Model Updating for a Continuous Concrete Girder Bridge Using Data from Construction Monitoring

Round 1

Reviewer 1 Report

Model updating for a continuous concrete girder bridge using data from the construction monitoring

 

Manuscript Number:

In the present paper authors provide a FE model updating method using data from construction monitoring is proposed in the article. Comparisons has been conducted between static structural responses calculated using the FE models and the data measured on the physical truth demonstrate the effectiveness of the proposed model updating approach and the superiority of the new method to the traditional method.  However, the paper requires some improvement before it can be recommended for publication, it is proposed to re-submit a thoroughly revised version of the manuscript, considering the following comments.

 

1.     Title is ok 2.     Overall recommendation should be reported in one sentence at the end of the abstract 3.     There are many studies in the related topic. So, introduction needs enrichments 4.  The authors should overview the recent progress made in the relevant area in the past two years or so. Such as: https://doi.org/10.1080/15732470412331289323; https://doi.org/10.3390/buildings13010185 5.     Emphasizing the importance of research in introduction 6.     Methodology: Authors should state in the methodology, procedures, geometry modeling  and concrete material model should be discussed comprehensively 7.     Clarity of Figure 7 must be improved 8.     The paper is well written and it is easy to follow, only the authors needs to go thoroughly revised version to correct the typo-mistake. 9.     Author should highlight the assumptions and limitations and future research direction of the study.

 

Author Response

Reviewer 1

 

In the present paper authors provide a FE model updating method using data from construction monitoring is proposed in the article. Comparisons has been conducted between static structural responses calculated using the FE models and the data measured on the physical truth demonstrate the effectiveness of the proposed model updating approach and the superiority of the new method to the traditional method.  However, the paper requires some improvement before it can be recommended for publication, it is proposed to re-submit a thoroughly revised version of the manuscript, considering the following comments.

 

1. Title is ok

Response: Thank you.

 

2. Overall recommendation should be reported in one sentence at the end of the abstract

Response: Thank you for the good suggestion. As suggested, we have added an overall recommendation in one sentence at the end the abstract (see the part highlight in blue).

 

3. There are many studies in the related topic. So, introduction needs enrichments

Response: Thank you for the comment. As suggested, many related studies reported by literatures in recent years have been cited in the revised manuscript [18-22], and they are discussed to enrich the introduction part (see the part highlighted in red).

 

4. The authors should overview the recent progress made in the relevant area in the past two years or so. Such as: https://doi.org/10.1080/15732470412331289323; https://doi.org/10.3390/buildings13010185

Response: Thank you for recommending these two important articles published recently. We have read and cited all of them [16,17], and overviewed the recent progresses based on them.

 

5. Emphasizing the importance of research in introduction

Response: The importance of the present research has been emphasized at the end of the introduction part (see the part highlighted in green).

 

6. Methodology: Authors should state in the methodology, procedures, geometry modeling and concrete material model should be discussed comprehensively

Response: Thank you. The missing information (the numerical simulation procedure, the geometric modeling and the concrete material model adopted) has been added in the revised manuscript to make it more informative (see the part highlight in purple).

 

7. Clarity of Figure 7 must be improved

Response: Thank you. We have revised the figure for clearness.

 

8. The paper is well written and it is easy to follow, only the authors needs to go thoroughly revised version to correct the typo-mistake.

Response: Thank you. We have reread the manuscript several times for typos. Some language mistakes have been corrected accordingly. Thank you for helping us improve the quality of our manuscript.

 

9. Author should highlight the assumptions and limitations and future research direction of the study.

Response: Thank you. We have presented the assumptions, the limitations and the future research direction of the study in the revised manuscript (see the last paragraph of the conclusion part).

Reviewer 2 Report

The paper is well written; however, there are some points to address:

1. Based on Table 3, why is there a difference between the model with updating technique and without it in case of relative error? (see position 34.5 m)

2. Why is there a deviation in stress measured on the continuous concrete girder bridge?

3. Give some recommendations for future work.

Author Response

Reviewer 2

 

The paper is well written; however, there are some points to address:

 

1. Based on Table 3, why is there a difference between the model with updating technique and without it in case of relative error? (see position 34.5 m)

Response: Thank you for the comment. Yes, for some locations (e.g., 34.5 meter position), the relative error is greater for the proposed model updating method than for the traditional method and the initial model without model updating. This is because proposed method cannot guarantee the accuracy improvements of the model for all responses in all positions; however, it has been proven that the model has been overall improved after it is updated using the proposed method, since it is calculated that the RMS error of static structural responses is 32.74% for the model of Huangsha Harbor Bridge updated using the new method, and they are 41.07% and 38.48% for the initial model without model updating and the model updated using the traditional method, respectively.

