A TECDOC on Seismic Soil Structure Interaction – IAEA draft TECDOC (Chapter 1-7 of the R507 Version, 15. Dec. 2016))
COMMENTS BY REVIEWERReviewer: Philippe Renault
Country/Organization: Switzerland / swissnuclear Date: 8.2.2017 / RESOLUTION
JJJohnson /
Comment No. / Page/Section / Comment / Accepted / Accepted, but modified as follows / Rejected / Reason for modification/rejection /
1 / 15 / 1.3 / Within the scope it should be added/clarified that slope stability is excluded from the scope of the document. / X / Agree with Alain’s proposed addition: “Any earthquake induced hazard like liquefaction, slope instability is excluded from the TECDOC.”
ü2 / 18 / 1.3.3 / The TecDoc mentions the “hardened-core components / noyeau dur” which is not clear and common to non-French readers. It is not clear if this expression or concept is relevant to mention in this SSI TecDoc. If deemed to be relevant, then a reference should be provided or some more technical background explanations. / X / Agree with Alain’s reference to IRSN definition.
3 / 18 / 1.3.3 / The second paragraph mentions at the end “high energy sources”. The correct expression is “source term”. / X / Technically, it is a correct statement, since non-reactor nuclear facilities do not have high energy sources, so the consequences of a failure are generally much less than for a reactor. That is one of the arguments for DOE facilities to be Seismic Design Category less than 5. However, two paragraphs, one of which included the statement concerning “high energy source” were deleted by Boris.
4 / 20 / 2.1 / Add an explanatory, illustrative generic sketch as figure to help the reader to clarify and distinguish: within vs. outcrop motion, control point, top of grade vs. bedrock, … / X / Figure 6-1 provides the requested visual explanation (AP response – JJJ agree)
5 / various / Check for consistency of terminology: “excitation level” should be replaced by “strain level” where appropriate, in order to be explicit and clear. / X / Searched the Word Version 509; found “excitation level” only one time on p. 115 and “excitation level” is correct in this context.
6 / 22 / 2.3-2-7 / The chapters 2.3 to 2.6 are going to be merged and restructured which is very much welcome. Chapter 2.7 on Uncertainties highlights a very important aspect. As the site response has an interface to the SHA this aspect should be introduced here with the goal to point the reader to the potential issues related to the purpose of the analysis vs. the consistency of the various pieces. Especially, the issue of consistency of the bedrock / top of grade UHS (originating usually from a PSHA) and the engineering approach to evaluate response spectra for soil surface often based on a deterministic concept. In practice the consistent consideration of all the uncertainties of the soil in SSI is simplified for the sake of efficiency. Nevertheless, the resulting top of grade response spectra between a fully probabilistic based SHA up to the surface and a response spectra evaluated up to rock and with added SSI are not identical. A paragraph on the consistency should be added, as from a regulatory aspect it is unsatisfying to not have the same results and the need to always ask for justification why the simplified approach is still representative. Or to be put in Chapter 6. / X / Rewritten, included in Version 509_JJJ, and a separate Word file. Still need to add further discussion on double-counting and consistency of hazard definition as different locations in the site profile.
7 / 28-44 / 3.1-3.5 / The chapters 3.1 to 3.5 are very academic and exhaustive. For the sake of the total length of the document it is recommended to shorten those chapters more. The discussed topics are all relevant, but maybe citing more literature references can help to shorten without sacrificing the scientific completeness. / X / I concur with AP, this is valuable to have in one place and it is not burdensome.
8 / 41 / 3.3.2 / The paragraph to be written about the “calibration for site response” should pick up that the calibration is based on site measurements. / X / To be added.
