ML20056E981
| ML20056E981 | |
| Person / Time | |
|---|---|
| Issue date: | 08/19/1993 |
| From: | Thoma J NRC OFFICE OF NUCLEAR MATERIAL SAFETY & SAFEGUARDS (NMSS) |
| To: | Rowland T ENERGY, DEPT. OF |
| References | |
| REF-PROJ-M-32 NUDOCS 9308250374 | |
| Download: ML20056E981 (8) | |
Text
- _ _ _ _ _ _ _ _ _._ ____ ____ _ __ _______ __ ______ -__ ___
1 1
Project M-32 AUG 191993 Mr. Thomas P. Rowland, Director West Valley Project Office U.S. Department of Energy P.O. Box 191 West Valley, NY 14171
Dear Mr. Rc41and:
SUBJECT:
REQUEST FOR ADDITIONAL INFORMATION FOR " WASTE FORM QUALIFICATION PROGRAM FOR CEMENT SOLIDIFICATION OF SLUDGE WASH LIQUID" CONCERNING SOLIDIFICATION WITH TYPE I PORTLAND CEMENT.
The Nuclear Regulatory Commission has reviewed the subject qualification program (QP) notebook, submitted by the U.S. Department of Energy (DOE) in various components between 1991 and 1993. The technical review included information contained in Volumes 1 and 2 of the QP notebook, as well as additional information that was attained as part of the review process through meetings and correspondence.
We have concluded that the QP must be supplemented by additional information, as specified in the enclosed request for additional information (RAI), before additional review and consideration of NRC approval can proceed.
Except for those areas discussed by the RAI, NRC staff feels that the QP adequately describes the characterization and solidification of sludge wash liquid with Portland Type I cement at the West Valley Site.
In responding to the RAI, West Valley Nuclear Services (WVNS) is requested to discuss the questions with NRC through correspondence and/or discussions prior to initiation of any testing programs resulting from the questions.
Additionally, WVNS is requested to provide formal responses to the questions, or a detailed schedule for responding to the questions (if additional testing is required), no more than 60 days from the data of this letter.
"**I#' (Original Signed by
)
John 0. Thoma, Section Leader Technical and Special Issues Section Low-Level Waste Management Branch Division of Low-Level Waste Management and Decommissioning Office of Nuclear Material Safety and Safeguards
Enclosure:
As stated DISTRIBUTION: : Central File LLWM r/f RBangart WBrach JAustin JKennedy
~JSurmeier MTokar NMSS r/f MiWs'millIB6kisSi nTCedeliFrsnce781bekW Difin? DistVi tiut16n?CbW PFs f' spend's!
g.M"Bn E"0 FC" ^Ti he~shfen~CT C6WFFE' = ^ CayeF C Encl osbr eTN~= ^ No"C6 py
- SEE PREVIOUS CONCURRENCE OFC LLWB
- E FCLB
- E LLWB E
LLWBk [
NAME RLewis/km GComfort JThom D OT PLoh d )
DATE 8/06/93 8/16/93 h /S/93 N
2/h/93
/ /93 S:/LLWMTYPE/WVRAI.RJL OFFICIAL RECORD COPY I
In smal ox on "DATE" line enter: M = E-Mail Distribution Copy; H = Hard Copy l
PDR: YES NO Category:
Proprietary or CF Only ACNW: YES NO %
IG: YES N0 4 Delete file after distribution:
Yes N No __
Of M.()
i 9308250374 930819 P
PROJ fd$2 PDR
=
Project M-32 Mr. Thomas P. Rowland, Director West Valley Project Office U.S. Department of Energy P.O. Box 191 West Valley, NY 14171 1
Dear Mr. Rowland:
i
SUBJECT:
REQUEST FOR ADDITIONAL INFORMATION FOR " WASTE FORM QUALJF CATION PROGRAM FOR CEMENT SOLIDIFICATION OF SLUDGE WASH LI 30" CONCERNING SOLIDIFICATION WITH TYPE I PORTLAND CEMENT.
The Nuclear Regulatory Commission has reviewed the subjeck ualification program (QP) notebook, submitted by the U.S. Department 4 f Energy (00E) in various components between 1991 and 1993.
