Precise Control of Following Motion Under Perturbed Gap Flow Field
Round 1
Reviewer 1 Report
Comments and Suggestions for Authors
Overall Comments
This paper is on the topic of controller design for position control of something like a hydrodynamic bearing where there is an influential force from the fluid film which must be considered in the controller design. The method uses PID control augmented with a canceling disturbance observer and some other filters.
Experimental results are presented; however the paper is devoid of any experimental explanation, even to the point of not introducing what type of fluid film system is being considered. The deficiencies go far beyond simply leaving out experimental details.
Specific Comments
- Sensor noise is an important issue when practically implementing submicron control.
- Line 38 – Why does A need to follow B in vertical direction?
- Figure 2 – Where did this figure come from? What load, speed, geometry? Is this position or force? What sensor was used? At what location?
- Figure 4 – Show how the modal frequency ranges correlate to results in Figure 3.
- Figure 6 – Legend entries are mislabeled.
- Figure 7 – Are all these blocks SISO? What if the disturbance is not at the same location as the control actuation? Also, what is the actuator?
- Figure 8 – Make more consistent with Figure 7, e.g., show controller. It is confusing having the control effort look like an external disturbance. Also, disturbance canceling signal is labeled as disturbance. Don’t we care about the position output?
- Figure 11 – Pick off points and summing nodes look identical.
- Please give I/O pairs for all Bode plots. Also, I prefer real units to dB.
- Figure 15 – Figure looks fuzzy. Also, what is the order of the closed-loop system?
- Figure 16 – Vertical axis for snap mislabeled.
- Figure 17 – What is the disturbance? Also, show time scale to observe response to changing setpoint, i.e., 1.1 s, in addition to response from disturbance.
- Show photograph of test rig and concept sketch with pertinent details.
Author Response
Comments 1: Sensor noise is an important issue when practically implementing submicron control.
Response 1: Thank you for pointing this out. We agree with this comment. We have designed hardware filters and software filters to suppress the influence of sensor noise. We did not mention this and the correlated system delay, considering that the paper’s focus is on control techniques under wideband time-varying disturbances. We have modified accordingly in line 174, page 5, mentioning that the LPF is used to suppress sensor noise.
“[A low-pass filter was used to filter the PID controller output to suppress the influence of sensor noise.]”
Comments 2: Line 38 – Why does A need to follow B in vertical direction?
Response 2: Thank you for the question. Because in the usage scenario, B needs to perform a 6-degree-of-freedom motion. To maintain the parallel gap, A needs to follow B in the vertical direction. We now briefly mention this in line 33, page 1.
“[Part A needs to follow B in vertical direction to maintain the parallel gap.]”
Comments 3: Figure 2 – Where did this figure come from? What load, speed, geometry? Is this position or force? What sensor was used? At what location?
Response 3: We used a force measuring tool to measure disturbance force of the gap flow field between part A and the flat plate of the tool, as shown in the following figure, which is presented in Figure 1 (b) of the paper.
During measurement, component A is stationary while the flow field is flowing. In the motion scenario with components A and B both moving as presented in Figure 1 (a) of the paper, the disturbance force is more complex and difficult to directly measure. We have designed DOB to evaluate it. We have modified the content in line 99, page 3.
“[As a preliminary step, we first used a force measuring tool to measure disturbance force of the gap flow field between part A and the flat plate of the tool, as shown in Figure 1 (b). During measurement, component A was stationary while the flow field was flowing. The measurement band of sensors were 1.2 kHz, and the sampling frequency was 10 kHz.]”
Comments 4: Figure 4 – Show how the modal frequency ranges correlate to results in Figure 3.
Response 4: Thank you for pointing this out. As shown in Figure 3 (originally Figure 6), the first four orders of the modal frequency ranges are 7.14Hz, 297Hz, 434Hz, and 867Hz. Due to the small amplitude of high-order modes, this article only deals with the first three orders, as shown in the old Figure 3. There are some errors in the old Figure 4. Considering the opinion of another reviewer, we have removed the Figures originally 3 and 4 and only retained the textual description. Meanwhile, Figure 6 is retained.
