Statistical Treatments of Chloride Threshold and Corrosion Propagation Rate

Round 1

Reviewer 1 Report

 This paper studied statistical treatments of chloride threshold and corrosion propagation rate, aiming into the need of an engineering judgement in function of the quality of the concrete and the exposure class. However, there are some problems in this paper, including lack of evidence and obvious errors. It is recommended that this manuscript is possibly considered after making minor revision taking into account the following remarks. Some other errors are also necessary to be corrected.  

1. I cannot see the highlights of the abstract in the present form. Please revise it to provide some results.

2. Figure 1 is difficult to distinguish, please improve the quality and revise the superscript and subscript .

3. What is the meaning of Fig.5? It there any connection of the carbonated structure to chloride threshold? Please provide reasons.

4. In section 3.2.2, the relation between coefficient of variation and relative humidity/ temperature is not evident. Please provide evidence or relevant references.

5. The subscript in Fig. 6, 8 and 9 should be revised. The errors in captions of Fig. 8 and Fig. 9 should be modified.

6. There are too many data in the conclusions, but still unclear. The comparison between laboratory and real structures should be added. Please modify it.

7. There are many errors in this manuscript. Please check carefully throughout the paper.

8. The format of references are in a mass. The authors should take it seriously.  

Comments for author File: Comments.docx

Author Response

 This paper studied statistical treatments of chloride threshold and corrosion propagation rate, aiming into the need of an engineering judgement in function of the quality of the concrete and the exposure class. However, there are some problems in this paper, including lack of evidence and obvious errors. It is recommended that this manuscript is possibly considered after making minor revision taking into account the following remarks. Some other errors are also necessary to be corrected.
Following the comments of the reviewer 2 I have changed completely the order of the chapters. I am sorry for the inconveniances.  
1.    I cannot see the highlights of the abstract in the present form. Please revise it to provide some results.
I do not understand which are the highlights of the abstract. 
2.    Figure 1 is difficult to distinguish, please improve the quality and revise the superscript and subscript 
Original figure has been introduced. The data are difficult to eb distinguished because they overlap. Only total chlorides is now represented to avoid the overlapping although is less illustrative
3.    What is the meaning of Fig.5? It there any connection of the carbonated structure to chloride threshold? Please provide reasons.
The treatment of the corrosion rate is independent of the chloride threshold in the sense that there are treated chloride contaminated and carbonated structures. All kind of strictures were considered for the statistical treatment.
4.    In section 3.2.2, the relation between coefficient of variation and relative humidity/ temperature is not evident. Please provide evidence or relevant references.
I do not understand on why is not evident. The CoV have been calculated from the values of figure 7 (now figure 2). The evidence is there in figure 7 (or now 2). These are the data in which mean and CoV was calculated.
5.    The subscript in Fig. 6, 8 and 9 should be revised. The errors in captions of Fig. 8 and Fig. 9 should be modified.
I have changed a bit the caption of figure 2 (present version). The other seem to me correct. I ask the referee to eb more precise
6.    There are too many data in the conclusions, but still unclear. The comparison between laboratory and real structures should be added. Please modify it.
There are several perspectives in the conclusions. The comparison between laboratory and site has been added as a general conclusion
7.    There are many errors in this manuscript. Please check carefully throughout the paper.
I have done my best to correct errors
8.    The format of references are in a mass. The authors should take it seriously
I do not understand “are in a mass”. In any case I have tried to fit into the rules although now it is very easy to find the references in internet and too many details are not needed. First author and year is enough. It takes a lot of time to be rigorous with the format. 

Author Response File: Author Response.pdf

Reviewer 2 Report

The manuscript by C. Andrade and D. Izquierdo describes the effect of choride concentrations on corrosion rates in concrete. The manuscript contains a good review of previous literature but before it should be considered for publication, I recommend the autothrs to adress the follwing points:

(1) the English should be improved, the manuscript contains a large number of grammatical errors (e.g. line 16: In all both ... --> In both ...; line 17: decades of differences --> differs by several order of mangitude ...) and typographical mistakes.

(2) more details are needed on the experimental procedure Rilem Recommendations (main manuscript text) / RILEM Recommendations (references). The reader should be able to reproduce the experimental results by reading this manuscript and not by getting referred to other works.