NRC review'was performed as if the QP notebook had been a vendor submitted topical rpp6rt.
The technical review i
included infornation contained in Volumes 1 and as additional information that was attained as'j2 of the QP notebook, as well part of the review process through meetings and correspondence.
/
We have concluded that the QP must be supplemented by additional information, as specified in the enclosed request fpf additional information (RAI), before additional review and consideration those areas discussed by the RAI, M.orf NRC approval can proceed.Except for C staff feels that the QP adequately describes the characterization an,d solidification of sludge wash liquid with Portland Type I cement at the, West Valley Site.
In responding to the RAI, Wes't Valley Nuclear Services (WVNS) is requested to discuss the questions witivNRC through correspondence and/or discussions pricr to initiation of any testing programs resulting from the questions.
Additionally, WVNS is requested to provide formal responses to the questions, or a detailed schedule for responding to the questions (if additional testing is required), no mor'e than 60 days from the date of this letter.
Sincerely, John 0. Thoma, Section Leader Technical and Special Issues Section Low-Level Waste Management Branch Division of low-Level Waste Management and Decommissioning Office of Nuclear Material Safety and Safeguards
Enclosure:
As stated DISTRIBUTION: Central File LLWM r/f RBangart WBrach JAustin JKennedy JSurmeier MTokar NMSS r/f Mark Smsll? Boxesiin Conpuhehce/ Block to]DefinelDistHbut;io(losure; N =
CopyfPrefeFenc~el In small Bbion "0FC" line enter: C = Cover; E~=~ Cover & Enc 0FC LLWB 6
FCLB A cc E_
LLWB LLWB NAME RLewis/km/A4 GComfort JThoma PLohaus DATE f/6/93 N
f /6/93
/ /93
/ /93
/ /93
/
S:/LLWMTYPE/WVRAI.RJL OFFICIAL RECORD COPY In small Box on "DATE" line enter: M = E-Mail Distribution Copy; H = Hard Copy PDR: YES NO Category:
Proprietary or CF Only ACNW: YES NO IG: YES NO Delete file after distribution:
Yes No _
v s.
REQUEST FOR ADDITIONAL INFORMATION NUMBER 1 WASTE FORM OUALIFICATION PROGRAM FOR SOLIDIFICATION OF SLUDGE WASH LIOUID IN TYPE I PORTLAND CEMENT j
Submitted By:
West Valley Nuclear Services Through:
United States Department of Energy In reviewing the qualification program (QP) notebook, the U.S. Nuclear Regulatory Commission found four areas of concern for which additional clarification, justification, and/or testing will be necessary before NRC review of the Type I recipe can continue.
NRC approval for the solidification process will be predicated on additional review assuming acceptable West Valley Nuclear Services (WVNS) response to the concerns listed below.
The four areas of concern include:
(1) characterization for organics and sulfates in the concentrated decontaminated sludge wash, (2) compressive strength development of qualification test specimens, (3) validity of immersion test protocols and results, and (4) determination of process control plan (PCP) verification cube minimum strength.
Detailed description of each of NRC staff concern follows.
Waste Stream Characterization In terms of waste stream characterization, the solidified sludge wash will not be expected to exhibit the degree of variability that a typical low-level waste stream would.
In the QP notebook, WVNS does not seem to emphasize concern for identification of sulfate and organic content in the sludge wash.
The concentration of organic matter (1) determines if branch technical position (BTP) irradiation and biodegradation tests need to be performed, and (2) has been shown to retard set in previous WVNS recipes (i.e., the supernatant recipe).
Considering that one or more organic constituents (most likely oxalic acid) contributed to set retardation of the decontaminated supernatant recipe at concentrations of about 39 parts per million (ppm), WVNS should have included as part of waste characterization (1) analyses for total organic carbon (TOC), and (2) some provision for a more detailed analysis to identify the organic species if further analysis i deemed necessary as a s
result of verification procedures. WVNS continually refers to organic concentrations as "less than 500 ppm", and uses this concentration as an acceptable ceiling to justify omission of BTP tests (i.e., irradiation and biodegradation); however, the waste characterization text and table presented in the QP does not include any mention of testing for organic concentrations or species identification.
Furthermore, the BTP makes no mention of any acceptable concentration below which testing is unnecessary (though one could probably be justified).
