ML20023E228
| ML20023E228 | |
| Person / Time | |
|---|---|
| Issue date: | 05/25/1983 |
| From: | Boyle R, Clark A NRC, NRC OFFICE OF NUCLEAR MATERIAL SAFETY & SAFEGUARDS (NMSS) |
| To: | Bell M, Rouse L NRC, NRC OFFICE OF NUCLEAR MATERIAL SAFETY & SAFEGUARDS (NMSS) |
| References | |
| REF-PROJ-M-32 NUDOCS 8306150216 | |
| Download: ML20023E228 (9) | |
Text
s Distribution: Original concurrence
-to be returned to FBrown SS 396 LP.rojectM-32 ATClark PDR NMSS R/F RBoyle LPDRs gy g,. M FCAF R/F RDHurt JGDavis WBurkhardt DBMausshardt WTing RECunningham LA file Project M-32 1
DRChapell SCornell/GBeveridge MEMORANDUM FOR: Leland C. Rouse, Chief. FCAF Michael J. Bell, Chief. WMHL FROM:
Regis Boyle, WHHL A. Thomas Clark, Jr., FCAF
SUBJECT:
TRIP TO SAVANNAH RIVER ON DEFENSE WASTE SOLIDIFICATION APPLICATION TO WEST VALLEY
Background
The West Valley Demonstration Project Act (Act) directs the U.S. Department of Energy (00E) to carry out a high-level radioactive waste solidification demonstration project at West Valley. Because the objective of the Defense Wastes Processing Facility (DWPF) at the Savannah River Plant is similiar to that of the West Valley Project, DOE has been considering the applica-bility of high-level waste management techniques used at its Savannah River Plant to the West Valley project.
Through arrangements with the DOE's Savannah River Operation's Office, we visited the Savannah River site to observe test equipment operation and discuss various aspects of waste management techniques with DOE personnel and with representatives of the Department's prime contractor at the site, the DuPont Corporation. A list of discussion participants is attach 6d.
DuPont's Engineering Department has contracted with the Bechtel Corporation to design a high-level waste solidification facility, the Defense Waste Processing Facility, DWPF. Construction on this project is nomirally stated to begin in 1985 with hot startup about 1989.
Huch of the information we obtained is related to the design and operation of the DWPF.
Its direct appitcability to the management of wastes at l
West Valley is uncertain. However, DOE has indicated to us that they in-tend to rely on SRP technical documents to support decisions on the West Valley project. Since the Department's reference alternative for the i
l solidification demonstration at West-Valley is the separated salt / sludge -
boros111cate glass process, it is our view that much of the information will be useful. _Our visit to Savannah River was planned to coincide with the development testing of a half-scale melter at the TNX-ETF location at SRL.
8306150216 830525 i
PDR PROJ M-32 PDR omcEt su:NAME)
......................l.......................
DATE )
OFFiClAL RECORD COPY usam mi-mmo Nec ronu sia oo-aci nacu c24o
+
Leland C. Rouse GAY 5 M Michael J. Bell Process Aspects
.Although sufficient information has been available to begin the design of the DWPF, the DuPont Company has underway a considerable effort to simplify and improve the design of the DWPF in many aspects. One of the key efforts is directed at performing more of the waste pretreatment directly in the waste tanks in the " farms" at the site.
Some of this work is ongoing, in an
" Interim Program."
Currently, waste supernatant solutions, consisting primarily of cesium-137 and smaller amounts of strontium-90 in a sodium nitrate solution, are being withdrawn from older tanks, evaporated to incipient crystallization, and transferred to newer tanks. Sludges, which contain the remaining activity are to be slurried and transferred to newer tanks. Older tanks, cleaned of their, contents, will be retired. DOE indicated that if the DWPF starts solid-ifying wastes in 1989 as scheduled, there would be no need for additional new storage tanks. This program then offers some further in-tank processing as follows.
~
l The sludges, which are freed of the supernatant, can be treated with a caustic wash to dissolve most of the aluminum present in the waste. This step was originally planned for the DWPF.
