Monitoring the Calibration Status of a Universal Testing Machine Through the Implementation of Acoustic Methods: Development of Equipment and a Suitable Interface
Round 1
Reviewer 1 Report
Comments and Suggestions for Authorsthis paper presents a concept to confirm a stress-strain instrument calibration by comparing the results on a test sample with the mechanical properties of the same sample measured by an acoustic method.
The method has some merit and is well presented. Overall, the pictures could be higher resolution so they can be more readable. In my opinion, the GUI gets too much attention. About a full page is dedicated to it; the emphasis should be on the algorithm. While I understand that GUIs are always overlooked in most software, and I understand that one may want to recognize their usefulness, the one presented here (as much as I can tell, without being able to read the text, due to picture quality) is nothing special.
The concept of integrating calibration confirmation and measurement in one package is good though.
That said, there are a few typos that should be corrected:
- line 346 has one too many "is"
- line 336 states "In each stage the sample was loaded and unloaded 25 times, thus at the end of the third stage the sample was subjected to a cyclic loading of 100 times". How does 25 x 3 = 100?
- line 157 either "shall" was meant to be "falls" or a word is missing.
There are some comments that could be done about nonlinearities generated by micro-cracks, that would likely explain the issues in the AL sample.
Resonant ultrasound spectroscopy is another way to determine the elastic constants that would be more precise than the method presented here, but that may not be necessary. Additional details about the acquisition length would have been good to inform on the spectrum resolution.
Overall, it is a good paper and could be publish as is, but I think the few corrections mentioned above would make it better.
Author Response
This paper presents a concept to confirm a stress-strain instrument calibration by comparing the results on a test sample with the mechanical properties of the same sample measured by an acoustic method.
The method has some merit and is well presented. Overall, the pictures could be higher resolution so they can be more readable. In my opinion, the GUI gets too much attention. About a full page is dedicated to it; the emphasis should be on the algorithm. While I understand that GUIs are always overlooked in most software, and I understand that one may want to recognize their usefulness, the one presented here (as much as I can tell, without being able to read the text, due to picture quality) is nothing special.
Answer: Necessary changes have been made in the draft. The clarity of figure has been improved overall, and the algorithmic description has been improved as well.
The concept of integrating calibration confirmation and measurement in one package is good though.
That said, there are a few typos that should be corrected:
- line 346 has one too many "is"
- line 336 states "In each stage the sample was loaded and unloaded 25 times, thus at the end of the third stage the sample was subjected to a cyclic loading of 100 times". How does 25 x 3 = 100?
- line 157 either "shall" was meant to be "falls" or a word is missing.
Answer: The above issues have been rectified. The is’s has been removed from lines 388, 389, 390.
Loading time calculation has been corrected in section 4.2.
The grammatic error in line has been corrected, “shall be”. (line 171)
There are some comments that could be done about nonlinearities generated by micro-cracks, that would likely explain the issues in the AL sample.
Answer: The remark is duly noted.
Resonant ultrasound spectroscopy is another way to determine the elastic constants that would be more precise than the method presented here, but that may not be necessary. Additional details about the acquisition length would have been good to inform on the spectrum resolution.
Answer: The explanation has been added in the draft . The corrections can be found in section 3.3. (from line 282 to 296)
Overall, it is a good paper and could be publish as is, but I think the few corrections mentioned above would make it better.
Please refer the corrections on un-tracked version of the draft.
Author Response File: Author Response.docx
Reviewer 2 Report
Comments and Suggestions for AuthorsThis study provides a novel method for monitoring the calibration status of universal testing machine, which uses acoustic means to determine Young's modulus. This model has achieved decent recognition accuracy, which has certain academic and practical value. However, it needs some modifications before it can be accepted.
1. The format of some formulas, charts and pictures in the article is not standardized, and the author needs to carefully revise.
