A Combined Safety Monitoring Model for High Concrete Dams

Round 1

Reviewer 1 Report

The authors have developed a model based on the Incremental Extreme Learning Machine, Least Median Squares Method and Extreme Learning Machine Method, for eliminating the Gross errors in the structural health monitoring - reliability analysis system and further to predict the behavior of structural concrete dams. The work has been carried out in the detail explaining all the literature review, identified the gaps, provided detail mathematics equations for each of the methods and finally validated the proposed model with a case study. 

Very Minor Changes are suggested 
Figure 3 - the picture is not clear, needs revisions and improved quality 

Figures 7,8,9,10,11 - Legend - "actrual" must be "actual"

Author Response

Point 1: Figure 3 - the picture is not clear, needs revisions and improved quality

Response 1:  We fully agree with your valuable views. Thank you for your comment. We improved the clarity of Figure 3 and enlarged the picture as required.

Point 2: Figures 7,8,9,10,11 - Legend - "actrual" must be "actual"

Response 2:  We fully agree with your valuable views. We thank reviewer for bring this advice. We have modified the figures as requested.

Author Response File: Author Response.docx

Reviewer 2 Report

LMS-IGG-ELM method for identification gross errors and predicting behavior of concrete dams

 

Yanbo Wang1,2, Hao Gu1,2,* , Yating Hu1,2, Meng Yang3 , Chongshi Gu1,2, Wenhan Cao1,2, Zheng 4 Fang1,2

 

The use of abbreviations in the title creates some discomfort and if it is not a difficulty the authors can reconsider it.

 

Abstract:

 

The thread in question is a necessary means of tracking the traditional behaviour (because we don't have many others) of civil engineering systems. It is the most popular means of evaluating the working mechanisms of systems and their surrounding environment. As such means, based on multiple measurements, it is necessary to use a system of statistical methods to evaluate the obtained results, the elimination of gross errors, as one of the most important tasks for the analysis of experimental measurements in all fields of civil engineering for multiple facilities. I would recommend the authors to pay attention in the abstract to the technical elements and measurement techniques, because the results depend on them the most. I recommend revision.

 

Keywords:

Keywords include all aspects of the elements that are the subject of this study and are very well thought-out.

 

Main text:

.

Paragraph 1

The examination of the monitoring system has the task of giving an assessment of the working mechanisms of the engineering systems of a similar class, given the scale and impact on the human environment. This is an extremely important goal. It can be pursued by many means, but the most important thing is to take preventive measures that are markers of danger and used as a means of evaluating the working mechanism in order to preserve the durability of the structural system. What will be the means to be used is undoubtedly important, but it is essential here to analyse these results. Based on the results, the important engineering conclusions about the health of the structural system are made. My expectation is to state the goals and only then to analyse the methods as a means of compensation.

 

I recommend that the authors review this publication - GABBAR, Hossam A., et al. CTIMS: Automated Defect Detection Framework Using Computed Tomography. Applied Sciences, 2022, 12.4: 2175.

Line 58 - The main application of measurement results is related to reverse analysis. Their credibility was determined by the credibility of the input data, and this is the most important argument.

 

Line 97 - Please explain the abbreviation RBF.

Line 111 - Please see the analysis in this material

Desire L. Massart, Leonard Kaufman, Peter J. Rousseeuw, Annick Leroy, Least median of squares: a robust method for outlier and model error detection in regression and calibration, Analytica Chimica Acta, Volume 187, 1986, Pages 171-179, ISSN 0003-2670,

 

Line 123 - The use of the concept of robustness requires a wide range of studies, including physical and mechanical parameters of materials. Here, however, you can also note the fast Fourier analysis as a tool for determining the value with the greatest frequency in the occurrence of serial measurements of a parameter.

 

Paragraph 2

 

Line 157 - You could also pay attention to some fundamental aspects of the Least Median of Squares Regression method Hodges (1967).

ANDREWS, D.F., BICKEL, P.J., HAMPEL, F.R., HUBER, P.J., ROGERS, W.H., and TUKEY, J.W. (1972), Robust Estimates of Location: Survey and Advances, Princeton, N.J.: Princeton University Press.

 

Paragraph 3

Row 269 – From a geotechnical point of view, deformations that may be caused by the ground base of the wall and deformations of the left and right support blocks are of interest. It is unlikely that temperature deformations will have a greater contribution.

