Prediction of RC Bridge Member Resistance Decreasing in Time under Various Conditions in Slovakia
Round 1
Reviewer 1
1. Line 38: the reliability margin in equation (1) has a unit. Whereas the reliability ratio in equation (2) has no unit. Please clarify. (Also Line 76)
We agree, the units were completed and explained.
2. line 44 did the collapse occur during prestressing? Maybe you should write … in case of a prestressed pedestrian bridge….
Yes, should be “prestressed” instead of “prestressing”.
3. line 52: please check the words in the beginning of the sentence (The There…)
Yes, should be only “there are ...”.
4. line 53: please define both ends of the distance
It was completed: the distance between an initial surface of a metal and the nearest concrete surface in time.
5. line 65: serviceability Limit State (check spelling)
We agree, should be “States“ instead of “Staves“.
6. line 83: “selected” how can you select the type and rate of corrosion? Please clarify.
We agree. Type and rate of corrosion “were selected“, not “have to be selected”.
7. equation 9a and equation 9b are equal. Please clarify
Equation 9(b) was changed – it was wrong equation due to copy.
8. equation 10b: please check the units. The constant 0.0116 must have a unit in order to achieve a length of corrosion. (µA/cm² * s) ≠ m
Yes, we agree, it is valid that 1 μA/cm2 is equal to 11.6 μm/year of corrosion – it was completed.
9. line: from material point of view: yes. Is this matter also true for geometry also? (I-beam and circular rod embedded in concrete)
We have to apologize, but which line? We do not know to answer or correct information.
10. how do the models consider the spalling of concrete cover that is caused by volume increase during corrison?
The models take into account only the increase of the rust volume and the reduction of the cross-section of the reinforcement, but do not take into account the spalling of the concrete cover – it was added at the end of paragraph 2.
11. line 145: for comprehensibility it would be good to present the background of the Model at least in an annex of the publication.
We agree, but the first model (linear) was presented in [22], so not to present it again, we chose the form of reference. On the other side, the second power-linear model is presented first time.
12. line 178 to 182: is the measured chloride content valid for road bridges, where de-icing agents are used? Please clarify.
The chloride content in Slovakia is measured only in rural or industrial environments, not near roads. That is the reason why the limit values were taken from code or literature – see section 3.5.
13. line203: equation 29 is the relation between SO2 and SO2air. please check!
Yes, it should be formula (28) instead of (29).
14. line 206: how about chloride from de-icing agents?
Only chlorides calculated from input data from SHMI were taken into account in this part. The chloride from de-icing agents is considered in section 4.
15. Figure 5: why does the first-year corrosion rate change from 2004 to 2017? Does SO2 decrease? Please explain/discuss. Please also explain how chloride from de-icing agents was modelled in these figures.
The discussion is in section 3.4. Yes, the average value of SO2 decreases in time (see [24,25]) was added in figure 7 (b) to show it. The values of rcorr were calculated according to formula (28), which take into account also chloride ions.
16. line 303f: how was the increased corrosion rate chosen? Please describe.
It follows from fig. 5 . It was added into text.
17. line 314f and table 2: how were the increased values chosen? Please describe.
The values of corrosion rate rcorr= 10.12 – 38.00 (μm/year) were calculated using formula (28) and follows from concentration of chloride Cl- = 11.20 mg/(m2·day) – that value was recommended in [29].
18. line 320: please specify if Fig. 10 and Fig 11 are calculated by using equation 15.
Yes, the values were calculated using equations (15a) and (15b) - it is written in lines 288-290 (end of the first paragraph of section 4).
19. Please describe in detail: how was the protection of the reinforcement by the concrete cover modelled. With an appropriate concrete cover, the bending moment resistance should not decrease until the chloride/carbonation front reaches the reinforcement.
The protection of reinforcement by the concrete cover was not modelled – in the parametric study, for simplifying, the length of the passive stage (t0) was neglected, so it means that the reinforcement corrosion starts at the beginning (see line 321). By means of the parametric study, the main aim was to express the influence of reinforcement corrosion on possible decrease of resistance of the element – by neglecting the passive stage, extreme values were obtained.
20. Please extend the conclusions.
Conclusions were extended and some of your following comments were discussed.
21. What did your research focus on – what did you assume – what is neglected in this research.
Please discuss if your model is able to predict an increased corrosion rate caused by increased crack width due to reduced reinforcement cross section. Please discuss if the permeability of the concrete cover and its thickness is considered in the model. Please discuss: Does concrete strength fc change during service life? Is the design concrete strength lower than the real concrete strength at the bridge? Might there be additional safety (reliability) due to compressive strength that can compensate reduced bending resistance due to cross-section losses at the reinforcement?
All dimensions and material characteristics were measured on real structure using destructive or non-destructive testing. The concrete strength was also tested using Schmidt hummer. The design documentation was not available and the original values of the material strengths were also not known, so it was not possible to compare them with the measured values. Thus, only the measured geometric and material properties of the bridge were taken into account in the parametric study.