2. As for the actual situation of the experiment, the author needs more pictures to explain.
3. In the Introduction, the authors should quickly elaborate on the specific issues which this paper focuses on and explain the status of recent research. The current introduction makes it difficult for the reader to understand.
4. This paper presents an acoustic-based method for measuring the Young's modulus (E) of materials, but it is obvious that the measurement results will change due to the change of the clamping position. How is this change controlled and what impact will its uncertainty have on the measurement results?
Comments on the Quality of English LanguageIt can be further improved
Author Response
This study provides a novel method for monitoring the calibration status of universal testing machine, which uses acoustic means to determine Young's modulus. This model has achieved decent recognition accuracy, which has certain academic and practical value. However, it needs some modifications before it can be accepted.
- The format of some formulas, charts and pictures in the article is not standardized, and the author needs to carefully revise.
Answer: Necessary changes have been made in the draft.
- As for the actual situation of the experiment, the author needs more pictures to explain.
Answer: Necessary changes have been made in the draft. A new picture of the test equipment has been show, fig 2.
- In the Introduction, the authors should quickly elaborate on the specific issues which this paper focuses on and explain the status of recent research. The current introduction makes it difficult for the reader to understand.
Answer: The current stage of the research has been highlighted in paragraph 2 of the paper. The structure of the introduction is as follows:
Paragraph 1: Explanation of UTM and its relevance in the testing industry.
Paragraph 2: Current state of calibration in UTMs and its drawbacks, highlighted towards the end of the paragraph.
Paragraph 3: Relevance of E-modulus as a calibration parameter
Paragraph 4: Methods for determining E-modulus, conventional methods.
Paragraph 5: Issues with the current methods to determine E, and a logic on how the same could be used to check machine calibration status.
Paragraph 6: Acoustic methods to determine E.
Paragraph 7: How both the methods could be combined and a methodology developed to check calibration status of the machine.
Paragraph 8: Structure of paper is elaborated here.
- This paper presents an acoustic-based method for measuring the Young's modulus (E) of materials, but it is obvious that the measurement results will change due to the change of the clamping position. How is this change controlled and what impact will its uncertainty have on the measurement results?
Answer: The change in clamping position is controlled through standardized procedures, such as using precise, repeatable markings or fixtures to ensure consistent position and force. Pre-calibration with reference samples helps account for clamping variations. To minimize the impact of clamping uncertainty, repeatability tests are performed to quantify variations in resonant frequencies and their effect on the calculated Young’s modulus. Correction factors based on calibration data and finite element simulations can further mitigate residual uncertainties, ensuring reliable and accurate measurements.
Please track the changes in the untracked version of the article.
Author Response File: Author Response.docx
Reviewer 3 Report
Comments and Suggestions for Authors1. «Traditionally it was obtained from a straight line drawn on the linear part of the stress-strain curve, but with the advent of newer machines and new data-acquisition systems, better curve fittings are obtained to calculate E» (lines 62-65) – it is necessary to give examples of which new data-acquisition systems the authors had in mind.
2. The following designations in Table 1 must be explained: ,
,
,
. A description can be provided at the bottom of the table.
3. «The E measured from such an equipment would eventually deviate from the value measured on the day the equipment was calibrated, and as per established norms this shouldn’t be more than 5%» (lines 83-85) – In this case are we referring to differences in E for the reference material against which the calibration was performed? The text should be clarified.
4. The quality of all the images presented in this paper is very poor and it is sometimes so difficult to study them. It is required to provide images in a higher quality.
5. In equation 5, it is necessary to explain why the "-" is located in the numerator of the ratio for the 3rd and 4th components of the sum.
6. Equation 10 is not mentioned in the text. Explanations similar to equations 7-9 should be added.
7. In Chapter 3.1, it would be helpful to provide a brief explanation of the reasons why these two materials were selected for the non-destructive reference samples.
8. Also, the following statement is given in Chapter 3.2: "The gauge length measurement had an uncertainty of 100 µm " (line 218). To better understand this parameter, it is recommended to include a schematic representation of the gauge length in Figure 1.
