Please find attached in this message in a single file the reply to reviewers and the new manuscript version.

My best regards,

Ignacio

Author Response File: Author Response.pdf

Reviewer 3 Report

The paper compares three different current control techniques for 6 phase drives. Below are my observations:

1. How are the controller parameters obtained? It would be good to mention this.

2. There are other important metrics to use for comparing controller performance. For example harmonic distortion, disturbance rejection, machine parameter uncertainty, controller complexity or execution time, effect of cross coupling, flux ripple, torque ripple etc. In the paper the metrics used are response time (rise time, settling time), DC link utilization and sampling time. The authors can consider adding other metrics for comparison to add more value to the paper.

3. It is not clear what is dead-time.

4. The phase votlage applied in case of MPC in Fig 8 and 13 are quite different than the IRFOC controller. Does MPC apply different voltage vectors? What would be other consequences of that? D-q axis voltages would be better to show, rather than Va. D-q axis voltages are the direct output from the controller. Also the waveforms are at low speed. It is preferable to obtain current step response at rated or near rated speed. What is the rated speed, torque, pole numbers of the machine?

5. Sec 5.2.2 - how are the gains for only P controller chosen? Lower rise time for P controller means higher bandwitdth is possible for the P controller. Why is that the case? P and PI controller should have same bandwidth, if designed properly. The effect of integral pole is cancelled by the zero of the PI controller. At high frequencies, both have same gain.

6. Why slope of i_q reference is important to be quantified? That will depend on speed controller and sampling time. Higher slope of i_q reference will have no impact performance if slope of actual i_q current is same. Slope of actual i_q would be important.

7. In all the tests, you can mention the voltage limitation of each of the controller. Then it will be clear from the waveform, that the controller is limited by the maximum possible applied voltage.

8. What happens to the xy plane currents and voltages in each of the tests?

9. Overall, there does not seem to be any improvement in our understanding of the different controllers. The comparison ignores important aspects. The conclusions are already well understood.

Author Response

Dear Reviwer,

Firstly, thank you for your efforts to enhnace the quality of our work.

Please find attached in this message in a single file the reply to reviewers and the new manuscript version.

My best regards,

Ignacio

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

Answers and modifications are convenient for publication. Even if results are logically predictible and so can be obtained by simulation, the support of experimental results with adequate  tests  is a more convincing proof.

The conclusion could distinguish the differences  which are intrinsic and associated to the algorithm from those associated to  the state of art of technology (maximum switching frequency with for example wide-gap components with low losses and capacities of  processors) .  

One value is surprising in table1 with iq* slope<iq slope.

Other minors remarks  and proposal in attached file

Comments for author File: Comments.pdf

Author Response

Dear Reviewer,

Thank you so much for your efforts to enhance our work. Please note that you can find in the attached file the reply to reviewers and the new manuscript version.


My best regards,

Ignacio

Author Response File: Author Response.pdf

Reviewer 3 Report

Thank you for the detailed response to my comments. I have no further suggestions.

Author Response

Dear Reviewer,

Thank you so much for your efforts to enhance our work. Please note that you can find in the attached file the reply to reviewers and the new manuscript version.


My best regards,

Ignacio

Author Response File: Author Response.pdf