9 / 45 / 3.5.1 / As introduction and basic motivation for the site instrumentation, it should be explicitly stated that the recommended ADDITIONAL instrumentation is not an operational issue and only conceived to collect data to refine models in the mid- to long-term. This aspect is not clear and national regulators might easily misuse this chapter to require more instrumentation which might also be connected to the seismic alarm system. But the two are two different systems which serve two different purposes and the additional instruments should be optional to improve the models in order to reduce uncertainty. / X / I concur with AP and his addition. Here is a slightly changed version: Site instrumentation, as described in this section, is in addition to seismic instrumentation installed at the site to record the level of ground motion and in-structure response due to an earthquake occurrence. The basic motivation for this additional instrumentation is to record very useful and important information to assess potential site effects and their modelling, and, consequently, reduce modeling uncertainty. Two different types of instrumentation may be implemented based on passive measurements of ambient vibrations or active measurements of seismic events.
10 / 46 / 3.5.2 / Use for the PEGASOS Refinement Project the 2013 reference instead of the 2009:
Swissnuclear (2013). Probabilistic Seismic Hazard Analysis for Swiss Nuclear Power Plant Sites - PEGASOS Refinement Project. Final Report, Vol. 1-5, http://www.swissnuclear.ch/de/downloads/. / X
11 / 46 / 3.5.3 / Add also density measurement / X / I concur with AP response.
12 / 55 / 3.8 / The practical aspect how to “interpolate” several measurements distributed over the site area to come up with an “average” (material, soil profile) is very important and should be discussed in form specific recommendations. / X / To be discussed.
13 / 56-62 / 4 / The whole chapter should be revised and discuss the interface to SHA instead of describing the theory of SHA. A reference to SSG-9 for the background of SHA should be sufficient. Within the interface discussion emphasis needs to be put on the potential double-counting of uncertainties, which is still a problem and not very well addressed in literature.
Table 4-1 and 4-2 should be kept. Remark: In Table 4-2 in the first row the frequency of 100 Hz or PGA should be added, respectively. / X / Rewritten, included in Version 509_JJJ, and a separate Word file.
There was no Table 4-2 in Versions 504 or 509 that I could find – I interpreted Philippe’s comment to mean existing Figure 4-1, which I retained. I added an additional table from SSG-9..
14 / 61 / 4.5 / The text of this section is redundant / a duplicate of the second paragraph of Section 4.1. / X / Rewritten.
15 / 67 / 5.2 / Proposal for reformulation: of the first paragraph: (see text below the table) / X
16 / 72 / 5.3.2 / The formulation “show great promise” is ambiguous. Or it is research state and not yet applicable for industry application or it is mature. A sentence should be added to make this distinction if the TecDoc is already recommending to use those models/codes or not. Suggestion is the latter, but the current formulation is too open. / I concur with BJ’s modification: “Free field ground motions obtained using large scale regional models have been validated (Taborda and Bielak, 2013; Dreger et al., 2015; Rodgers et al., 2008; Aagaard et al., 2010; Pitarka et al., 2013, 2015, 2016) and are currently used to develop seismic free field motions for a number of large scale regions in the USA, mostly on the west coast.”
17 / 96 / 6.3.1 / Sub-section on Convolution: Add 1 Hz to the discrete frequencies.
Add “in log-log space” to the sentence “…are connected by segmented lines IN LOG-LOG…” / X / Both additions made to Version 509.
18 / 97 / 6.3.1 / Reformulate: - Select ground motion response spectra from data bases representing the (M,R);
In
Select ground motion response spectra from data bases representing the shape of the UHS / response spectrum;
Because M, R are secondary parameters to achieve the same goal: get the correct and representative earthquake for the given conditions. The shape of the spectrum is the target and not the M and R at the origin of it. / X / Not what is being done in practice.
19 / 99 / 6.3.2 / A more specific recommendation / alternative for “other APPROPRIATE methods need to be used” is to increase the standard deviation of the generic profile in order to capture/envelope the site-specific features. / X / Added in Version 509: “One approximate way of dealing with this situation is to increase the variability (standard deviation) of the soil properties of the soil profile for site response analysis and SSI analysis.”