9. For Table 6, it is necessary to re-describe the equation for calculating the uncertainties in load measurements (preferably as part of consideration of a single force measuring device used in the testing machine).
10. «A suitable adhesion material has to be used as coupling agent between sensor and the sample, and for this purpose bee wax was used» (lines 258-260) – the authors need to clarify whether the specified adhesive material contributes to the formation of acoustic signal interference and what the magnitude of this contribution is, as well as justify the choice of beeswax as an adhesive.
11. «The rest of the modes are omitted by the algorithm» (line 305) – it should be clarified which modes are excluded from consideration when implementing the algorithm.
12. «In each stage the sample was loaded and unloaded 25 times, thus at the end of the third stage the sample was subjected to a cyclic loading of 100 times. The change in E was not very significant and when the plots were checked, there was no significant deviations from one another» – the statement that there are no significant differences between the grouped data (0,1 and 0,15%) should be supported by appropriate statistical analysis. This includes assessing the normality of the distribution of the variable E for different deformation levels, as well as using parametric or non-parametric statistical models. Since there were 25 times, the statistical algorithms are well-suited for this analysis.
13. «Though there was a distinct third mode (803 Hz) seen, there were several other peaks which never corresponds to the frequencies seen in figure 8» – is it really necessary to include Figure 8 (Simulated modal frequencies of the reference sample (both Al6060, and M700).) in the text, or is there a mistake here?
14. «But, the stress-strain curves showed deviations within the first two stages, and when the E was ascertained, there was stark difference observed in each curve» – comments similar to paragraph 12.
15. «The validation of the proposed system has shown that the geometric parameters of the reference sample have a significant impact on the algorithm's output» – it is not clear on what basis the conclusion was drawn. According to the 3.1 Reference Sample (Figure 1 Reference sample dimensions), samples of the same geometric shape were used in the study. However, the thesis states that "the geometric parameters of the reference sample have a significant impact". Additional comments are required.
Comments for author File: Comments.pdf
Author Response
- «Traditionally it was obtained from a straight line drawn on the linear part of the stress-strain curve, but with the advent of newer machines and new data-acquisition systems, better curve fittings are obtained to calculate E» (lines 62-65) – it is necessary to give examples of which new data-acquisition systems the authors had in mind.
Answer: The correction has been made in line 62 to 69.
- The following designations in Table 1 must be explained: , ,, . A description can be provided at the bottom of the table.
Answer: The correction has been made in the table. (Line 71)
- «The E measured from such an equipment would eventually deviate from the value measured on the day the equipment was calibrated, and as per established norms this shouldn’t be more than 5%» (lines 83-85) – In this case are we referring to differences in E for the reference material against which the calibration was performed? The text should be clarified.
Answer: This has been clarified. Line 87 to 89.
- The quality of all the images presented in this paper is very poor and it is sometimes so difficult to study them. It is required to provide images in a higher quality.
Answer: Images with low resolution have been replaced in the draft.
- In equation 5, it is necessary to explain why the "-" is located in the numerator of the ratio for the 3rd and 4th components of the sum.
Answer: In the 3rd and 4th the partial differentiation is with respect to A0 and . Since the derivation is quite big, we didn’t want to put it in the paper. But in short I can explain with respect to 3rd component.
E is a function of F(force), (gauge length), cross-sectional area (, and change in gauge length , please see equation (A)
(A)
With respect to component 3, a partial differentiation of equation (A) with respect to will give the following:
= (B)
Using equation (A), one can substitute the expression (B) in the 3rd component as follows;
(C)
The E seen in equation (C) would be in each component and the uncertainties accruing from each component will be added to it.
- Equation 10 is not mentioned in the text. Explanations similar to equations 7-9 should be added.
Answer: The correction has been made in the draft.
- In Chapter 1, it would be helpful to provide a brief explanation of the reasons why these two materials were selected for the non-destructive reference samples.