Enclosure
s.
i WVNS needs to demonstrate a method for determination of the level of organic species in the sludge wash waste stream, and justify how and why.this level is insufficient as to require the irradiation and biodegradation qualification tests.
WVNS indicates that at a December 1990 meeting between WVNS and NRC at West Valley,"it was decided that such omissions were acceptable to NRC.
NRC records and personal recollection do not reflect this conclusion. WVNS-TRQ-029 and WVNS-TRQ-034 called for analysis of sludge wash concentrate on pH, TOC and total inorganic carbon (TIC), as well as on several inorganic species.
The results of this test request were never submitted to NRC because the-program was discontinued when the original recipe failed to pass the immersion test.
For the case of sulfates, WVNS indicated that a 25% sulfate spike was added to.
WVNS-TP-051 drums to provide for a margin of operation. The incomplete dissolution of the sulfate and its subsequent uneven distribution in the waste forms was partly blamed for the failure of these drum cores during immersion.
Campaign 22 production drum cores did not contain a sulfate spike. Thus the margin of operation intended to be provided during XXX-051 testing has been lost.
Furthermore, WVNS has indicated to NRC that the sulfate concentration i
as received from the evaporator shows a high degree of variability.
Since the formation of ettringite due to sulfate attack has been determined to.be one of the prime factors affecting stability for the sludge wash waste stream, NRC feels that the range of dissolved sulfate in the concentrated waste needs to be accounted for during qualification work.
WVNS is asked to comment on how they intend to characterize and control the sulfate concentration of the concentrated decontaminated sludge wash during production and qualification work. The nominal sulfate concentration of Campaign 22 drums is given to be 9.5% of total salt; what 'was the sulfate concentration of drum 82538, which was the Campaign 22 drum subjected to immersion testing? How would a high sulfate concentration in the concentrated i
waste affect immersion test performance?
1 Evaluation for Time-Development of Comoressive Strenath Compressive strength development in time was' monitored using the campaign 22-15 production drums.
Cored samples were tested at 49, 50, 65, and 82 days for different production drums. The same drum (drum 82212) was used for the 65 and 82 day tests, when it was noticed that the 65 day cores were noticeably damp, pe n ibly due to incomplete curing.
The 49 and 50 day cores exhibited an average strength of 1497 200 psi, and the 82-day cores exhibited a strength of 1620 1 297 psi. Actually, one of the three 82-day specimens showed a l
significantly higher strength, which raised their average. The 65 day drum (drum 82112) had an average strength of 1140 psi, or 70% of its strength 17 days later (i.e., at 82 days).
Comparison of drum 82112 strengths to those of other drums taken at 49 and 50 ' days after casting would indicate that samples are essentially (>75%) cured; however, such comparisons would need further
.iust i fication.
2
. _,, _. -., -, - -.... _ _, _ _. _ _ _ -.... ~... _.. -, _.. _ _.. _...,.. _.
Since WVNS has established that drum 82112 had-not completely cured in the tested time frame (in fact, evidence suggests drum 82112 was still rapidly curing at 65 days after casting), comparison of the drum 82212 results to other Campaign 22 drum results is inappropriate.
Not only is drum 82212 still curing at a rapid rate, there is apparently some other mechanism at work retarding its cure as compared to other drums.
Therefore, the evidence presented by WVNS does not sufficiently support the conclusion that Campaign 22-IS BTP testing was performed on essentially cured specimens.
There is no guarantee that the 82 day strength exhibited by drum 82112 is a value that would be exhibited by an essentially cured Camapign 22 drum core specimen.
Cores were taken from WVNS-XXX-044 drums after 38, 42, and 74 days of cure.
The XXX-044 drums were full-scale drums in which surrogate sludge wash at 20%
TDS was cemented with a nominal W/C of 0.64.
After like amounts of curing time, XXX-044 samples showed compressive strengths on the average of 250 psi less than the actual sludge wash drums of Campaign 22-1S. There is a slight but not significant increase in strength between 38 day (1202 145 psi), 42 day (1305 34 psi), and 74 day (1322 147 psi) cores for the XXX-044 drums, indicating that the cores have achieved an essentially cured condition after approximately 40 days cure, provided one assumes that ingredients, mixing, and curing conditions remained constant for all drums. This assumption is easily justified for these laboratory-created and controlled specimens.