Comments 5: Figure 6 – Legend entries are mislabeled.
Response 5: Thank you for this comment. We have fixed Figure 6 (now Figure 3).
Comments 6: Figure 7 – Are all these blocks SISO? What if the disturbance is not at the same location as the control actuation? Also, what is the actuator?
Response 6: The motion control of both component A and component B is multi-input and multi-output (MIMO). To study motion control under wideband time-varying disturbances, we decouple MIMO into a number of single-input and single-output (SISO). This article describes the controller design method for SISO in Z direction. In fact, Due to the disturbance force acting in the Z direction and not located at the control point, disturbance torque in Rx and Ry directions will be generated. The MIMO control under disturbance torque is our next research topic. As a common handling method, in order to emphasize the description of DOBC, the actuator and sensor components have been omitted in Figure 7. We have modified the content in page 3, line 94 to explain this.
“[The motion control of both component A and component B is multi-input and multi-output (MIMO). To study motion control under wideband time-varying disturbances, we decoupled MIMO into a series of single inputs and single outputs (SISO). This article describes the controller design method for SISO in the Z direction of part A.]”
Comments 7: Figure 8 – Make more consistent with Figure 7, e.g., show controller. It is confusing having the control effort look like an external disturbance. Also, disturbance canceling signal is labeled as disturbance. Don’t we care about the position output?
Response 7: To better illustrate the construction process of the H∞ model, we have modified the original Figure 8 into Figures 5(a) and 5(b). Figure 5(a) explains how to construct an extended DOB loop from Figure 7. Due to the design concept of DOB, it is possible to decouple the design of external PID loop and internal DOB loop. So when designing the internal DOB loop, we set the input uc of the external PID loop to zero. Then, the output of Q-filter is the disturbance force evaluation value , and ue equals to d- . In H∞ model, the position output is the output of the Plant and the input of , which is considered an internal variable. It should be pointed out that there is an error in the old Figure 8. Z2 (t) is ud, not ue. In the new Figure 5(b), We have corrected the error.
Comments 8: Figure 11 – Pick off points and summing nodes look identical.
Response 8: In fact, we designed PFs in parallel mode. There is a summer here, not a linking point. We have modified the Figure 8 (originally Figure 11).
Comments 9: Please give I/O pairs for all Bode plots. Also, I prefer real units to dB.
Response 9: Thank you for pointing this out. We agree with this comment. We have modified all Bode plots.
Comments 10: Figure 15 – Figure looks fuzzy. Also, what is the order of the closed-loop system?
Response 10: We have modified Figure 15 (now Figure 13). The order of the closed-loop system is 84.
Comments 11: Figure 16 – Vertical axis for snap mislabeled.
Response 11: Thank you for pointing this out. The vertical axis for snap is shown as djerk. To keep the same, we have modified all "snap" to "djerk" in this article.
Comments 12: Figure 17 – What is the disturbance? Also, show time scale to observe response to changing setpoint, i.e., 1.1 s, in addition to response from disturbance.
Response 12: We have modified Figure 17 (now Figure 16) accordingly. We added setpoints and disturbance force evaluation values in the figure. Due to the inability to measure the disturbance force in the following motion scenario, we have provided an evaluation value of DOB.
Comments 13: Show photograph of test rig and concept sketch with pertinent details
Response 13: We have added the photograph of test rig and related description on Page 10.
Reviewer 2 Report
Comments and Suggestions for Authors
This is an interesting paper. However, there are some issues in this paper as follows:
- The abstract section could have more numerical results.
- I can not see the conclusion section. This section must be.
- What are the limitations of your study? Please mention them before the conclusion section.
- It would be better to add a table to compare your method and the traditional method in the literature review section.
- Please add a focused paragraph in the introduction highlighting exactly what aspect of this integration is new or significantly improved over previous works.
- It would be beneficial to include a subsection or appendix that describes the experimental or simulation setup, sampling time, real-time controller platform (if applicable), and hardware-in-loop validation.
Author Response
Comments 1: The abstract section could have more numerical results.
Response 1: Thank you for this comment. We have modified the content of the abstract section.