Answer: This correction has been made with relevant references, as to why these two materials were adopted. (Line 211 to 229)
- Also, the following statement is given in Chapter 2: "The gauge length measurement had an uncertainty of 100 µm " (line 218). To better understand this parameter, it is recommended to include a schematic representation of the gauge length in Figure 1.
Answer: The correction has been made in the draft.
- For Table 6, it is necessary to re-describe the equation for calculating the uncertainties in load measurements (preferably as part of consideration of a single force measuring device used in the testing machine).
Answer: The values in Table 6 were provided by the firm that calibrated the machine and the calibration certificate provided thereafter. These values were used in conjunction with our sample data to calculate the uncertainties.
- «A suitable adhesion material has to be used as coupling agent between sensor and the sample, and for this purpose bee wax was used» (lines 258-260) – the authors need to clarify whether the specified adhesive material contributes to the formation of acoustic signal interference and what the magnitude of this contribution is, as well as justify the choice of beeswax as an adhesive.
Answer: The explanation has been added to the draft along with relevant references. (Line 288 to 295)
- «The rest of the modes are omitted by the algorithm» (line 305) – it should be clarified which modes are excluded from consideration when implementing the algorithm.
Answer: An additional explanation is added following the statement. (line 340 to 342)
- «In each stage the sample was loaded and unloaded 25 times, thus at the end of the third stage the sample was subjected to a cyclic loading of 100 times. The change in E was not very significant and when the plots were checked, there was no significant deviations from one another» – the statement that there are no significant differences between the grouped data (0,1 and 0,15%) should be supported by appropriate statistical analysis. This includes assessing the normality of the distribution of the variable E for different deformation levels, as well as using parametric or non-parametric statistical models. Since there were 25 times, the statistical algorithms are well-suited for this analysis.
Answer: We have included a table (table 8) in the draft copy showing the first five values of E recorded in each stage, along with the standard deviations. In adhering to this comment we are in the process of collecting more data with our reference samples from different makes of the machine (QUASAR). This is to built a training data for a set of machine learning algorithms (random forest, linear regression), and thereby help predicting when the machine should be next calibrated.
- «Though there was a distinct third mode (803 Hz) seen, there were several other peaks which never corresponds to the frequencies seen in figure 8» – is it really necessary to include Figure 8 (Simulated modal frequencies of the reference sample (both Al6060, and M700).) in the text, or is there a mistake here?
Answer: The statement was wrong and has been duly corrected. (Line 407 and 408)
- «But, the stress-strain curves showed deviations within the first two stages, and when the E was ascertained, there was stark difference observed in each curve» – comments similar to paragraph 12.
Answer: A correction has been carried out with respect to this statement.
- «The validation of the proposed system has shown that the geometric parameters of the reference sample have a significant impact on the algorithm's output» – it is not clear on what basis the conclusion was drawn. According to the 1 Reference Sample (Figure 1 Reference sample dimensions), samples of the same geometric shape were used in the study. However, the thesis states that "the geometric parameters of the reference sample have a significant impact". Additional comments are required.
Answer: There has been a slight error in what the authors wanted to convey here. The statement has been deleted and a new explanation has been inserted in its place conveying the actual meaning. (Line 512-514)
All the corrections with respect to line can be seen in the untracked version.
Author Response File: Author Response.docx
Reviewer 4 Report
Comments and Suggestions for Authors
The paper describes a methodology for applying two testing procedures and a reference sample that can be used non-destructively.
In the introduction it is stated how the research combines established testing methods with a specific testing material for checking UMT’s calibration statuses. The use of a software application and a relative GUI. When it is stated that “to the authors knowledge this technique was not implemented in the past” is this related the combination of whole the elements previously mentioned, or it is referred to the software application only? This should be better clarified.
The main contribution of the research resides in the establishment of the methodology, in the development of the non-destructive testing sample, or in the software application? This should to be clarified in the introduction.