Neat cement samples prepared in XXX-044 indicated complete curing after 28 days (i.e., no trend in strength increase was discernable for the given data).
XXX-051 drums were cored between 29 and 35 days with one exception.
Drum E was cored at 33, 40, and 47 days after casting.
This drum showed a slight decrease in average compressive strength over this time period from 1757 108 psi to 1670 115 psi, although this difference is statistically insignificant. No further testing was performed under XXX-051 that would indicate compressive strength time-behavior.
In conclusion, it is apparent that the surrogate waste recipe created under XXX-044 meets the essentially cured specimen guidance criteria of the BTP.
Insufficient data has been provided to indicate that the actual waste recipes of Campaign 22 or XXX-051 satisfy the as-cured waste form conditions.
Normally, the proof of essential cure using surrogate recipes is sufficient to meet NRC guidance criteria; however, in WVNS' case the immersion test performance of the qualified waste form is tantamount to the pre-immersion condition of the Campaign 22 drums. Therefore WVNS is asked to address the situation. One possible method of addressing this would be to remove additional cores from those Campaign 22 drums immersion tested to determine their strength today. (Note that due to the extended cure time of these specimens, a 75% strength is not necessarily what NRC would consider as the acceptedlimit).
Another possible way to address this concern is to use results of the archival testing program, if applicable. Of course, WVNS is free to address this question in any manner they wish.
Note: The high strengths exhibited by the Campaign 22-1S drums following immersion and other environmental testing tend to indicate that the integrity 3
b 1
l l
of the cored samples may have been allowed to develop sufficiently before l
testing to warrant conclusions as to these waste forms' stability in the post-environmental test condition.
Empirical data would be necessary to further verify this hypothesis and show, for example, that the specimen integrity is not being compromised by extended periods of immersion.
4 4
Campaian 22 Cores Immersion Testina 1
Cores were taken from one Campaign 22 1-S production drum after the failure in immersion of the XXX-051 drums. The Campaign 22 cores were expected to j
exhibit immersion test behavior similar to that of XXX-044 cores, because of the similarity in pH of the two test programs.
Nine (9) cores were removed from the top, middle, and bottom positions of drum 82538 (20% TDS/0.66 W/C) and subjected to immersion testing in simulated sea water.
Additionally, these immersion tests were performed (as opposed to XXX-051) at ambient laboratory temperatures. This drum cured 49 days prior to immersion test initiation.
The pre-and post immersion test results for this drum are presented in Table l.
Results indicate that the average strength after immersion decreased by 14% from 1603 psi to 1230 psi.
Consideration of only the 90 day immersion results indicates that 77% of the pre-immersion compressive strength has been attained.
Samples from the middle portion of the drum indicated a decrease in strength greater than that allowed by the BTP (i.e., decrease of 25% or more).
In fact, the sample demonstrating the lowest compressive strength (1030 psi) is only 60% of its pre-immersion strength of 1720 psi after 90 days immersion.
WVNS has not demonstrated that after 49 days of curing, the drum being tested has achieved an essentially cured condition (see above comments). A drum j
which has not completed curing will continue to cure while immersed, at the same time that it degrades (due to the action of the immersion fluid).
These i
competing effects can cause the immersion test performance of an specimen to be overestimated. Since after 90 days of cure the WVNS specimens show retention of 77% of their strength, even the slightest overestimation could cause samples to not meet the BTP guidance criterion of 75%.
4 Given the performance of the past sludge wash recipes in the immersion test:
(1) it is unclear why WVNS did not pursue additional testing (more 90 day, or possibly extended 180 day) for the Campaign 22 drum cores; also (2) the decision to test only the cores from one drum out of the many produced is questioned.
Is this individual drum statistically representative of all i
Campaign 22 drums? Subject to the interpretation of the BTP guidance, the tested cores are marginal in meeting the BTP test guidance criteria.
For NRC review to proceed, WVNS must demonstrate or discuss how and why sampling and immersion testing an individual drum is representative of all, or even some or most, of the drums produced using the Type I recipe at 20% TDS.
