Semi-Active Vibration Control Based on a Smart Exciter with an Optimized Electrical Shunt Circuit
Round 1
Reviewer 1 Report
The manuscript above investigates the effectiveness of smart exciter along with resistive inductive in structural vibration control/reduction. The authors simulated the vibration reduction through numerical and experimental simulations. The results and effectiveness of the system are presented and evaluated.
The presented study is an interesting topic in engineering field and is of merit for further investigation with more extensive and various environmental situation.
However, the paper is well organized and can be useful for other researchers. I recommend the authors to add more detail in the conclusion section, presenting numbers/percent of vibration reduction comparing systems with and without the proposed technology before publishing the paper.
Author Response
Reviewer 1:
Comment:
The presented study is an interesting topic in engineering field and is of merit for further investigation with more extensive and various environmental situation.
Answer:
The authors are thankful for the revision by Reviewer 1, which will improve the quality of this work. Below you find our answers on your comments, parts of the manuscript (marked grey) and our corrections/additions (marked blue).
Comment:
I recommend the authors to add more detail in the conclusion section, presenting numbers/percent of vibration reduction comparing systems with and without the proposed technology before publishing the paper.
Answer:
Please find the results summarized in the suggested way.
The main findings according to the feasibility of the proposed semi-active damping approach can be summarized as follows. The results of the numerical investigations prove that the time span of the step response of a panel with SE and RL shunt is reduced by a factor 10 compared to a panel without semi-active damping device. Furthermore it has been found that it is possible to reduce the steady state response of the panel (excited in its first natural mode) down to 32.4dB, if the proposed semi-active control approach is taken into account during time-harmonic simulations. The experimental results prove that the proposed approach is capable to reduce the overall vibration level in the frequency band 0Hz – 400Hz down to nearly 8.5dB, if the SE is placed in the centre of the panel. Under free-field conditions a reduction of sound pressure level close to 2dB has been detected at the position of a microphone that has been placed in 1m distance to the panel.
Author Response File: 
Author Response.pdf
Reviewer 2 Report
The article submitted to Applied sciences is devoted to current topic. Noise/ vibration reduction is important and wide research field.
The paper is divided into five chapters including introduction and conclusions. One appendix is added but
The methodology is observed. Theoretical part in Section 2 is followed by Section 3 with numerical modelling. The obtained results in Section 3 are followingly supported by experimental analysis. Thank to that, clear conclusions are mentioned.
I am not sure about the necessity of the appendix.
I recommend omitting abbreviations in keywords. And then it is better to use the full name/ nomenclature and the appropriate abbreviation in Introduction section. It relates to the line 34 SE and the line 40 – MIMO-system
Line 47 – There is a reference to Fig. 1 (c) and this is the first reference (also before the references to (a) and (b) parts – Because of that, I recommend to change order of the parts of Fig. 1 and after that Relaxation-Isolator will be part (a) and two other small pics will be (b) and (c).
Lines 90 up to 95 – Equations (1) to (3) probably should be centred (the same for the equation (4) and (5)
Line 106 – Additional info about the panel could be added (mainly material parameters as weight or Young’s modulus in order to repeatability)
Line 110 – Rayleigh damping can be described in more detail or at least referenced.
Line 145 – yellow formatting probably by mistake.
Line 172 – please consider zoom of 0.1s which means amplitude reduction in another figure.
Figure 6 – Is it necessary to publish first second, because there are zeros in that time interval. Please consider use of mm instead of m for Y-axis
As you stated that FEM was used, therefore I recommend to supplement Fig. 7 by publishing pictures of four mode-shapes with/wo SE.
Line 199 Why have not used 142.2 Hz instead of 143 Hz for angular frequency calculation. Is not a reason for not exact resonance graph in Fig. 8. The following note is related to this.
Line 207 – You are comparing 4.5 10-3 and 0.6 10-4 but 0.6 value is at the end of the time interval or we can say steady value but 4.5 is at the beginning.