Its elimination from the DWPF design will be a cost savings. The sludges will be washed further with water prior to slurrying for transfer to the DWPF.
A further treatment of the. separated salt solutions is being considered.
This treatment would also take place in the waste tanks. Special chemicals,'
to precipitate Cs-137 and absorb strontium-90, respectively,0 H) are added
(
sodium tetraphenylborate (NaTPB) and sodium titanate (NaTi i
5 The slurry is continuously pumped through a bank of Hott sintered steel filters, releasing
\\
i l
a clear liquid radially from the tubes, the." enriched" slurry returning to -
,(n the tank. The slurry is continuously circulated until a 10 percent slurry
'is obtained.
[\\
i t
Both of these process techniques, i.e., sludge treatment and in-tank i
T l
precipitation, simplify the processing to be done in the DWPF.
In the next A~
l section we discuss some ramifications of these process improvements on melter
(
operation.
i
'i
(
l i
i i
1
-)
l l
l l
l' l
.a.,
orricey
.............. - - ~ ~
sua==e >
....................a..
our>
l xac ronu ais oo-ao> uncu ono OFFICIAL RECORD COPY usce mi-a.A
1 JL
~.
Leland C. '.suse Michael J. Bell-i Melter Operation
. The melter is the heart of the high-level waste solidification process.
Savannah River has been concentrating its efforts on the Slurry Fed Ceramic i
Melter _(SFCM). The SFCM is _ similar in concept to the. liquid fed ceramic l
melter being developed at Battelle's Pacific Northwest Laboratory. The
. calcining step has been eliminated, i.e., rather than converting to a solid powder feed to the melter, a liould slurry is fed directly to the melter and the liquid is quickly evaporated.
Some. feed preparation-is necessary prior to feeding the melter. The principal step occurs in the Slurry Receipt Adjustment Tank (SRAT).
Formic acid is added to the sludge in the SRAT. The formic acid reduces mercuric ions to the metal and is rernoved from the sludge, eventually to be returned to reprocessing operation. The formic acid'also helps to suppress foaming in the melter, improve slurry rheology, and convert ruthenium to a less volatile lower oxidation state.
As proposed for the DWPF design, e melter will consist of a 304L stainless steel shell lined with Fiberfrax
, a thermal. insulatory material. The brick is highly corrosion'resistantpofrax k3. All metallic inserts into i
the melter are made of Inconel 690. This incl selectrodeplatesgnd j
rods and various instrument probes. Inconel 690_
melts at about 1340 C.
The basic melter design has been likened to a teakettle, i.e., a simple pot with a pour spout (riser). Product from the melter is drawn from the spout. The experimental melter we observed at the TNX complex could be i
tipped abcut a horizontal axis (as in pouring from a' teakettle), rotating either forward or backward to increage or gecrease pour rate. Very 11ttie tilt is required. (cn the order of 1 to 3.)
Pressure control on the spout is being considered for DWPF design to achieve the same purpose.
[
Various spout designs are being considered.
t Heating or cooling a melter to change temperature is a tedious procedure.-
A general rule is - once hot, always hot.
It is likely that the DWPF L
melter.will be thoroughly tested at THX, cooled, installed in the new plant, heated and used continuously at full heat until its end-of-life..
At operating temperature heat.is principally supplied by resistance (Joule);
i I
heating of the glass melt using diametrically opposed electrodes. The,
resistivity.of the melt is important and is too high at room temperature to.
heat the glass. Therefore, an auxiliary heater is used in the lid of the i
melter to heat the contents to the point at which the electrodes can supply sufficient heat for melting. The lid heater always supplies significant i
energy for process control. The pour spout is also independently heatea.
t
\\
Inconel-sheathed Silicon carbide lid heaters have been used in the THX
\\
l' experimental melter (two-thirds scale linear dimensions, 46% scale by melter area). Other materials are being considered for the DWPF design.
i
. If the resistivitg of the melt is too low (high ionic composition),lthere.