The responses to the above concerns about essentially cured pre-immersion 4
i
j..
i j
i 1
specimens will have impacts on the interpretations and analyses of the immersion test results.
4 Table 1.
Campaign 22 Drum Immersion Performance a
Test Condition Core Location and Comp. Strength, psi Top Middle Bottom Average Pre-immersion 1680 1720 1410 1603 49 days cure Post-immersion 43 days 1290 1490 1390 1
60 days 1560 1180 1500 1413 90 days 1410 1030 1250 1230 Determination of PCP Minimum Strenath NRC staff disagrees with the determination method for PCP minimum strength described by WVNS-TSR-053. This method involved preparation of 10 cubes with the nominal recipe (i.e., 20% TDS and W/C of 0.66) and 20 other cubes with slight variations from the nominal recipe.
Specifically, 5 cubes were made at 19 %TDS/ 0.64 W/C, 5 at 19/0.68, 5 at 21/0.64, and 5 at 21/0.68. The compressive strength, gel-time, and bleed water results obtained from these qualification specimens will be used for comparison purposes with PCP verification specimens obtained during production. The intent of selecting the chosen recipes was to blanket the possible range of variables that will be encountered in the cement solidification system during production.
Analysis of the results yields that, as expected, samples having lower %TDS and those having lower W/C demonstrate higher compressive ~ strengths. The lowest compressive strength (average 632 psi) was exhibited by the samples having the highest %TDS (21) and the highest W/C (0.68).
Because two input variables are changing that are not linearly 1% 'ndent, qualitative trends are all that can be identified through analyt of t strength data.
To determine the PCP verification specimen min.
c4 no.. compressive strength, WVHS averaged the compressive strengths for the thirty prepared cubes. The data was then tested for a normal distribution using the Chi-Square test,.and the 95% confidence level standard deviation was determined.
As the BTP suggests, the PCP minimum strength was then set at the average minus two standard deviation level. This value was found to be 383 psi.
the 95% standard deviation was 204 psi.
NRC staff feels that WVNS has inappropriately applied a normal distribution to the samples.
Break data from a set of cubes prepared using a single 5
,r
)
formulation (i.e., %TDS and W/C) would be expected to exhibit normally distributed strength results.
However, comparison of samples with different ingredients would not generally be expected to follow a Gaussian distribution.
For example if several more of the high W/C, high %TDS samples, which had the lowest strength, were prepared, the resulting distribution would show a positive skew.
Because equal numbers of samples below and above the nominal recipe were created, the effect of the skew of the curve was not observed; however, the distribution resulting from the 30 samples tested could only be expected to be Gaussian if a change in %TDS from 19-21 has the same effect on strength as a change in W/C ratio-from 0.64 to 0.68 does.
There is no.
evidence that this is the case (in fact, analysis of the data seems to suggest that W/C ratio changes have a greater effect on the strength).
The effect of applying a Gaussian distribution to the several kinds of samples was that the standard deviation was extremely large.
This large standard deviation caused the average minus two sigma level to be such that essentially any prepared specimen could pass the criterion.
It is not clear how this limit will effectively screen for nonconforming specimens. Consider that the worst strength obtained from the worst case (21% TDS/ 0.68 W/C) qualification sample yielded a strength of 548 psi, well above the PCP minimum of 383 psi.
This sample composition is at the very edge of the process control window to be used during production.
WVNS is asked to justify their choice of PCP specimen minimum strength and the method by which it was determined. Alternatively, the PCP minimum strength could be redefined through additional testing.
Two possible courses of action are (1) some method of quantifying the effects of the variables on the compressive strength and then weighing the obtained values, or (2) defining a PCP minimum using an alternative technique, such as using only the nominal recipe or defining the PCP minimum at each %TDS and its highest allowed W/C ratio (this sample would be expected to have the lowest strength for the given TDS).
WVNS is asked to discuss the choice of 30 samples as a representative and sufficient statistical population. While 30 may be sufficient, it is unclear from the notebook if this was considered and what the justification, if any, there is for such a choice.
NOTE: The concerns presented in this particular question will also apply to the Type V recipes presented in Volume 3 of the QP notebook, although review of those recipes is still being performed.
l i
6