Line 224 – Why has been used the panel with different dimensions as that is described in Section 2.
Line 241 – please describe setting of DAQ system of B&K type 3050, e.g. sampling rate at least.
Figure 14 – The position 3 is depicted on Fig. 14 (b). I recommend using picture from the stage where SE was attached to P2. Then, the results from this position continue in Fig. 15 and Fig. 16.
References – They are mistakenly pasted twice. I recommend at least 5 more references.
Author Response
Reviewer 2:
Comment:
The methodology is observed. Theoretical part in Section 2 is followed by Section 3 with numerical modelling. The obtained results in Section 3 are followingly supported by experimental analysis. Thank to that, clear conclusions are mentioned.
Answer:
The authors are thankful for the revision by Reviewer 2, which will improve the quality of this work. Below you find our answers on your comments, parts of the manuscript (marked grey) and our corrections/additions (marked yellow).
Comment:
I am not sure about the necessity of the appendix.
Answer:
The appendix is given to support both repeatability (Appendix A) and understanding (Appendix B). The latter is helpful to illustrate the dynamic behaviour of the panel. However, it is also possible to follow the main text without this extra information. For this two reasons we decided to work with an appendix.
Comment:
I recommend omitting abbreviations in keywords. And then it is better to use the full name/ nomenclature and the appropriate abbreviation in Introduction section. It relates to the line 34 SE and the line 40 – MIMO-system
Answer:
Many thanks for this recommendation. We followed your advice.
Comment:
Line 47 – There is a reference to Fig. 1 (c) and this is the first reference (also before the references to (a) and (b) parts – Because of that, I recommend to change order of the parts of Fig. 1 and after that Relaxation-Isolator will be part (a) and two other small pics will be (b) and (c).
Answer:
Many thanks for this recommendation. We followed your advice.
Comment:
Lines 90 up to 95 – Equations (1) to (3) probably should be centred (the same for the equation (4) and (5)
Answer:
Many thanks for this recommendation. We followed your advice. All equations are centred now.
Comment:
Line 106 – Additional info about the panel could be added (mainly material parameters as weight or Young’s modulus in order to repeatability)
Answer:
Many thanks for this recommendation. We followed your advice. Materia parameter have been added as follows.
The panel material has been characterized by a density of 137.14kg/m³, a Young’s modulus of 6.5e8Pa and a Poisson’s ratio of 0.35. In addition, stiffness proportional Rayleigh damping, theory and new developments have been reported in [22], has been selected to model the damping effect of the panel material. This model parameter has been set to b=3.2e-6. Prescribed Displacement/Rotation boundary condition (simply supported) have been applied.
Comment:
Line 110 – Rayleigh damping can be described in more detail or at least referenced.
Answer:
Many thanks for this recommendation. A reference has been added as well as the Rayleigh parameter used in our study. Please compare with the answer to your previous comment.
Comment:
Line 145 – yellow formatting probably by mistake.
Answer:
Has been removed.
Comment:
Line 172 – please consider zoom of 0.1s which means amplitude reduction in another figure.
Answer:
Many thanks for your suggestion. To show the full time-span in figure 6 without a zoom is intended to visualize how the proposed semi-active damping approach reduces the needed by the system “to come to rest” after a step input is given. We addressed your comment as follows:
Comparing the results in the time interval t=1s and t=10s it is obvious that the amount of damping is significantly increased for the system with SE.
Comment:
Figure 6 – Is it necessary to publish first second, because there are zeros in that time interval. Please consider use of mm instead of m for Y-axis
Answer:
Many thanks for your question. We believe that also showing the time span 0s – 1s is important, because the reader can easier detect the step input. Furthermore, it is easier to compare the step input “level”. In the text we addressed your comment as follows:
After comparing Fehler! Verweisquelle konnte nicht gefunden werden.(a) and (b) we find that the initial conditions at t=1s (time of step input) are identical for both systems. We also find that the SE with RL-shunt can reduce the vibrations in a short period of time.