\\
is insufficient I R loss. A water jacket is planned for the DWPF melter for contro l of thermal 1pdrafts in the operating cell.
5 omes)
..,.......7....-.
I. SURNAME)
.a--..*...-...
pare)
.. ~. - -... -.
Nac ronu ais oow nncu om OFFICIAL RECORD COPY usam asi km
Leland C. Rouse MY ' e mq3 Michael J. Bell i d
At TNX we observed melter operation. Quartz viewing ports and a closed-circuit television system (CCTV) permitted observation of the surface of the melt the pour through, and the fine stream pouring directly into the waste containers. The waste containers used for testing are of carbon steel; those planned for the DWPF product will be of stainless steel. Stainless steel containers have been used at TNX for the canister testing programs.
1 The physical chemistry of the melt is quite interesting. A " solid" state is maintained at the surface of the melt, termed a " cold cap".
The thickness of the cold cap varies from ane to six inches and the surface temperature of n
the " cold cap" is nominally 100 C.
The slurry mixture of water, waste, and glass frit is slowly, but continuously poured directly on tne " cold cap".
The slurry fed temperature is approximately 40"C.
The feed consists of approximately 40% sludge and frit and 60% water. The sludge and frit consists of 65% of frit and 35% sludge. The mixture trickles and spreads over the surface, the water evaporating, the remaining solids slowly being absorbed into the melt. As the solids enter the " cold cap" the frit melts and chemical changes occur in the waste. The reducing climate of the melt releases hydrogen, carbon oxide, and oxygen gases from the melt which bubble to the surface. The surface of the " cold cap" is always changing its configuration as hot melt bubbles up from below forming small pools at the surface.~ This action helps the melt absorb the recently added material.
~ Beneath the " cold cap" thermal currents (convective flow induced by temperature differences)areinducedinthemelttofurthermixthecontentsandobtain a homogeneous mixture. These thermal currents are induced by providing a current flow in one set of juxtaposed electrodes which is greater tnan the current in the other set (there are two sets of two electrodes with two electrically separated, out-of-phase, power supplies). The thernal cu'rrents and off-gas bubbling contribute a source to the off-gas system, which naintains a vacuum of two inches of water or greater in the vapor space of the melter. We were told that in order to assure a homogeneous mixture, a melter residence time of 12 hours1.388889e-4 days <br />0.00333 hours <br />1.984127e-5 weeks <br />4.566e-6 months <br /> is recommended. The density of the melt at operating temperature is about 2.4 grams per cubic centimeter.
Viscosity of the melt is important and is a somewhat complicated function of temperatures (see DP-1507). The viscosity of the melt should be below 50 poise for good pouring. Low viscosity may increase electrodg corrosion.
The temperature of the melt must be maintained at abogt 1150 C.
Below g
950 C crystals of iron spinel will precipitate. 1050 C is considered a minimum operating temperature. Corrosion rates of the melter electrodes (which may control the life of the melter) are expected to range from 0.5 to 1.8 mils per day. The contents of the melter can be drained through a freeze plug at the botton of the melter.
In the next sections we discuss ongoing tests on thermal current control and off-gas decontamination and control, errice >
sunnus >
DATe)
OFFICIAL RECORD COPY usom mi_us..so anc ronu sia no-so) nacu o24o
4 Leland C. Rouse Michael J. Bell M4y 2 5 B63 Physical Hodel using chemical engineering principles in the equations of material, energy, and electropotential conservation and balance, the TNX engineers have entify as a Physicci developed an analog model of the melter which they gtank small enough to Model. The Physical Model consists of a Plexiglas fit on a conference room table. The tank contains a solution of lithium chloride in glycerine. This solution provides a room temperature asplog of the melt.
In the model cooling water jackets, again of Plexiglas W, were provided at the bottom, sides, and riser. Cooling water can also be supplied to an approximately 1/2-inch thick stainless steel header which simulates the " cold cap".