22. please check: all figures should be in high enough resolution. (fig 9 fig 3…)
It was increased.
23. I am not a native speaker, even though I do not feel able to judge language I strongly recommend to check spelling and grammar. E.g. (areas with lover first year….)
We use “less value“ instead of “lover value”.
Reviewer 2
This paper presents the decrease of structural member's strength due to steel corrosion in concrete, which is caused by chloride ion and sulfur dioxide. However, it is not suitable for publication due to the lack of originality.
The reasons are as follows. The models used are from the previous studies, especially from ISO standards. You assumed uniform corrosion of steel in concrete, even though pitting corrosion is more important for structural member's strength because of stress concentration. In actual concrete structures, uniform corrosion hardly occurs because concrete cover is not homogenous as well as environmental conditions are different due to large size of structure and shape. It is clear that high chloride ion and sulfur dioxide increase the steel corrosion in concrete from the previous studies. Overall, the results can easily be expected and the accuracy of the results is not sufficient for publication.
The literature suggests that it is recommended to use a linear model of corrosion (equations 10a and 10b) for corrosion of concrete reinforcement, which implies a derived model of RC resistance change (equation 14), as published in [22] and elsewhere. However, recent measurements show that in some cases it is better to use a power model for the corrosion of the concrete reinforcement, as recommended for structural steel (hence the literature review has also been done in this area). This is also demonstrated by our own measurements, which have been published at conferences and magazines. Power models and their parameters were derived quite a lot, but only for structural steel, not for concrete reinforcement. For this reason, the recommended model given in EN ISO 9224 [21] was chosen (equations 12a and 12b and the resulting resistance change model - equations 15a and 15b). We consider this idea (use of power model and its verification for concrete reinforcement) to be relatively new. The linear model can basically be a power model when b + ∆b≈1 (as in the paper).
The paper summarizes the authors' work over the last 2-3 years, presents the latest corrosion maps, which have not been created in Slovakia so far, and deals with the impact of the environment in Slovakia on the values of corrosion losses. As pointed out in the paper, it is also pointed out that the obtained levels of corrosion rcorr (from the measured values of aggressiveness in Slovakia) do not take into account the increased aggressive environment around the road network. As the increased aggressive environment has not yet been measured in Slovakia, literature values and standards have been taken to verify the impact of this environment on RC bridge elements. A parametric study has also shown a significant impact that has been experimentally measured and presented in [38].
The literatures suggest that it is recommended to use a linear model of corrosion (equations 10a and 10b) for corrosion of concrete reinforcement, which implies a derived model of RC resistance change (equation 14), as published in [22] and elsewhere. However, recent measurements show that in some cases it is better to use a power-linear model for the corrosion of the concrete reinforcement, as recommended for structural steel (hence the literature review has also been done in this area). This is also demonstrated by our own measurements, which have been published at conferences and magazines. Power models and their parameters were derived quite a lot, but only for structural steel, not for reinforcement. For this reason, the recommended power-linear model given in EN ISO 9224 [21] was chosen (equations 12a and 12b and the resulting resistance change model - equations 15a and 15b). We consider this idea (use of power-linear model and its verification for reinforcement) to be relatively new. The linear model can basically be a power model when b + ∆b≈1 (as it is shown in the paper).
We agree that a simpler uniform corrosion model was used instead of a pit corrosion model, but uniform corrosion was experimentally found on the presented bridge in Kolárovice [38]. The paper summarizes the authors' work over the last 2-3 years, presents the latest corrosion maps, which have not been created in Slovakia so far, and deals with the impact of the environment in Slovakia on the values of corrosion losses. As pointed out in the paper, it is also pointed out that the obtained levels of corrosion rcorr (from the measured values of aggressiveness in Slovakia) do not take into account the increased aggressive environment around the road network. As the increased aggressive environment has not yet been measured in Slovakia, literature values and standards have been taken to verify the impact of this environment on RC bridge elements. A parametric study has also shown a significant impact that has been experimentally measured and presented in [38].
Reviewer 3
1. Corrosion of steel bars depends on pH. pH of concrete is influenced of carbonation process and carbonation depends on quality, bulk density of concrete. Are these parameters implemented in model? I thing that validity of corrosion model is limited. On the other hand corrosion maps in general basis could be helpful.
The model neglects the influence of permeability of concrete, concrete cover thickness, pH factor of concrete, carbonation – these parameters were not implemented in model. It was focusen only on corrosion of reinforcement and its influence on RC member resistance change in time.
2. Line 52: Delete first word
It was deleted.
3. Figere 3 should be in English
It was changed.
Round 2
Reviewer 1
1. How did you verify that you selected the correct type of corrosion model please clarify. Of course you selected a corrosion model. But how is this model verified to be correct in this application? Did you verify it? Did someone in literature verify it?.