Within the theoretical consideration chapter, in section 2.1, it is mentioned that 2 out of the 4 categories for sources of uncertainties are used. If the influence of the discarded categories cannot be discounted, what is the rationale behind selecting the first two and do not consider the third and fourth? This should be clarified.
I suspect you might have a missing verb after “shall” in lone 157.
In chapter 3, section 3.1, the testing sample is presented. This is presented in the paper as one of the points of impact of the conducted research. Nevertheless, the rationale behind the choices of materials or their dimensions is not discussed. As this is presented as a very important foundational point of the project, the rationale behind the choice of testing material should be discussed. I suggest expanding section 3.1 to include a more in details description of the rationale behind the testing samples.
In chapter 3, section 3.3, it is described how “different methods were employed to introduce sound”. Given the acoustics measurement approach, it is important to specify that the methodology involves a mechanical excitation of a sample under test with the purpose of recording the sound produced by the object itself, and that sound is not applied to the testing sample in any way.
In the introduction, it is mentioned how acoustic methods make use of microphones, piezo, or vibrometer. In the proposed approach the authors employ an accelerometer as a way of recording the sound produced by the vibrating test sample. The rationale behind this choice should be presented.
In line 270, it is indicated that the sampling frequency was chosen to be 16kHz. Why this choice? The choice of sampling frequency is highly dictated by the range of frequencies expected to be in the recorded signal. How was it predicted what kind of frequencies (or in what range) would appear at the accelerometer to inform the choice of sampling frequency?
In chapter 4, section 4.2, it is described how the spectral content of the measurement is affected by repetitive testing and increase strain of the test sample. In line 344 it is stated that “this is not ideal when using aluminium […]”. Why is that? The reasoning behind the situation not being ideal should be mentioned to link to what the appearance of new frequency indicates.
In the following page (lines 362 to 366) It is remarked how the resulting spectra reinforce the observation that new peaks in the frequency content does not point towards the best result. What do the spectra indicate? There should be an explanation of what these peaks indicate and what kind of negative information they convey. E.g. a representation of loss of stiffness over time, structural damages, etc.
Chapter 5 (line 436) states that the study “presented a new method for non-destructive monitoring of UTM calibration using acoustic techniques. This should also mention the tension test data and the use of a dedicated test sample to be comprehensive of all its aspects.
In line 444 it is said that Al6060 resulted limited in suitability for continuous monitoring. These aspects should be analysed and discussed further in the results section.
Line 456 says that the reference sample has retained its elastic properties. Is this referring only to the M700 or both? If the loss of elasticity is what it is demonstrated by the spectral analysis it would support the result section to analyse this outcome.
Figures should have a better resolution. Figure 8 specifically is very hard to read.
Author Response
The paper describes a methodology for applying two testing procedures and a reference sample that can be used non-destructively.
In the introduction it is stated how the research combines established testing methods with a specific testing material for checking UMT’s calibration statuses. The use of a software application and a relative GUI. When it is stated that “to the authors knowledge this technique was not implemented in the past” is this related the combination of whole the elements previously mentioned, or it is referred to the software application only? This should be better clarified.
Answer: Here we are referring to the whole process together, as when we checked literature no such combined strategies exist for checking machine calibration status. But this has now been updated in the article. Please see line 118 to121.
The main contribution of the research resides in the establishment of the methodology, in the development of the non-destructive testing sample, or in the software application? This should to be clarified in the introduction.
Answer: This has been clarified in the paper. This has been clarified between line 101 to 121.
Within the theoretical consideration chapter, in section 2.1, it is mentioned that 2 out of the 4 categories for sources of uncertainties are used. If the influence of the discarded categories cannot be discounted, what is the rationale behind selecting the first two and do not consider the third and fourth? This should be clarified.
Answer: This has been clarified in the section. Line 136 to 142.
I suspect you might have a missing verb after “shall” in lone 157.