 

Row 269 – Figure 5 does not read well, could be improved.

 

Row 300 - Figure 6. You can also add the reactions of other control positions. Thus, with only one illustration, confirmation of the measurements cannot be sought. Figure needs a legend, out-of-trend points need discussion.

 

Row 318 - Table 4. It is not clear why the assessment of gross errors was carried out by only two of the methods. It requires clarification

Row 323 - Table 5. It is appropriate to give another group of gross errors and the percentage of their filtering. In my opinion, a full conclusion cannot be drawn from only one result.

 

Row 355 - Figure 11. This comparison is interesting and highlights the upper and lower limits of the rating band. In considering formula 25, it would be interesting to take into account other elements of the unknown quantities - Mood, A.; Graybill, F.; Boes, D. (1974). Introduction to the Theory of Statistics (3rd ed.). McGraw-Hill.

 

Paragraph 4

 

The contribution of gross errors is indisputable. The real question remains, is the computational (analytical approach) sufficient for identification, is there no other physical cause (technique, tools, etc.) or do we need a preceding qualitative one with the task of determining the physical causes of the processes that caused the settlement or displacement.

 

I have no special recommendations here, although the emphasis of this post is on the method and procedure itself, and my expectations, according to the title, were quite different.

 

Author Response

Point 1: The use of abbreviations in the title creates some discomfort and if it is not a difficulty the authors can reconsider it.

Response 1: Thank you for your advice. We fully agree with your valuable views.

We have revised the title based on the use of abbreviations in the title.

Point 2: I would recommend the authors to pay attention in the abstract to the technical elements and measurement techniques, because the results depend on them the most. I recommend revision.

Response 2: Thank you for your advice. We fully agree with your valuable views. Firstly, we have revised the abstract. Secondly, we have reviewed the technical elements and measurement techniques for the used method. It is a combination of several machine learning algorithms, using the advantages of different algorithms, to improve the efficiency and accuracy of computing. For the internal algorithm, it is actually a black-box model, and their principle has been developed in detail in the second chapter “Method”, and we fully verified the generalization ability and reliability of the method through the example part, so as to obtain the superiority of the method used.

Point 3: Keywords include all aspects of the elements that are the subject of this study and are very well thought-out.

Response 3: We thank reviewer for bring this advice. We fully agree with your valuable views. We reconsidered the keywords, added and corrected them to make them more consistent.

Point 4: I recommend that the authors review this publication - GABBAR, Hossam A., et al. CTIMS: Automated Defect Detection Framework Using Computed Tomography. Applied Sciences, 2022, 12.4: 2175.

Line 58 - The main application of measurement results is related to reverse analysis. Their credibility was determined by the credibility of the input data, and this is the most important argument.

Response 4: We fully agree with your valuable views. We read this literature, and found it very helpful for our research, adding it as a reference, and adding "The credibility of the main application of measurement results is determined by the credibility of the input data." to make the significance of the article even more significant.

Point 5: Line 97 - Please explain the abbreviation RBF.

Response 5: We fully agree with your valuable views. We thank reviewer for bring this advice. We added the abbreviation RBF as radial basis function.

Point 6: Line 111 - Please see the analysis in this material.

Desire L. Massart, Leonard Kaufman, Peter J. Rousseeuw, Annick Leroy, Least median of squares: a robust method for outlier and model error detection in regression and calibration, Analytica Chimica Acta, Volume 187, 1986, Pages 171-179, ISSN 0003-2670.

Response 6: We fully agree with your valuable views. We thank reviewer for bring this advice. We checked and revised similar errors and added this reference after reading it.

Point 7: Line 123 - The use of the concept of robustness requires a wide range of studies, including physical and mechanical parameters of materials. Here, however, you can also note the fast Fourier analysis as a tool for determining the value with the greatest frequency in the occurrence of serial measurements of a parameter.

Response 7: We fully agree with your valuable views. We read the relevant literature of the fast Fourier analysis, and this method can determine the value with the greatest frequency in the occurrence of serial measurements of a parameter as reviewers’ thought. We added “Rasheed, F utilized fast fourier transform to highlight the areas with high frequency change to identify outlier regions.” in line 63.

Point 8: Line 157 - You could also pay attention to some fundamental aspects of the Least Median of Squares Regression method Hodges (1967).