(3) The novelty of the manuscript is not clear to me. Is this a review article? Is the statistical evaluation the novel contribution? This confusion may be partially due to the fact that the authors mix introduction and results/discussions. For instance the first two paragraphs in section 3.1.1. should go to the inntroduction (stating existing knowledge from literature) and not into the results section which is about the presentation of own results and discussing these in the existing research framework.

(4) Figure formatting is different for all figures and it is suggestd the authors use one consistent formatting style throughout their manuscript. Remove figure titles as these should be part of the accurate description in the figure captions

(5) Figures (and tables) are to be treated as stand-alone objects! This means that the references must be understood by the reader without looking for it in the manuscript text. Just as one example (out of many) - it is unclear which reference is meant with D. Izquierdo et.al. 2004.

(6) line 135: what is meant with "Cl_" ?

(7) log-normal distributions are used for events that are occurring on a logarithmic (time-)scale. What is the statistical approach to this fit? For corrosion, I would have expected distributions such as Weibull, Gumbel or in rarer occasions Gamma-distributions

(8) section 3.2 plenty of typographical errors and missing spaces between words , please do correct these

(9) line 200: please remove the apostrophes in spatially (or remove the entire word since "homogeneous" indicated a spatial distribution

(10) line 205: please remove the apostrophes around temporal

(11) figures 4 and 5: please add references to this graph

(12) ilne 263: "teh epriod" --> the period

(13) figure 9: the text in the figure caption should be improved as it is hard to understand for the reader what the authors are referring to

(14) line 320: remove the "Please add:"

(15) reference 9 is incomplete and missing page umbers

(16) reference 28 and other onlien references need a date when the link was accessed since this is referring to a dynamic content

(17) there are a lot of self-citations by the authors; agreed that it may be hard to find reliabale literature, the authors are nevertheless encouraged to give credit to other groups that worked on a rather common issue of chloride induced corrosion in concrete

(18) the purpose of the results in figure 9 is not clear to me; if this figure is indeed realted to the statistical fit and finding the appropriate distribution to the cumulative probability of individual measurement values, one might expect a standardised stochastic framework to identify the quality of fit --> for instance a Smirnov or Kolmogorov-Smirnov test instead a general statement on "Bayseian statistics". It is also not clear as to how the authors have derived this fit. This is especially important because it seems to represent the novelty of this work!

(19) figure 1: can you please comment of  the local pH effect for the different potentials?