We also believe that using basic SI-units is helpful for the reader. For this reason we would like to use m instead of mm.
Comment:
As you stated that FEM was used, therefore I recommend to supplement Fig. 7 by publishing pictures of four mode-shapes with/wo SE.
Answer:
Many thanks for your comment. Because de-tuning the system usually only alters the natural frequency but not the associated mode shape, we believe that presenting the first mode shapes in Appendix B for the plate without SE is sufficient in order to help the reader to trace our contribution. As also known from modal analysis, every mode shapes (eigen-vector) is determined up to a proportional factor. Also for this reason we decided not to compare the eigen-vectors obtained from the analysis of free-vibrations.
Comment:
Line 199 Why have not used 142.2 Hz instead of 143 Hz for angular frequency calculation. Is not a reason for not exact resonance graph in Fig. 8. The following note is related to this.
Answer:
Many thanks for your comment. The exact position of the resonance peak every-time depends also on the FFT-resolution. We tried to choose an excitation frequency that ensures that the panel vibration vibrates in a forced vibration mode that is close to the first eigen-vector. The results shown in Figure 9 prove that this has been achieved by using 143Hz as excitation frequency. For this reason it has been possible to discuss the narrowband effect of the proposed semi-active control approach. This has been the intention of subsection 3.2.2. However, your comment help us to formulate more precise about the natural frequencies seen in Figure 7. To argue Appendix B is also helpful. We used the following wording.
In Fehler! Verweisquelle konnte nicht gefunden werden.(a) we can find that 142.2Hz, 290.5Hz, 419.6Hz and 567.3Hz are the first four dominating natural frequencies of the panel. A resonance at 540Hz is of lower magnitude. After comparing Fehler! Verweisquelle konnte nicht gefunden werden.(a) and (b) we see that the vibration reduction effect is most significant at the first natural frequency whilst at the other three natural frequencies the amplitude remains nearly unchanged.
Comment:
Line 207 – You are comparing 4.5 10-3 and 0.6 10-4 but 0.6 value is at the end of the time interval or we can say steady value but 4.5 is at the beginning.
Answer:
Many thanks for your comment that has been taken into account as follows:
The maximum amplitude of the panel without SE is reduced from m down to m. Hence, the vibration level has been reduced by 37.5dB. A reduction of 32.4dB is to be found, if the steady state is compared at t=9.
Comment:
Line 224 – Why has been used the panel with different dimensions as that is described in Section 2.
Answer:
Many thanks for your comment that has been taken into account as follows:
The dimensions of the panel are not identical to the dimensions described in section 2. However, to proof the feasibility of semi-active vibration control concept experimentally in a first step, it has not been necessary to match the parameter of the numerical model.
Comment:
Line 241 – please describe setting of DAQ system of B&K type 3050, e.g. sampling rate at least.
Answer:
Many thanks for your comment that has been taken into account as follows:
Additionally, an impulse hammer (B&K type 8206), a microphone, a multi-input-frontend (B&K type 3050 using 5kHz as sampling frequency and an integrated digital high-pass filter with 0.7Hz corner frequency for all inputs) and three accelerometers (B&K type 4508 B) have been used.
Comment:
Figure 14 – The position 3 is depicted on Fig. 14 (b). I recommend using picture from the stage where SE was attached to P2. Then, the results from this position continue in Fig. 15 and Fig. 16.
Answer:
Many thanks for your comment. We are sorry that we cannot provide such a photo. For this reason your comment has been taken into account as follows:
Figure 1. Experiment’s setup: (a) back side (b) front side with SE mounted on position P3
Comment:
References – They are mistakenly pasted twice. I recommend at least 5 more references.
Answer:
Many thanks for your comment. Please find the corrected list of references with seven relevant additional entries.
Author Response File: Author Response.pdf