 

2. Why is there a deviation in stress measured on the continuous concrete girder bridge?

Response: According to our experiences, two issues lead to the deviation in stress measured on the concrete bridge: first, the uneven temperature field on the structure exerts the temperature gradient-induced stresses on the measured data; second, stresses measured in concrete using vibrating wire stress transducers might be inaccurate when the vibrating wire stress transducers are used over a certain period due to the inherent weakness of the measuring technique. To deal with the first issue, we use a FE analysis-based method to compensate for the temperature-induced deviation. For the second issue, we update the stress arrays measured based on GM (1,1) models. Thank you for your incisive comment.

 

3. Give some recommendations for future work.

Response: Thank you. We have presented the limitations and the future research direction of the study in the revised manuscript (see the last paragraph of the conclusion part).

Reviewer 3 Report

The paper corresponds to the topics of scientific journal.

Paper requires correction.

1. The abstract is representative of the content, but has no goal, scientific novelty. 

2. The literature review is a weak part of this paper. Author did not provide comparations of gained results with similar up-to-date studies in the field. The author did not conduct a critical analysis of contemporary publications in order to justify the relevance. There are many publications on this topic.

3. The author talks about the stress measurement using the vibrating wire stress transducers. It is known that deformations are measured. Stress is a measure of internal forces, it is impossible to measure it directly. After that, the author proposes to improve the finite element model. And he again works with stresses. Why, when measuring deformations, recalculate them into stresses and then compare them with stresses obtained in a finite element package. After all, the finite element package also first counts the deformations, and then, according to the physical law, recalculates them in stresses. Why does the author not carry out calculations on deformations? Thus, it can eliminate a number of errors in the mathematical models of the materials used.

4. The author does all the improvement of the finite element model on the basis of 3.2. "Compensation for the deviation of the measured stress due to the uneven temperature". However, the finite element modeling of this important part of the article was not carried out quite correctly. First, the author applied a large finite element mesh. According to the thickness of the object (in the direction of temperature change), the author generated only one finite element. This does not allow for the actual distribution of temperatures. It is necessary to split at least into three finite elements. And so - there is a finite node outside, there is inside - and what parameters between them are not clear? There is no description of the accepted models, material models in the finite element package, the dimension of the problem. There are no physical and mechanical characteristics of materials. Are there no boundary conditions? 

5. The author talks about experimental research. There is no processing of the experiment: what are the deviations, mathematical expectation and other parameters. It is necessary to specify the experiment processing parameters, such as mathematical expectation, deviations and etc. 

Based on the results obtained, the author applies the Gray theory. Therefore, these results must be confirmed.

6. Discussion section needs the revision for the scientific standard:

 – what can be considered the advantages of this study compared to analogues (and it is mandatory to indicate alternative methods as a basis for comparison)

 – what are the shortcomings of the study

 – what could be the development of this research and why exactly in this.

Author Response

Reviewer 3

 

The paper corresponds to the topics of scientific journal.

 

Paper requires correction.

 

1. The abstract is representative of the content, but has no goal, scientific novelty.

Response: Thank you for the comment. As suggested, the goal and the scientific novelty of the present study have been added in the revised abstract (see the part highlighted in bold).

 

2. The literature review is a weak part of this paper. Author did not provide comparations of gained results with similar up-to-date studies in the field. The author did not conduct a critical analysis of contemporary publications in order to justify the relevance. There are many publications on this topic.

Response: Thank you. As suggested, many related important articles published in recent years have been cited in the revised manuscript [18-22], and they are discussed to enrich the introduction part (see the part highlighted in red).

 

3. The author talks about the stress measurement using the vibrating wire stress transducers. It is known that deformations are measured. Stress is a measure of internal forces, it is impossible to measure it directly. After that, the author proposes to improve the finite element model. And he again works with stresses. Why, when measuring deformations, recalculate them into stresses and then compare them with stresses obtained in a finite element package. After all, the finite element package also first counts the deformations, and then, according to the physical law, recalculates them in stresses. Why does the author not carry out calculations on deformations? Thus, it can eliminate a number of errors in the mathematical models of the materials used.

