Artificial Neural Network for Vertical Displacement Prediction of a Bridge from Strains (Part 2): Optimization of Strain-Measurement Points by a Genetic Algorithm under Dynamic Loading

Round 1

Reviewer 1 Report

 

The manuscript involves the objective to build a method to optimize the number and location of strain-measurement points using the GA. The paper is generally well-organized and its structure is well-designed.

However, the method using ANN and GA are not innovative. Thus, Section 2.1 and 2.2 just illustrating general description of these traditional techniques can be reduced or removed. This paper also includes several parts just explaining general descriptions of several terminology and techniques. They should be reduced.

Datasets used in this study, 2083 for training and 201 for tests are sufficient to assist reliable outcomes. In addition, the case study data are sufficient and reliable.

This paper seems a cast study paper using the traditional ANN and GA method for optimizing the number and location of strain-measurement points. This reviewer cannot find any new intellectual findings of this methodology section.

Thus, the authors would need to emphasize the given case study to show the implications of the proposed method. In addition, this paper would need to conduct an extensive literature review for showing what contributions and intellectual merits this paper provides.

For explaining the adopted GA method for this study, the authors need to explain how Selection Operator, Crossover Operator, and Mutation Operator are used and defined in this study. The authors can find some information in Lines 138-140 but cannot find the Mutation Operator in this study, which is a key operator inserting random genes in offspring.  

The conclusion part needs more elaboration. This reviewer is hard to find out any new and innovative thoughts in the conclusion section. This manuscript requires more insightful conclusions to be published.  

 

 

Author Response

Thank you for your review

Based on your comments, We revised the manuscript.

Please review our revised manuscript.

Thank you

Sincerely

 

 

Author Response File: Author Response.pdf

Reviewer 2 Report

This manuscript is an extension of the previous Part1 [18]. The reviewer can see a very good coherence throughout the research work. Detailed comments can be found below.

Abstract needs to be improved. (1) Too many “strong ability” phrases, this has to be replaced. (2) What is GA is not clear, it is not suitable to abbreviate terminologies in the abstract.

Line 45 – “various studies have been conducted to measure bridge displacements.” à” through various direct measurement approaches”.

The methods of using ABAQUS and MATLAB for simulation and optimization are not clear. It has to be more specific (e.g. boundary conditions, models, etc.).

Figure 2 is not centered.

Line 155-159 is a repetition of the previous content. Please either delete or modify it.

Page 10 (figures, table and text) is similar to the previous work in [18]. Please either delete or modify it to avoid repetition.

What is the cost of having more sensors on the infrastructure? This has to be detailed. Many researches have been done to explore the approaches of having a large number of sensors for structural health monitoring.

Glišić, Branko, et al. "Strain sensing sheets for structural health monitoring based on large-area electronics and integrated circuits." Proceedings of the IEEE 104.8 (2016): 1513-1528.

Chen, Xiyuan, et al. "Characterization of Distributed Microfabricated Strain Gauges on Stretchable Sensor Networks for Structural Applications." Sensors 18.10 (2018): 3260.

Please cite these references to answer the question from another perspective.

The reviewer understand that the verification of the proposed method applied to an actual bridge is listed as a future work. However, the verification and validation of the proposed method applied to a simple beam should not be a hard experiment. Please include that in company with the numerical analysis for better confidence of the proposed method achieved by taking advantages of the ANN and GA algorithms.

Author Response

Thank you for your review

Based on your comments, We revised the manuscript.

Please review our revised manuscript.

Thank you

Sincerely

Author Response File: Author Response.pdf

Round 2

Reviewer 2 Report

The reviewer can see great improvements made by the authors. However, there are still minor changes that need to be addressed before publication.

Firstly, there are many grammatical errors in the abstract. For example, "a laboratory experiment is 24 carried out with the vibrating cantilever beam", "the" should be changed to "a", "optimized locations" --> "the optimized locations", "good agreement" --> "good agreements", "measuring" --> "measurements".

Besides, the newly added section 5 also has many problems. "The section of beam is 50x5 mm rectangular shape." "section" --> "cross-section", "rectangular shape" --> "rectangle". Line 429, "of beam end part" --> "at the end of the beam", "occurred" --> "introduced". Line 430, "was" --> "were". Line 432, "were occurred" --> "were found/measured" or "occurred". Line 433, "with 200 Hz" --> "with 200 Hz sampling frequency”. Line 434, "1020mm" --> "1020 mm" missing a blank space between the number and the unit. Line 439, "for two number and the optimized locations were 100 mm and 700 mm" --> "to be 100 mm and 700 mm". Line 441, “optimized locations" --> "the optimized locations". Line 444, "200Hz" --> "200 Hz".

Please also reformat the section 5 to make the text justified to be in accordance with other texts.

Author Response

Thank you for your review.

 

Based on your review, 

1) We changed grammatical errors in this paper based on your comments

2) We reformatted the section 5 to make the text justified to be in accordance

   with other texts.

 

Please, review my manuscripts.

Thank you.

Sincerely