A Novel Double-Sided Electromagnetic Dog Clutch with an Integrated Synchronizer Function

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

The manuscript deals with the description, simulation, implementation and experimental verification of an electro-magnetic control system for synchronizing  a dog clutch sleeve before coupling.

The title problem is well presented, the various passages are well discussed and documented, and the solution/workaround for the inevitable issues are  pretty instructive.

A few minor points:

• p.15, l.477 the 0.7 Nm friction torque estimation seems evaluated based on static (vs. dynamic) friction, whose coefficient usually differ. Is there an experimental way to overcome this issue?
• p. 16, l.524 the apparently same friction torque grows to 4 Nm at 1000 rpm speed; is it possible to at least coarsely predict the variation of friction torque with speed, e.g. based on the expected loss variation at journal vs. roller bearings?
• p. 15-16: the "app." expression is widely employed, and somehow disturbing in reading. Could a simple tilde, e.g. "~0.43s", be employed instead?

Author Response

Dear Reviewer,

firstly, I want to thank you for your very positive feedback! Secondly, I am really impressed with the deepest understanding of our research, that you have shown! It surely requires a large amount of knowledge not only in the area of different electromagnetic energy converters and actuators, but also in the field of automotive transmissions. This allows some extended discussion to answer your questions.

 

The friction torque is one of the main problems in the proposed system. In contrast to electric machines, where it only defines the efficiency, the friction in the rotating parts of a gearbox is a crucial parameter, which defines, whether the proposed actuator can be applied or not.

 

Initially, the idea to add overlapping teeth on the complementary rings has arisen to increase the disengaging force in a common dog clutch. A little later it became clear, that these teeth open up the possibility of a contactless speed synchronization without an external torque source, if the coil currents can be properly controlled. In the very first designs, the torque amplitude was too low (less than 1 Nm), so that it required a development of an analytical flux model [1] and a design method [2] to obtain torques that allowed at least to suppose that the synchronization can be successful. Nevertheless, the torque density is still low comparing to common electric machines [2]. On the other hand, a gearbox requires a dog clutch that looks like a dog clutch, not a linear-rotary electric machine, and the proposed solution still represents a dog clutch. However, as it has been shown in the article, its average torque can become lower than the friction torque in some cases.

 

Thus, you have raised very important but also difficult questions. I have conducted a literature research to answer, whether the friction torque can be predicted with a relative simple analytical expression or not. Based on the results, I can answer that an analytical modeling of friction torques is difficult. For bearings only, a relative complex model is proposed [3]. Probably, it would require the use of AI to model the entire gearbox [4, 5], since many other sources of friction have to be modeled in the belt transmission [6]. During the experiments, the belts became significantly warmer after operation at higher speeds. Presumably, the belt transmissions represent the main sources of friction in the designed gearbox, not the bearings. The use of belt transmission instead of gear trains was a requirement of the institute manager (F. Poltschak) to make the gearbox more universal.

 

Based on the experimental results, analytical expressions for the friction torques have been derived in a limited range of the rotations speeds. A particular subsection that considers the friction torques has been added to the experimental results. Your question about different measurements of static and dynamic frictions has been also answered.  Based on the experimental results, it can be concluded that the difference between the static and dynamic friction coefficients is not very high in the described system [7].

 

P.S. The mentioned values of frictions 0.7 and 4 Nm represent different resulting frictions. The first value is the resulting friction on the countershaft, while the second is the resulting friction on the right gearwheel, which is approximately 64/44 times higher due to the gear ratio. The resulting friction on the countershaft grows in this case to ≈ 4*44/65 ≈ 2.75 Nm. Different symbols have been added for these values to avoid further confusion. Moreover, I just noticed that the value 4 is false, it must be 3. This will be fixed in the final version of the manuscript.

 

P.P.S. There were indeed to many “app.”, all of them have been replaced by ~

 

Sincerely,

Bogdan Miroschnitschenko

 

References

 

1. Miroschnitschenko, Bogdan. "Magnetic flux calculation in a novel linear-rotary electromagnetic actuator using 3d magnetic equivalent circuit." 2023 11th International Conference on Control, Mechatronics and Automation (ICCMA). IEEE, 2023.