20 / 101-102 / 6.3.3 / The 3D methods have still an issue with correct representation of high frequencies (>10 Hz) or are not validate above that range. This note should be added in order to point out today’s limitations of those methods/codes. / X / 1. I agree with BJs addition: “In addition, it should be noted, that due to computational requirements, large scale regional models are usually restricted to lower frequencies (below 3Hz) while there are current projects (US-DOE) that will extend simulations to 10Hz, for very large regions (200km x 150km x 4km). Another problem is that seismic source, fault slip models currently cannot produce high frequency motions, and stochastic high frequency motions need to be introduced.
2. However, we need to recognize that the phrase “have shown great promise” appears in the same paragraph – should be changed as in Philippe Renault’s Comment 16.
21 / 102 / 6.3.3 / The sub-section on “3D versus 1D Seismic Models” lacks of clear guidance and is problematic in the sense that it “invalidates” past practice and from the practical point of view this is legally an issue for existing approved evaluations. / X / Re-structure.
22 / 103 / 6.3.3 / Last paragraph before 6.4: “1D waves is not appropriate”. Same comment as in #21: Guidance instead and invalidation of past. / X / Re-structure.
23 / 109 / 6.5 / Three approaches are used in practice today: Add as third – Generation of synthetic, simulation based time histories for the given source and site conditions. The synthetic do not need any scaling or spectral matching. / X / Added “Three” approaches; added third bullet as stated. (group discussion needed)
24 / 110 / 6.5 / Conditions for the three spatial components: Add for clarity in the second bullet: STRONG MOTION duration of the … / X / Added two places. Done.
25 / 111 / 6.5 / Last bullet on the pager about the statistical independency: This bullet should be removed as the approach is outdated in practice (even if still present in US-NRC RegGuide 800). Real records don’t show the specified values of the correlation coefficients. R. Kennedy supports this point of view. / X / Difficult to remove, because it is in U.S. NRC SRP Section 3.7.1 (Rev. 4, December 2014): When time histories are used, each of the three ground motion time histories should be shown to be statistically independent from the others. Each pair of time histories are considered to be statistically independent if the absolute value of their correlation coefficient does not exceed 0.16.
and in ASCE 4-16, Section 2.6.2: “…. The three (orthogonal) components in a ground motion history shall be statistically independent. The correlation between any two acceleration time series is defined as the absolute value of ρ12 computed by Eq. (2-3). A suite of ground motion histories shall be considered statistically independent when the mean correlation of the set is no greater than 0.16, with no single correlation of any two records greater than 0.3 (Hadjian, 1981; Hadjian, 1984; Huang et al., 2011).
26 / 112 / 6.7 / Delete “and confidence levels” in the section title, as there is no link to uncertainties in the text below and the context is not clear. / X / Done
27 / 119-126 / 7.2.1 / The sections below “Finite Elements” until the end (before 7.3) should be significantly shortened for the sake of showing the SSI specific issues and recommendations / X / Agree. Only highlighted in Version 509 for reminder.
28 / 168 / 7.4.13 / Add/mention consistent infinitesimal finite-element cell method (CIFECM) (Wolf and Song 1996, Emani and Maheshwari 2009) and scaled boundary finite element method (SBFEM) (Wolf 2003) as simplified models, but more sophisticated than Winkler springs. / X / I agree with BJ’s addition: There are a number of other methods that have been used to model SSI phenomena, that vary in sophistication and complexity. We note infinitesimal finite-element cell method (CIFECM) (Wolf and Song 1996, Emani and Maheshwari 2009) and scaled boundary finite element method (SBFEM) (Wolf 2003).
29 / 176-177 / 7.6.2 / Under the section “Example results”, please remove/anonymize the name of the NPP site (=Mühleberg) in the text and in the figure captions. There is no added value in explicitly mentioning the site name. / X / These results are from a publication that identifies the Mühleberg site and NPP.
30 / General / Check for citation of literature references and remove those which are not cited in the text
Check orthograph again and especially consistence of spelling technical terms across the different chapters (coming from different authors) in order to harmonize. / X
Comment 15 – Text proposal:
There are four basic approaches that are used to develop ground motion models: empirical GMPEs, point source stochastic simulations, finite-fault simulations (FFS), and the hybrid empirical method (HEM).