The header is fed by a network of variable-feed distribution tubes, perhaps 1/8-inch in dianeter. Through dimensionless analysis, the solution, cooling water feed rates and temperatures, and the electrical potential and current flow provide a reliable analog to the actual melter. The variables studied in the Physical Model are melt temperatures and composition (honogeneity) as related to current and voltage and heat transfer properties. The Physical fiodel will be useful for melter design and operating characteristics and limits.
In addition to the Physical Model some computer analysis has been undertaken with sub-contractors in two dimensions. A three-dimension analytical model is being considered.
Off-Gas System An estimated one percent of the melter radioactive contents may become airborne by the physiochemical actions in the molter as previously described.
Considerable effort has been expended at THX to provide an adequate technical basis for the design of the melter off-gas system for the DUPF. During melter test operations at TNX, off-gas system tests are also underway. Tne off-gases will be drawn by an exhauster through a series of components which cool the gas, and remove particulate and semi-volatile material.
A venturi scrubber will likely be ged to cool the gas. Testing is presently underway on a multistage HydrosoniFscrubber for initial high-efficiency decontamination. Other similar gas atomized scrubbers are being evaluated. Two HEPA filters in series are planned for further decontamination downstream.
A high-efficiency mist eliminator is also presently planned for the system.
8 A minimum decontamination factor of greater than 10 might be expected from such an off-gas treatment system with additional filtration from the ventilation system. We expect to follow the progress of this development work for its possible direct application at West Valley.
omce>
SURNAhnEh oue) nac ronu sis oo-soi nncu oaa OFFICIAL RECORD COPY usamw _a3m
l Leland C. Rouse *
, Michael J. Bell ygy 3 g 99g3 Container Operations At' feed' rates greater than 50 pounds per hour and minimum pour temperatures of 1050 C no voids have formed in the 24-inch diameter containers.
It is planned to remove containers from beneath the' melter pour spout after each ~
is filled.
(Note: Containers will not actually be filled to the top; as'
- a precaution a freeboard will be available in each container.) The containers will be cooled at a " controlled rate" to minimize cracking of the glass.
After cooling and decontaminating the exterior surface of the containers, p
they w111-be moved to a welding station. At TNX we observed the results obtained from sealing a container using.a resistance-upset wslder.
In this technique,a stainless steel plug is inserted in the canister throat. A surge of high amperage current (240,000) is applied across the plug container gap while the plug is under 60,000 pounds force. The weld which is formed is virtually undetectable in a section specimen, looking like one continuous piece.
Waste Form Programs i
DuPont's testing program for waste form is divided into three subprograms.
l These include (1) a borosilicate glass / water interactions program,- (2) labor-atory tests on waste / repository interactions and (3) actual in-situ testing of i
l borosilicate glass performance in the Stripa facility (SwedeliT.
i The objective of the first subprogram is to provide guidance to the repository l
developers by testing for repository conditions that will improve the per-l formance of the waste form. For example DuPont has found by conducting
- tests at various pH values for the' buffer surrounding the waste form that boros111cate glass performs better in a neutral environment. DuPont has also 1
l
'found in preliminary testing that use of metallic lead in the package or back-
-fill material reduces the leachability of the glass by about a factor of 100.
Their tests have shown that this reduction in leachability is obtained when-the metallic lead is placed in the backfill material or the package. However, L
placing the lead directly in the borosilicate glass as Pb ions has not resulted l
in a reduced leach rate. These tests are apparently continuing so that a better understanding of this phenomenon can be obtained.
i L
In addition to the above testg. DuPgnt is also conducting tests which vary i
temperature (tests in the 120 C-130 C range will be conducted), waste surface area to volume ratio Eh, radiation levels, and glass composition.
i~
.The second subprogram being conducted at SRP relates to the waste / repository interactions. We were told that the principal document on this subject was
-titled "An ' Assessment of _ Savannah River Boros111cate Glass in the Repository Environment " (DP-1629). This report was provided to us in' July 1982 when we received the SRP Environmental Assessment on the selection of borosilicate glass as the waste form for SRP wastas. At that time we were critical of.