 

2. Miroschnitschenko, Bogdan, Wolfgang Amrhein, and Florian Poltschak. "Design and Optimization of a Linear-Rotary Electromagnetic Actuator Based on Analytical Model of Magnetic Flux." 2024 International Symposium on Power Electronics, Electrical Drives, Automation and Motion (SPEEDAM). IEEE, 2024.

 

3. The SKF model for calculating the frictional moment. https://cdn.skfmediahub.skf.com/api/public/0901d1968065e9e7/pdf_preview_medium/0901d1968065e9e7_pdf_preview_medium.pdf

 

 

4. Hou, Yu, et al. "Prediction of Frictional Moment of Cylindrical Roller Bearing Using Experimental Data-Driven Artificial Neural Networks." Journal of Tribology 145.9 (2023): 091103.

 

5. Baş, Hasan, and Yunus Emre Karabacak. "Machine learning-based prediction of friction torque and friction coefficient in statically loaded radial journal bearings." Tribology international 186 (2023): 108592.

 

6. Kubas, Krzysztof, and Andrzej Harlecki. "Dynamic analysis of a belt

transmission with the GMS friction model." Meccanica 56.9 (2021): 2293-2305.

 

7. Youqin, J. "Coefficients of Friction---Static versus Dynamic." (2020).

 

Reviewer 2 Report

Comments and Suggestions for Authors

This is a very interesting study, and my questions and suggestions are as follows:
1. There are many types of clutches, such as wet clutches that can be used for power shifting. It is recommended that the authors provide a brief introduction and comparison of various clutches in the introduction section, highlighting the application scenarios of the clutches described in this study.
2. The power transmitted by the clutch during the experiment seems to be very limited. What is the maximum load torque it can withstand under the premise of smooth gear shifting? Can it be practically applied in vehicles?
3. Does the clutch need to be continuously powered during use? How much heat will it generate and how much energy will it consume?
4. It is suggested that the authors directly provide the component names corresponding to each numerical index in the title of Figure 8.
5. It is suggested that the authors explain the limitations of this study and its potential application scenarios in the conclusion section, and provide an outlook for the future.

Author Response

Dear Reviewer,

firstly, I want to thank you for your positive feedback and finding our study very interesting!

Secondly, I find that the quality of the paper has been significantly improved based on your comments. All of them were implemented in the revised manuscript. Below, I leave some remarks about the changes that have been made.

 

1. There are many types of clutches, such as wet clutches that can be used for power shifting. It is recommended that the authors provide a brief introduction and comparison of various clutches in the introduction section, highlighting the application scenarios of the clutches

• Indeed, it was not sufficiently explained in the introduction, why a dog clutch is preferred in this case. Now, other clutch types are briefly mentioned together with their typical application in the automotive area. Moreover, the advantages and disadvantages of dog clutches are now described.

 

2. The power transmitted by the clutch during the experiment seems to be very limited. What is the maximum load torque it can withstand under the premise of smooth gear shifting? Can it be practically applied in vehicles?

 

• These questions are now briefly considered in the paper. The operation of the clutch under load is described with more details in the forthcoming publication [1]. Its relevant part is attached to this message to answer your questions without a significant increase of the reviewed paper.

 

3. Does the clutch need to be continuously powered during use? How much heat will it generate and how much energy will it consume?

• That is a very good point. Yes, the clutch must be continuously powered to keep the shift sleeve engaged at higher loads. The reason, why this is necessary, is also described with more detail in the attached file. However, the power losses and heat generation are negligible. A brief explanation for that is now added in the text.

 

4. It is suggested that the authors directly provide the component names corresponding to each numerical index in the title of Figure 8.

• This is very important comment. It is much better, to see the meaning of all indexes in one place, and not to search for them in the text. All indexes are now explained directly in the caption of this Figure.

 

5. It is suggested that the authors explain the limitations of this study and its potential application scenarios in the conclusion section, and provide an outlook for the future.

 

• This point is also crucial. The study is only focused on the possibilities of speed synchronization. Many other aspects were not considered. Two different methods have been developed to allow sensorless operation of the clutch [1, 2] and will be published soon. However, there are many other interesting topics that can be studied. The limitation and the outlook for the future are now presented in the Conclusion.

 

All significant changes are marked in the revised paper.