Answer: Corrected (line 171)
In chapter 3, section 3.1, the testing sample is presented. This is presented in the paper as one of the points of impact of the conducted research. Nevertheless, the rationale behind the choices of materials or their dimensions is not discussed. As this is presented as a very important foundational point of the project, the rationale behind the choice of testing material should be discussed. I suggest expanding section 3.1 to include a more in details description of the rationale behind the testing samples.
Answer: A detailed description has been incorporated is the said section. (Line 211 to 223)
In chapter 3, section 3.3, it is described how “different methods were employed to introduce sound”. Given the acoustics measurement approach, it is important to specify that the methodology involves a mechanical excitation of a sample under test with the purpose of recording the sound produced by the object itself, and that sound is not applied to the testing sample in any way.
Answer: The section has been corrected. (Line 268 to 270).
In the introduction, it is mentioned how acoustic methods make use of microphones, piezo, or vibrometer. In the proposed approach the authors employ an accelerometer as a way of recording the sound produced by the vibrating test sample. The rationale behind this choice should be presented.
Answer: The rationale for adopting an accelerometer has been explained. (Line 281 to 295)
In line 270, it is indicated that the sampling frequency was chosen to be 16kHz. Why this choice? The choice of sampling frequency is highly dictated by the range of frequencies expected to be in the recorded signal. How was it predicted what kind of frequencies (or in what range) would appear at the accelerometer to inform the choice of sampling frequency?
Answer: Yes, you are right, the choice of sampling frequency is highly dependent on the range of frequencies expected to be in the recorded signal. This was predicted from the modal analysis that was carried out in the paper. For example, in the case of M700, the range was between 200 and 1000 Hz. That’s, another reason we conducted the modal analysis first and then we knew what frequencies we are looking for and hence the choice of the sampling rate.
In chapter 4, section 4.2, it is described how the spectral content of the measurement is affected by repetitive testing and increase strain of the test sample. In line 344 it is stated that “this is not ideal when using aluminium […]”. Why is that? The reasoning behind the situation not being ideal should be mentioned to link to what the appearance of new frequency indicates.
Answer: The argument is updated in the draft. (Line 384 to 390).
In the following page (lines 362 to 366) It is remarked how the resulting spectra reinforce the observation that new peaks in the frequency content does not point towards the best result. What do the spectra indicate? There should be an explanation of what these peaks indicate and what kind of negative information they convey. E.g. a representation of loss of stiffness over time, structural damages, etc.
Answer: A further argument with this regard has been added to the section with new reference. (Line 399-415)
Chapter 5 (line 436) states that the study “presented a new method for non-destructive monitoring of UTM calibration using acoustic techniques. This should also mention the tension test data and the use of a dedicated test sample to be comprehensive of all its aspects.
Answer: This has been rectified (from line 489 to 505)
In line 444 it is said that Al6060 resulted limited in suitability for continuous monitoring. These aspects should be analysed and discussed further in the results section.
Answer: This has been rectified with more arguments in the result section substantiated with new references.
Line 456 says that the reference sample has retained its elastic properties. Is this referring only to the M700 or both? If the loss of elasticity is what it is demonstrated by the spectral analysis it would support the result section to analyse this outcome.
Answer: An argument has been included in the result section along with inclusion of M700 after the above-mentioned statement, “Even with continuous use, the reference sample (M700) has maintained its elastic properties”
Figures should have a better resolution. Figure 8 specifically is very hard to read.
Answer: The picture has been updated, but now the figure is Figure 9.
Please refer to the untracked version of the paper with reference to all the lines.
Author Response File: Author Response.docx
Round 2
Reviewer 2 Report
Comments and Suggestions for AuthorsThe issues have been addressed.
Author Response
Thank you for the review
Reviewer 4 Report
Comments and Suggestions for AuthorsI am happy with the present state of the paper post reviews.
Author Response
Thank you for the review