Author Response

The manuscript by C. Andrade and D. Izquierdo describes the effect of chloride concentrations on corrosion rates in concrete. The manuscript contains a good review of previous literature but before it should be considered for publication, I recommend the authors to adress the follwing points:
Thanks. 
(1)    the English should be improved, the manuscript contains a large number of grammatical errors (e.g. line 16: In all both ... --> In both ...; line 17: decades of differences --> differs by several order of mangitude ...) and typographical mistakes.
I did my best for correcting errors. Regarding the English,  tried to improve.
(2)    more details are needed on the experimental procedure Rilem Recommendations (main manuscript text) / RILEM Recommendations (references). The reader should be able to reproduce the experimental results by reading this manuscript and not by getting referred to other works.
I have introduced some more details but to explain for example how chemical analysis of chlorides is described I think that extends the paper and is very known.
(3)    The novelty of the manuscript is not clear to me. Is this a review article? Is the statistical evaluation the novel contribution? This confusion may be partially due to the fact that the authors mix introduction and results/discussions. For instance, the first two paragraphs in section 3.1.1. should go to the introduction (stating existing knowledge from literature) and not into the results section which is about the presentation of own results and discussing these in the existing research framework.
There are several novelties although the statistical characterization is the most evident. Also, I have elaborated further previous conclusions on chloride threshold as there are recent papers suggesting that the variability introduces confusion. In my opinion the variability is normal and should be understood.
I have changed the structure of the paper following the classic of Introduction, Materials. Results, Discussion. I did a “thematic” structure separating chloride threshold and corrosion rate, but I realized that this has not been correctly understood and then I have changed the structure and, as an example, I have moved all the comments on bibliography to the introduction.
(4)    Figure formatting is different for all figures and it is suggested the authors use one consistent formatting style throughout their manuscript. Remove figure titles as these should be part of the accurate description in the figure captions
Thanks made
(5)    Figures (and tables) are to be treated as stand-alone objects! This means that the references must be understood by the reader without looking for it in the manuscript text. Just as one example (out of many) - it is unclear which reference is meant with D. Izquierdo et.al. 2004.
(6)    I do not understand completely the suggestion, and I did not find in any caption “D. Izquierdo et.al. 2004”. In the captions I used only the normal manner for the references
(7)    line 135: what is meant with "Cl_" ?
It was an error thanks. It is Cl-
(8)    log-normal distributions are used for events that are occurring on a logarithmic (time-)scale. What is the statistical approach to this fit? For corrosion, I would have expected distributions such as Weibull, Gumbel or in rarer occasions Gamma-distributions
Log-normal distributions are those suggested to avoid negative values when making the probabilistic calculations. In any case here I did not fit the data in any distribution. This was made before by each author or in [15] by us.
The distribution of corrosion is not fitted. It is the natural trend it appeared from the direct representation of the data. It seems LN but I did not yet the fitting with a program.
(9)    section 3.2 plenty of typographical errors and missing spaces between words , please do correct these
Sorry but I did not find many typographical errors in that section
(10)    line 200: please remove the apostrophes in spatially (or remove the entire word since "homogeneous" indicated a spatial distribution
I have maintained spatially because homogeneous is not enough as homogeneous can be referred to the evolution in time. Sorry but it seems to em better as it is.
(11)    line 205: please remove the apostrophes around temporal
Sorry but I left because it si a manner to emphasize the different  variations.
(12)    figures 4 and 5: please add references to this graph
They have not references as they are bnew in this paper. This is the first time I publish these data
(13)    ilne 263: "teh epriod" --> the period
thanks corrected

(14)    figure 9: the text in the figure caption should be improved as it is hard to understand for the reader what the authors are referring to
The figure is the same than the figure 6 but removing those distributions that seem less representatives. One distribution is from carbonated structures and the other from chloride contaminated ones. It is remarkable the closer they are in spite of have been built from different structures. It is added the average value found.
(15)    line 320: remove the "Please add:"
Thanks, made
(16)    reference 9 is incomplete and missing page umbers
I did not find the pages in the journal. I added the DOI
(17)    reference 28 and other onlien references need a date when the link was accessed since this is referring to a dynamic content
The reference 28 has a date. I tried to complete as much as I could
(18)    there are a lot of self-citations by the authors; agreed that it may be hard to find reliabale literature, the authors are nevertheless encouraged to give credit to other groups that worked on a rather common issue of chloride induced corrosion in concrete
There are 10 of 31 references. The problem is that the subjects is not treated very much by other authors (statistical treatment of data). For example, there are not corrosion rate values apart from ours in the literature.
(19)    the purpose of the results in figure 9 is not clear to me; if this figure is indeed related to the statistical fit and finding the appropriate distribution to the cumulative probability of individual measurement values, one might expect a standardised stochastic framework to identify the quality of fit --> for instance a Smirnov or Kolmogorov-Smirnov test instead a general statement on "Bayesian statistics". It is also not clear as to how the authors have derived this fit. This is especially important because it seems to represent the novelty of this work!
The representation is directly the values. We are not interested in this moment to verify which kind of statistical distribution fits batter. This is not the purpose. On this first step is to show what we have obtained. Nobody can do similar plotting. We are the only ones having the data. How to use it: by making a Bayesian updating when new values are obtained, but I removed this possibility because the referee sems not be happy with the Bayesian updating. Let say by now that this is the straightforward representation of the results. It is a fact. No more by now. 
(20)    figure 1: can you please comment of  the local pH effect for the different potentials?
In the paper [14] and in several other papers we have given the probability not only as total chloride but also a free chloride and as Cl/OH. The role of pH is related to the threshold but measuring pH in a concrete is difficult and the Cl/OH ratio is neither used in practice nor specified in the codes. 
I thank the referee for the interesting suggestions and the detection of errors. I fully change the structure of the paper to follow the suggestions made.

 

Author Response File: Author Response.pdf