We agree. There were used two types of corrosion models. According to our paper, there is shown that in some cases is more correct linear model (in the case of splashing transport mechanism) and power-linear model (in the case of inland areas and spraying transport mechanism). It is hard to say which is more correct.
2. Equation 10b. The units seem still not to match. A subtraction from the diameter must have the dimension “Length”
[Diameter]=mm
[r.corr]=µA/cm²
[0.0116]=µm/year
[dt]=years
[Diameter]=2*[r.corr]*[0.0116]*[dt]
mm≠2*µA/cm²*µm/year*year=A/cm
Please carefully check!
According to literature, it is valid that 1 μA/cm2 is equal to 11.6 μm/year of corrosion. So, it means, that only icorr is in unite [μA/cm2], but full expression 0.0116.icorr is in units [μm/year]. It is empirical formula from which follows that number 0.0116 should have unit [(μm.μA/(year.cm2)], but anywhere it is not shown. Formula and its unit 10b) is correct – it was taken from literature.
3. Line 133+134 “since structural steel and reinforcement are both carbon steels (according to EN 9224 [21]), the power model can be used also for reinforcement.”
Is this true for reinforcement covered by concrete? Is the model developed on an I beam valid for a rod?
It is only assumed that it can be used because both of them are carbon steels. It was reworded.
4. You tried to clarify but: The model, does not increase the volume at all. The volume of the rod is constant. But the ratio of corroded volume to uncorroded volume changes. Please clarify that the total volume of the rod (corroded and uncorroded) does not change (which is not true for corroding steel).
The model takes into account the change of reinforcement diameter ø(t) (decreasing) in time using rcorr or icorr, which are reflected in the parameters km1 to km4. So, the volume (diameter) of rod is changed in time – decreased.
5. Line 345) you did neglect protection of the reinforcement due to concrete cover. Protection of concrete cover is one main concepts of reinforced concrete. Proper design according to EN 1992-2 achieves more than 5-15 years passive stage of the reinforcement due to concrete cover. At least in most European countries it does. Please therefore add to the abstract and the conclusion (near line 381) that your study neglects the effect of concrete cover and therefore models the degradation process for unprotected bending reinforcement
We agree with you – protection of concrete structures is mainly assured by concrete cover. So due to improve quality of paper, we did and filled up also parametric study with chosen (not calculated) lengths of passive stages (10-40 years) – see figs 12, and 13. Moreover, also influence of pitting corrosion was included for comparison (fig. 14 and texts).
6. The resolution of the figures can still be improved (especially fig 9 and fig 3) or the figures can be redrawen
Fig 9 was improved, Fig. 3 was downloaded from the project ICP materials [26]. We do not know to improve the picture. If necessary, we can delete the Fig. 3 and renumber the following Figures.
Reviewer 2
I think that it is not suitable for publication due to the lack of originality and accuracy.
This paper assumed uniform corrosion, which means that corrosion of steel is uniform on the steel surface. However, it is not consistent with real reinforcement in concrete. The carbonation of concrete starts from the concrete surface and grows into the inside. Therefore, corrosion starts from the reinforcement surface close to the concrete surface and the rest parts of reinforcements surface acts as cathodes. This makes significantly different results of the analysis.
In conclusion part of the paper, it is clear that high chloride ion and sulfur dioxide increase the steel corrosion in concrete from the previous studies. Although the measurement results were used for two to three years, the analysis results do not seem to be much different from the existing research results. Therefore, it is judged that the originality of this study result is insufficient.
Overall, this paper is not suitable for publication
We agree that pitting corrosion is more aggressive and has higher influence as uniform corrosion. For this reason, the pitting corrosion was also included in parametric studies – see text and Fig. 14. Also, the influence of the passive stage was included to get more relevant results for comparison.
According to our experimental results from the real bridge in Kolarovice (mentioned in the paper). The pitting corrosion starts first due to chloride ions, but after the time (splashing mechanism is seen on the bridge) and after spalling of concrete cover, it was changed into uniform corrosion seen on the bridge. It was the reason for considering also the uniform corrosion. Moreover, the real corrosion rates measured on the bridge (from 2005-2018) were confirmed values used in the parametric study – see paragraph after Table 2).
The article presents our own measurements and research work (new corrosion maps, which have not been created in Slovakia so far) and new information suitably applied to already known procedures and models (comparison of the linear model and power model according to EN ISO 9224). The application of corrosion losses verified by experimental measurement on real bridge construction and taking into account three levels of corrosion (inland areas with minimum influence of corrosion, spraying mechanism and splashing mechanism) was not found in the literature in this form.
The other two opponents are of the opinion that the paper is suitable for publication (after revision of course).
Thank you for your time spent with the review, and we again hope to convince you to change your mind about the ability to publish the paper.
