New Perspectives on Excavation Disturbance Zones: Main Driving Forces

Round 1

Reviewer 1 Report

1) In the numerical simulation, the width and height of the model are both 66 m, and the 66 m rock layer is regarded as a same rock material. That seems unrealistic.

2) In fact, the crustal stress should be applied first, and then the tunnel should be excavated. In other words, whether ignoring the mining impact will have a significant impact on this study?

3) How to introduce environmental factors into the model, and what is the action mechanism of environmental factors?

4) Why is the stress condition of the model inconsistent with that of the actual situation of the on-site?

5) Many contours of numerical results are blurry, which makes the readers difficult to understand.

6) The phase heat in the humidity is suggested to take into account in the numerical model.

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Reviewer 2 Report

The subject is an interesting analysis of an applied and complex engineering problem. The theoretical background of the proposed model is consistently founded and has been conveniently applied to check its suitability based on the three working conditions of the exemplary cases handled with the help of the self-developed RFAP code. The different failure types are included in the code. The introduction as a background or state of the art of the subject is convincingly sustained with the extensive and suitable bibliography. The conclusions are justified and confirm the expected influence of environmental factors.

The authors come to the logical expectation that the damage process, as it applies to the failure or damage degradation of quasi-brittle materials, depends on the environmental factors, whose influence is reasonably reproduced

My only concern is twofold:

First: Although the problem is too complex to request already the application of probabilistic considerations, a comment about this issue should be brought in the paper for a possible future improvement of the model. The numerical analysis may provide an estimation of the damage evolution, once the model parameters are estimated, but it does not imply a probabilistic prediction of the damage evolution. The numerical model is based on several material parameters (see Table 1), whose estimation must be based on exploration, supposedly, with test drilling and, obviously, on the experience of the engineers’ team. Some of them may be more or less acceptably deterministic, others perhaps not. This entails hazard due to the variability and uncertainty associated with the properties of the rock as a non-uniform, quasi-brittle material, including the unexpected presence of cracks and faults. How does this affect the expected damage evolution of EDZs? and how could be this included in the RFPA code? This seems to be of utmost relevance in the predictive planning of the site in order to ensure hazardless forecasting of the damage process and to avoid unexpected costs.

Two: Damage related to time and environmental effects, as implied in the EDZ evolution, entails great complexity. I wonder if the consideration of a parallel, complementary methodology based on the phenomenological way of envisaging the damage evolution would be applicable. In other engineering failure degradation problems, damage related to time effects proves to be conveniently handled using sample functions, in the sense as proposed by Bogdanoff-Kozin [*] to capture the stochastic cumulative damage process.  A certain driving force (as could be stresses or more complex functions of stresses and strains) is selected as the main variable as a function of time, while temperature and humidity are considered the secondary variables: The latter, hopefully, may influence the damage evolution quantitatively but not qualitatively. In fact, the sigmoidal shape of the damage function evolution is expected to remain unaltered (see Figure 2). By the way: How is evolving the plotted function in Figure 2 after 90 days? Would it show a trend to a sigmoidal shape similar to the creep function plotted in Figure 1? If yes, this would reinforce the interest in the use of the sample functions approach.

The positive results achieved with the application of sample functions on damage evolution in concrete, for instance, prove this being a promising methodology to be also applied to the solution of the EDZ evolution in rocks.

Typos

Line 767: Replace “form” by “from”

Line 873: Replace “investingation” by “investigation”

 

[*] Bogdanoog J.L., Kozin F. Probabilistic models of cumulative damage, Wiley, 1985.

 

 

 

 

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Reviewer 3 Report

I have read your manuscript carefully, which deals with an interesting subject. It is generally well-written, and the methodology adopted is described.

This reviewer does not understand how the numerical model for EDZ was validated with real data, respectively, using reference [8] for humidity in a headrace tunnel and reference [36] for the temperature at a nuclear waste repository. These study cases are tunnels in very different rock masses according to parameters of Tables 1 and 2 but nothing is mentioned on the type of rock mass (e.g., are there discontinuities on the first one to explain how humidity degrades rock mass stability so quickly – 150 days) as well as the overburden for both cases because there are some statements [e.g., line 612 “with the exception of the stress factors”; lines 642/3: “the ratio of horizontal to vertical geostress was consistent with the actual on-site conditions (i.e., the lateral pressure coefficients of the two were the same)”] that are not very clear.

Perhaps conclusions should be reviewed in the way they are written (e.g., lines 623/5; 644/5) emphasizing that these are numerical models tested for specific conditions. 

Some minor remarks:

- Depth of the damage zone: do you mean thickness?

- Lines 68, 86, 101: c.f. usage of “several authors” (none, one or two references);

- Lines 105/6: please clarify;

- Lines 231/2: parameters letters are not in italic;

- Lines 472/8 - Same conditions (300/9 & Figure) 5 for both models, so why do you repeat?

Author Response

Please see the attachment.

Author Response File: Author Response.docx