 

Kind regards

Bogdan Miroschnitschenko

 

References

1. Miroschnitschenko, B. Sensorless Control of Linear Motion in a Linear-Rotary Reluctance Actuator Integrated into an Electromagnetic Dog Clutch. in publishing.
2. Miroschnitschenko, B.; Poltschak, F.; Amrhein, W. Sensorless estimation of relative angular position and speed in a linear-rotary actuator with high influence of eddy currents. In Proceedings of the 2025th International Conference on Electrical Machines, Drives and Power Systems (ELMA2025). IEEE, in publishing, 2025.

 

 

Author Response File: Author Response.pdf

Reviewer 3 Report

Comments and Suggestions for Authors

The manuscript describes a double-sided electromagnetic dog clutch to be used in gearboxes for automotive drives. The presented device is an evolution of a previously developed single-sided one. 

The topic is interesting and a few simulation results are shown, together with test results obtained on a prototype.

However, some changes and improvements are needed, as detailed in the following:

• fig. 3 shows torque and axial force diagrams as a function of the angular position, with current as a parameter: it should be explicitly declared wich calculation tool have been used (FEM 3D magnetostatic simulations? which software?);
• in the equations and in the text, the symbol used for the applied voltage (U) is too similar to the symbols used for the gear ratios (u_l and u_s): please, change one of the two symbols, in order to avoid misunderstanding;
• the control algorithm shown in fig. 5 is not sufficiently explained in the text: please, add explanations and comments;
• eq.s (11) and (12) are not clear and should be more widely explained, also referring to fig. 5;
• the simulated waveforms shown in fig.s 6 and 7 are not sufficiently explained and commented; moreover, the simulation environment should be cited (Matlab/Simulink?);
• in Paragraph 5, the experimental setup is described, with information as concerns several used components; however, no details are given concerning used measurement instrumentation and transducers : please, add their main accuracy data;
• rows 507, 508: it is written: " the speed synchronization takes app. 0.6;"; maybe, 0.6 should be seconds, so s is missing.

Author Response

Dear Reviewer,

firstly, I want to thank you for your positive feedback and finding our research interesting!

Secondly, I find that the quality of the paper has been significantly improved based on your comments. All of them were implemented in the revised manuscript. Below, I leave some remarks about the changes that have been made.

 

• “fig. 3 shows torque and axial force diagrams as a function of the angular position, with current as a parameter: it should be explicitly declared wich calculation tool have been used (FEM 3D magnetostatic simulations? which software?);”
• this is indeed important; now, it is mentioned that the characteristics were obtained using a 3D magnetostatic analysis in Ansys Electronics.

 

• in the equations and in the text, the symbol used for the applied voltage (U) is too similar to the symbols used for the gear ratios (u_land u_s): please, change one of the two symbols, in order to avoid misunderstanding;
• It is true, that these symbols can be confusing for the reader; now, symbols g_l and g_r are used for the gear ratios.

 

• the control algorithm shown in fig. 5 is not sufficiently explained in the text: please, add explanations and comments;
• The development of the algorithm has been described in detail, but not summarized to allow a fast understanding of Fig. 5. Now, it is summarized in a particular paragraph.

 

• s (11) and (12) are not clear and should be more widely explained, also referring to fig. 5;
• The torque components used in these equations and how they arise are now better explained in the text.

 

• the simulated waveforms shown in fig.s 6 and 7 are not sufficiently explained and commented; moreover, the simulation environment should be cited (Matlab/Simulink?);
• Indeed, the complex waveforms have been introduced for the first time in these Figures without any explanation. Now, they are explained in the text.

 

• in Paragraph 5, the experimental setup is described, with information as concerns several used components; however, no details are given concerning used measurement instrumentation and transducers : please, add their main accuracy data;
• The accuracy data have been added for all sensors to show that the measurement have been done with a sufficient accuracy; moreover, some more details about the used electronics have been added.

 

• rows 507, 508: it is written: " the speed synchronization takes app. 0.6;"; maybe, 0.6 should be seconds, so s is missing.
• Thank you, the missing “s” is now on its place.

All significant changes are marked in the revised paper.

 

Kind regards

Bogdan Miroschnitschenko

Round 2

Reviewer 3 Report

Comments and Suggestions for Authors

The reviewer suggestions and requests have been adequately fulfilled. The new/modified parts are significant and helpful for the reader. No more requests or comments.