Response: According to the operating principle of the vibrating wire stress transducers, the natural frequency of the vibrating wire is measured first; then, the strain is directly calculated by the transducer using the measured natural frequency based on a calibrated formula without providing the deformation information. Yes, in nature, when the mechanical state changes for a structure, the corresponding change of the natural frequency of the vibrating wire stress transducer is directly related to the structural deformation, and the structural deformation in turn induces the strain. However, the vibrating wire stress transducers are basically commercial available instruments, and we cannot change their functions by ourselves. Therefore, the deformations are not measured on location during the construction stages (the data listed in Table 3 are measured using a Total Station after the bridge is built), and we cannot use the deformations measured on location to update the FE model. I don’t know if I explain this clearly. Anyway, thank you for your incisive observation.

 

4. The author does all the improvement of the finite element model on the basis of 3.2. "Compensation for the deviation of the measured stress due to the uneven temperature". However, the finite element modeling of this important part of the article was not carried out quite correctly. First, the author applied a large finite element mesh. According to the thickness of the object (in the direction of temperature change), the author generated only one finite element. This does not allow for the actual distribution of temperatures. It is necessary to split at least into three finite elements. And so - there is a finite node outside, there is inside - and what parameters between them are not clear? There is no description of the accepted models, material models in the finite element package, the dimension of the problem. There are no physical and mechanical characteristics of materials. Are there no boundary conditions?

Response: Thank you for pointing out our mistake. As suggested, we have remeshed the model with refined elements (see Fig. 4) and recalculated. As can be seen in Fig. 4, in the direction of temperature change, the thickness of the object has been divided into 2~5 elements, which can reflect the temperature gradient on the thickness of the object. Besides, more details of the established model are reported for references, including the geometry, the material properties, and the boundary conditions (see the part highlighted in pink).

 

5. The author talks about experimental research. There is no processing of the experiment: what are the deviations, mathematical expectation and other parameters. It is necessary to specify the experiment processing parameters, such as mathematical expectation, deviations and etc.

 

Based on the results obtained, the author applies the Gray theory. Therefore, these results must be confirmed.

Response: Thank you for the comment. In Table 3, the statistical parameters (the expectation, the deviation, the kurtosis and the skewness) are added for data arrays listed in the table, from which it can be concluded that the model updated using the proposed method is closer to the physical truth, compared with the model without model updating and the model updated using the traditional method. However, we cannot give the statistical parameters for the stresses measured on location during the construction stages. This is because we have only collected one data from each transducer in each construction stage (there is no repetition of measurements), so the samples are not enough for conducting the statistical analyses. Anyway, thank you for your good suggestion.

 

6. Discussion section needs the revision for the scientific standard:

 

 – what can be considered the advantages of this study compared to analogues (and it is mandatory to indicate alternative methods as a basis for comparison)

Response: Thank you. We have presented the advantages of this study compared with the analogue, and the alternative method as the basis for comparison is also indicated (see the part highlighted in italic in the conclusion part).

 

 – what are the shortcomings of the study

• what could be the development of this research and why exactly in this.

Response: Thank you. We have presented the limitations and the future research direction of the study in the revised manuscript (see the last paragraph of the conclusion part).

Round 2

Reviewer 1 Report

The authors addressed most of the comments. The paper can be recommended for publication 

 

Author Response

The authors addressed most of the comments. The paper can be recommended for publication

Response: I sincerely thank the reviewer for the time and efforts invested in reviewing this manuscript. I hope the quality of the present manuscript can reaches the standard of a publication in Applied Sciences.

Reviewer 3 Report

The authors substantially revised the manuscript. All my comments were taken into account. I think that the article has become better and can be published.

I recommend to improve the quality of fig.

Author Response

The authors substantially revised the manuscript. All my comments were taken into account. I think that the article has become better and can be published.

Response: I sincerely thank the reviewer for the time and efforts invested in reviewing this manuscript. I hope the quality of the present manuscript can reaches the standard of a publication in Applied Sciences.

 

I recommend to improve the quality of fig.

Response: As suggested, I have improved the quality of Figs. 5-7 for clearness. Thank you for helping me improve the quality of my manuscript.