“[Compared to traditional control methods, the new method can increase the open-loop gain by 15 times and the open-loop bandwidth by 8%. We even observed a 150-time increase of the open-loop gain at the peak frequency.]”
Comments 2: I can not see the conclusion section. This section must be.
Response 2: Thank you for pointing this out. We have added the conclusion section.
Comments 3: What are the limitations of your study? Please mention them before the conclusion section.
Response 3: Thank you for the comment. We have modified accordingly in line 384, page 14.
“[This study has a few limitations. First, this study only discussed SISO systems, which are different from the MIMO systems often used in real semiconductor device. Second, external disturbance forces often do not act on the control point, which will result in disturbance torque. This is not fully discussed in this paper. In the future, more research should be conducted to optimize the control parameters and improve performance, as well as to investigate MIMO control under torque disturbances of the gap flow field.]”
Comments 4: It would be better to add a table to compare your method and the traditional method in the literature review section.
Response 4: Thank you for pointing this out. We agree with this comment. We have added Table 1 and modified accordingly in line 90, page 2.
“[Table 1 compares this H∞ DOBC method with traditional methods.]”
Comments 5: Please add a focused paragraph in the introduction highlighting exactly what aspect of this integration is new or significantly improved over previous works.
Response 5: Thank you for the suggestion. We have added the following in line 77, page 2, in response.
“[This method contributes a few innovations.
- Decoupling the Disturbance Observer (DOB) loop from the outer control loop, the method constructs the extended DOB loop with zero output from the outer control loop. Then, it uses H∞ mixed sensitivity shaping technology to design a Q-filter. The H∞ DOBC is used for the overall suppression of wideband time-varying disturbances, significantly increasing the open-loop gain in the low-frequency range and enhancing the phase margin.
- The method combines DOBC with lead lag correction, which improves the open-loop gain in the low to mid frequency range, compensating for the open-loop gain loss in the mid frequency range introduced by the H∞ DOBC.
Disturbance rejection filtering correction was used to suppress certain concentrated frequency bands in the flow field disturbances and improve the open-loop gain.]”
Comments 6: It would be beneficial to include a subsection or appendix that describes the experimental or simulation setup, sampling time, real-time controller platform (if applicable), and hardware-in-loop validation.
Response 6: Thank you for the suggestion. We have added the photograph of test rig and related description in page 10.
Reviewer 3 Report
Comments and Suggestions for Authors
The research is interesting and based on combining of many advanced control system approaches. It solves a practical problem of great importance.
My comments, questions and recommendations are the following.
1.Please explain how you determine precisely the so many parameters.
2. The Introduction contains unnecessary figures not related to any of the cited literature. Only the explanations to them may be preserved.
3. The figures should contain short captions and remain self-contained. All explanations in the current captions should be moved to the main text.
4. It is good to define the initial data necessary for the design.
5. There is no explanation on plant model verification and validation based on experimetal data.
6. The bottoms of Figures 7, 8 and 11 are cut , Figure 15 is not clear
7. Some abbreviations (MCK, DOBC) have to be explained when met for the first time in the main text and separately in the abstract, as each is self-contained.
8. Only present tenses are used in papers. Please correct.
9. How do you tackle the inverse plant transfer function which describes no real-world element?
10. The stability of the PID control system at presence of all filters and corrections (Fig.11) is not tested in case of inaccurate or changed parameters.
11. When all filters are designed in the frequency domain, the PID controller should also be tuned in the frequecy domain with respect to some equivalent plant that contains all filters and corrections. Please give details on the PID tuning.
Author Response
Comments 1: Please explain how you determine precisely the so many parameters.
Response 1: Thank you for pointing this out. In fact, the parameters of LPF、W1、W2、LL are designed. The setting of the LPF filter parameters should take into account the frequency band of wideband time-varying disturbances, measurement noise frequency band, and frequency band of the reference input. 1/W1 is designed as a low-pass filter, and its parameters should take into account the frequency band of wideband time-varying disturbances. 1/W2 is designed as a high-pass filter, and its parameters should take into account the frequency band of wideband time-varying disturbances too. LL is the lead lag correction loop, and its parameters should take into account the frequency band of open-loop transfer. The parameters of Notchs are determined based on mechanical transfer function test results. The parameters of PFs were determined based on disturbance force FFT result. The parameters of PID were determined based on open-loop bandwidth and empirical formulas. We have modified those contents accordingly.