ANDREWS, D.F., BICKEL, P.J., HAMPEL, F.R., HUBER, P.J., ROGERS, W.H., and TUKEY, J.W. (1972), Robust Estimates of Location: Survey and Advances, Princeton, N.J.: Princeton University Press.

Response 8: We fully agree with your valuable views. We read the relevant literature of the fast Fourier analysis, and this method can determine the value with the greatest frequency in the occurrence of serial measurements of a parameter as reviewers’ thought. We added “Rasheed, F utilized fast fourier transform to highlight the areas with high frequency change to identify outlier regions.” in line 63.

Point 9: Line 157 - You could also pay attention to some fundamental aspects of the Least Median of Squares Regression method Hodges (1967). ANDREWS, D.F., BICKEL, P.J., HAMPEL, F.R., HUBER, P.J., ROGERS, W.H., and TUKEY, J.W. (1972), Robust Estimates of Location: Survey and Advances, Princeton, N.J.: Princeton University Press.

Response 9: We fully agree with your valuable views. Thank you for your advice. We added this article as a reference to the literature.

Point 10: Row 269 – From a geotechnical point of view, deformations that may be caused by the ground base of the wall and deformations of the left and right support blocks are of interest. It is unlikely that temperature deformations will have a greater contribution.

Response 10: We fully agree with your valuable views. The object of our study is concrete dam. After many previous studies, the influencing factors mainly include water level, temperature and aging factor. For the extremely high arch dam in this paper, it is easy to produce structural deformation. Temperature is a very important influencing factor, especially in the working condition of low water level and temperature decline. Therefore, temperature is the object that we should focus on, and we should consider temperature as an important influencing factor.

Point 11: Row 269 – Figure 5 does not read well, could be improved.

Response 11: We fully agree with your valuable views. Thank you for your advice. We improved the clarity of Figure 5 and enlarged the picture as required.

Point 12: Row 300 - Figure 6. You can also add the reactions of other control positions. Thus, with only one illustration, confirmation of the measurements cannot be sought. Figure needs a legend, out-of-trend points need discussion.

Response 12: We fully agree with your valuable views. We thank reviewer for bring this advice. We have added a legend in Figure 6 to make it even more complete. And we presented a discussion of out-of-trend points in line 301 to line 304.

Point 13: Row 318 - Table 4. It is not clear why the assessment of gross errors was carried out by only two of the methods. It requires clarification.

Response 13: We fully agree with your valuable views. We thank reviewer for bring this advice. We have added a legend in Figure 6 to make it even more complete. And we presented a discussion of out-of-trend points in line 301 to line 304.

Point 14: Row 318 - Table 4. It is not clear why the assessment of gross errors was carried out by only two of the methods. It requires clarification.

Response 14: We fully agree with your valuable views. We thank reviewer for bring this advice. We carried out four methods to assess the gross errors. The LMS-IGG-ELM gross errors identification method is utilized to compare with the popular DBSCAN clustering algorithm, the traditional Romanovsky criterion and the 3method to verify and analyze the effects of different methods.

Point 15: Row 323 - Table 5. It is appropriate to give another group of gross errors and the percentage of their filtering. In my opinion, a full conclusion cannot be drawn from only one result.

Response 15: We fully agree with your valuable views. Thank you for your advice. The monitoring data of crown cantilever arch dam is the key data for the deformation safety of arch dam. We selected the largest deformation, the most representative measurement point PL13-3 as the research object, the gross errors selected four typical types of shock, ridge platform, jump and step , the four gross errors removal method were examined, at the same time, it is also made to introduce gross errors at different positions and different size gross errors at the same position, the results are basically consistent with table 5, we added "We adjust the size values of these gross errors, conduct five sets of experiments, and resulting the gross errors identification effect is basically consistent, which verified the generalization and reliability of the method. " in the paper,  to make the study more reasonable and accurate.

Point 16: Row 355 - Figure 11. This comparison is interesting and highlights the upper and lower limits of the rating band. In considering formula 25, it would be interesting to take into account other elements of the unknown quantities - Mood, A.; Graybill, F.; Boes, D. (1974). Introduction to the Theory of Statistics (3rd ed.). McGraw-Hill.

Response 16: We fully agree with your valuable views. Figure 11 means that effect comparison chart of the dam validation segment after eliminating gross errors by the four methods. It combined the results of the four methods into a picture to easily compare the advantages of the used method, not for the upper and lower limits of the rating band.

Author Response File: Author Response.docx