ML20206A212
| ML20206A212 | |
| Person / Time | |
|---|---|
| Issue date: | 02/12/1987 |
| From: | Martin D NRC OFFICE OF NUCLEAR MATERIAL SAFETY & SAFEGUARDS (NMSS) |
| To: | Brinkman J JACOBS ENGINEERING GROUP, INC. |
| References | |
| REF-WM-39 NUDOCS 8704070447 | |
| Download: ML20206A212 (3) | |
Text
,
pts it E@
WM-39/ TAC LETTER /MY/87/02/12 Jim Brinkman Jacobs Engineering Group First National Bank Building Suite 1516 5301 Central Avenue, N.E.
Albuquerque, NM 87108
Dear Mr Brinkman:
The NRC staff has received and reviewed the material pertinent to the infiltration stud; and the justification for using a 1-Dimensional computer code to simulate infiltration, dated. December 22 and 31, respectively.
The enclosed comments reflect the broad concern that the computer code and input data do not incorporate sufficient conservatism to compensate for uncertainty, when simulating the major processes known to influence infiltration. Depending on how this broad concern is addressed, numerous other detailed questions would arise.
The subject transmittals did not contain information on the application of the model to estimating rtdon exhalation rates. Therefore, comments pertaining specifically to this aspect of the project are not enclosed. We note, however, that there is considerable uncertainty inherent in procedures used to calculate long-term moisture contents. The conservatism necessary to overcome this uncertainty with respect to infiltration will make the program unsuitable for determining radon exhalation rates.
If you have questions regarding this transmittal please contact me, or Mr.
Michael Young, of my staff, on (FTS) 427-4450.
Sincerely, 4RIGING S3suso sr Dan E. Martin, Section Leader Uranium Recovery Projects Section Low-Level Waste and Uranium Recovery Projects Branch Division of Waste Management Office of Nuclear Material Safety and Safeguards
Enclosure:
WM Record file WM Project cc: John Anderson, DOE /Alb.
Docket No.
Andrew Soloby, Jacobs/Alb.
PDR 7 LPDR -
-Distribution:
8704070447 870212 PDR WASTE WM-39 PDR (Return ta WM[623 S,S)
L WM-39/MHY/ TAC LETTER i l
\\
NRC STAFF COMMENTS ON TAC-INFILTRATION STUDY 1.
TRANSFORMATION OF BOUNDARY CONDITIONS The instantaneous transformation of boundary conditions from psi = 0 to psi =
15 bars, and vice versa, does not conservatively represent physical changes of capillary potential or the subsequent effects on-infiltration.
From the version of S0ILM0IST presented to NRC staff on 15 November 1986, it appears the amount of evaporation would be higher than anticipated under field conditions due to this rapid change of boundary conditions. Hence, the simulated flux i
rate of water through the tailings will be less than expected in the field.
This condition may lead to erroneous conclusions of potential impacts to l
ground-water resources and inappropriate design specifications.
Is there a way to increase psi in a step-wise fashion when the source term (i.e. upper layer of water) is exhausted? Perhaps decreasing the magnitude of the time steps during this transformation will allow the program to more accurately simulate 4
changes in capillary pressure while maintaining numerical stability.
l 1
2.
CALCULATION OF DETENTION TIME The methodology presented for calculating the detention time of water on the pile appears overly simplified. The use of Darcy's Law and other travel time relationships, although theoretically correct, do not account for local imperfections in tne radon barrier and/or filter layer which can drastically alter flow rates.
It is more likely that local imperfections, such as q
depressions or small fractures, will cause water to remain on the cover longer and/or infiltrate through the tailings material faster. Would incorporating a j
' safety factor' that increased the theoretical detention time, improve the l
3 degree of conservatism with respect to infiltration and address the inaccuracy of using Darcy's Law? This safety factor, although not physically significant, I
l may improve the realism of the results.
t i
With respect to the transmittal dated 31 December 1986, an example describes the approach used to calculate detention time of water in the filter layer j
following a rainfall event. This example uses a slope of 0.2, which leads to a i
detention time of five (5) days. When NRC staff used this approach, j
incorporating the variable slope values representative of past disposal designs, the detention time increased to 33 days. This represents a i
1 significantly longer time period for which infiltration occurs, and a corresponding increase in water flux rates. Using two slope values-in the i
calculations more accurately represents the actual tailings pile. configuration.
3.
SIMULATION IN 2-DIMENSIONS l
S0ILMOIST simulates flow only in the downward direction, and considers lateral i
flow to be insignificant. To date, we have received no demonstration that i
I
J t
WM-39/MHY/ TAC LETTER
, lateral flow is insignificant throughout the radon barrier, unsaturated-tailings and foundation soil.
It would appear that the slope of the final enclosed disposal cell could result in lateral flow. This concern is similar to a comment made by the technical experts, as described in the transmittal.
This issue should be resolved, either by incorporating 2-Dimensional flow or demonstrating that 1-Dimensional simulation is appropriate.
4.
INCORPORATION OF CELL CHARACTERISTICS Although S0ILM0IST incorporates disposal cell characteristics such as slope and thickness of the cover material, there are no provisions for using non-uniform values in the program. For instance, only one slope value can be used in the program for a given simulation run, even though the majority of the disposal designs use two slopes. Also, site-specific conditions may require thicker layers of material on one side of the pile than the other. Thus, SOILMOIST may not be valid for use with piles designed with non-uniform parameter values or possibly unique disposal designs. Using site-specific design features as input data into the program may produce more accurate results.
5.
S0ILMOIST VALIDATION SOILM0IST has not yet been validated with reliable data or field studies. We recognize that both validated datasets and field studies are likely to be used in the future to validate S0ILM0IST. However, it appears that the milestones set for completion of this project may not accomodate thorough validation prior to the use of SOILMOIST in actual designs. NRC staff feel that validating the computer code is of primary concern, especially since disposal designs resulting from S0ILM0IST are pending. Therefore, we suggest you slow down the pace of the project and fully validate the computer code prior to using it as a design tool.
6.
DOCUMENTATION OF INPUT DATA The use of accurate input data for each simulation is as important to running the model as the program code itself.
Initial boundary conditions, characteristic soil properties, and appropriate climatological scenarios each contain levels of uncertainty which can greatly affect output results and subsequent designs. NRC staff are concerned that uncertainties in the input data may overshadow advances in S0ILMOIST. Therefore, NRC staff stress the need for fully documenting how all input data was acquired and incorporated into the program, along with assessments of possible error. This documentation should be available for simulations pertaining to both downward water flux and upward radon exhalation.