~ the document because it considered only salt and granite as possible. geologic media for a repository. DuPcnt is now beginning to conduct laboratory tests on how borosilicate glass would perform in a typical geological environment. We observed test specimens of basalt and tuff which will be used in tests planned
_ orrice) 1 su m te>
one>
nne ronu sia (10-80) NRCM 024o OFFICIAL RECORD COPY usam m-mm
1-Leland C. Rouse D 5 PJ03
~
Michael J. Bell-for the latter part of February or early Harch,1983. From our discussions with DuPont personnel and our observations at the laboratory, it appears that DuPont fully intends to carry out tests that would satisfy several of the concerns'which were expressed in our comments on the SRP Environmental Assessment.
Constituents of the waste are important to the final product " structural" quality. This is particularly true of aluminum in which case a high aluminum content decreases the durability of the glass.
We inquired about the effect of thorium on the waste glass. We were told that thorium is'likely to be less leachable than uranium but 1.he effect of thorium on' leach rates of boros111cate glass has not been studied..
with DuPont personnel.(in-situ testing at the Stripa facility) was not discussed The third subprogram Criticality Considhrable effort has been spent on nuclear criticality analysis. The
, principal fissile material of concern is plutoalum-239, which is present in small amounts in the waste. The thermodynamics of the melter indicate i
that almost all other materials would be reduced in the' melter (and thereby settle out) before plutonium dioxide. At about 0.018 weight per cent or l
less, the concentration of plutonium dioxide is well below the solubility
. limit of about 4 to 7 weight per cent. At any rate, under extreme conditions
'of operation, the infinite multiplication factor, K, has been calculated to be less than 0.15. The situation at West Valley should be similar.
Steam Explosion-Although, as mentioned above, the normal slurry feed rate to the welter is low compared with the melter inventory DuPont has still been investigating the plausibility of an accident involving a steam explosion in the melter.
L Studies done in the United States and abroad have generally concluded that a melt-water to vapor pressurization could never exceed the confinement
~
strength of the melter. However, the possibility of salt formation and reaction with water is still being pursued. We intend to follow this work.
Other Accidents A systecatic analysis has been performed of the accidents which may be credible for the DWPF. This analysis will _be very useful in our review of West Valley, at least from the standpcint of a collection of likely incidents.
Naturally, analysis for West Valley will have to consider the specific systems and site location for an accurate ar.alysis.
4 omcs >
tuanwa >
DATE)
A ac ronu sisyo-somacu ano _
N OFFICIAL RECORD COPY uscro: mi-a-=
Leland C. Rouse MAY 2 5 1963 Michael J. Bell Sludge Removal We were shown the facilities and equipment at TNX to develop and test sludge removal equipment. A full scale mock-up'of the bottom of a typical high-level waste storage tank is installed at THX. These tests form the basis for sludge slurry and renoval from tanks in the waste " farms" at the plant. DuPont has also constructed a 1/12th scale model of a tank for scale-up testing. We were shown several long-shaft pumps trsad for the tests.
I A. Thomas Clark, Jr.
Advanced Fuel and Spent Fuel Licensing Branch id Regis Boyle High Level Waste Licensing lianagement Branch
Enclosure:
As stated i
t Disc OB 3-7, A23
/d.9 FC F WM FCAF
.g...........
.c or,,m >
g sunn.cz >.Af B5x.a.lb..8hy.le.........Lchub one>..... 5&/83...
.. 5/9./83..
.. 5Af/83.......
5/.L B3...........
Nac rosu sis com> nacu ouo OFFIClAL RECORD COPY usomni-maa
r List of Persons Contacted at Savannah River SR00 DuPont T. B. Hindman, Jr.
B. G. Kitchen W. B. Wilson J. F. Ortaldo J. W. Geiger F. H. Brown C. T. Randall M. J. Plodinec M. D. Boersma D. J. Pellarin E.. J. - Hennelly L. F. Landon W. S. Durant A. J. Renwick K. R. Routt M. D. Dukes i
n l
i I
i f
l
-- _ _, _