Comments 2: The Introduction contains unnecessary figures not related to any of the cited literature. Only the explanations to them may be preserved.
Response 2: Thank you for pointing this out. We apologize for our negligence. We have modified the content of Introduction and deleted unnecessary figures.
Comments 3: The figures should contain short captions and remain self-contained. All explanations in the current captions should be moved to the main text.
Response 3: Thank you for this comment. We have modified all figure captions accordingly.
Comments 4: It is good to define the initial data necessary for the design.
Response 4: Thank you for pointing this out. We agree with this comment. We have modified the manuscript. External disturbance and plant model are now described in chapter 2.1 and 2.2.
Comments 5: There is no explanation on plant model verification and validation based on experimental data.
Response 5: Thank you for pointing this out. We agree with this comment. We have modified the content in line 123, page 4.
“[The fitting results were consistent with the modal testing results of the hammering method. Part A had structural modal frequencies of 297 Hz and 434 Hz and flexible connection modal frequency of 7.14 Hz.]”
Comments 6: The bottoms of Figures 7, 8 and 11 are cut , Figure 15 is not clear.
Response 6: Sorry about our negligence in these places. We have fixed the issues in Figure 4 (originally Figure 7), Figure 5 (originally Figure 8) , Figure 8 (originally Figure 11), and Figure 13 (originally Figure 15).
Comments 7: Some abbreviations (MCK, DOBC) have to be explained when met for the first time in the main text and separately in the abstract, as each is self-contained.
Response 7: Thank you for pointing this out. We agree with this comment. We have modified those abbreviations.
Comments 8: Only present tenses are used in papers. Please correct.
Response 8: Thank you for the comment. We apologize for our negligence about tenses, and have changed the theoretical parts of our paper into present tense. However, both our knowledge and a search through the literature, especially studies involving authors from or journals in English-speaking countries (for instance, [11], [21] and [22] in our References), indicate that a mixture of tenses is necessary for a paper that is coherent and easy to understand. In particular, past tense is usually used to describe the processes of experiments conducted. We have gone through another round of language revision with an English language advisor, doing our best to make sure that the tenses are appropriate.
Comments 9: How do you tackle the inverse plant transfer function which describes no real-world element?
Response 9: Thank you for pointing this out. In fact, the inverse plant needs to be used in combination with Q-filter, and the order of Q-filter needs to be higher than that of the inverse plant. We have added an explanation for this in line 259, page 8.
“[Meanwhile, the inverse plant needs to be used in combination with Q-filter, with the order of Q-filter higher than that of the inverse plant. In this paper, the order of the inverse plant is two, and the order of the Q-filter is four.]”
Comments 10: The stability of the PID control system at presence of all filters and corrections (Fig.11) is not tested in case of inaccurate or changed parameters.
Response 10: We have added stability assessment when there was a 10% uncertainty or drift in system parameters. We have modified the contents in line 334, page 12.
Comments 11: When all filters are designed in the frequency domain, the PID controller should also be tuned in the frequency domain with respect to some equivalent plant that contains all filters and corrections. Please give details on the PID tuning.
Response 11: Thank you for the comment. We have added Table 2 to compare the characteristic values of the four control methods. The characteristic values include Open-loop gain (@20 Hz), Open-loop bandwidth, and Phase margin. After designing the open-loop bandwidth, we calculated the initial PID parameters based on empirical formulas. Then, based on the transfer function test results on the experimental platform, we fine-tuned the PID parameters and notch filters parameters. This article focuses on the design of the controller, so it does not provide a detailed description of the parameter tuning process.
Round 2
Reviewer 1 Report
Comments and Suggestions for Authors
The paper is much better. The authors did an excellent job addressing the reviewer comments.
Reviewer 2 Report
Comments and Suggestions for Authors
The authors readdressed all issues.
