ML19249A460
| ML19249A460 | |
| Person / Time | |
|---|---|
| Site: | Crane  |
| Issue date: | 06/13/1979 |
| From: | Murley T NRC OFFICE OF NUCLEAR REGULATORY RESEARCH (RES) |
| To: | Mattson R Office of Nuclear Reactor Regulation |
| References | |
| NUDOCS 7908230004 | |
| Download: ML19249A460 (5) | |
Text
ENCL 0SURE 1
7908030CM
??4236
s hc UNITED MATES y?
NUCLEAR REGULATORY COMMISslON gf%
g WASHINGTON, D. C. 20555 e
June 13,1979 MEMORANDUM FOR: Roger Mattson, Director Division of Systems Safety FROM:
Thomas E. Murley, Director Division of Reactor Safety Research
SUBJECT:
HYOR0 GEN /0XYGEN GENERATION DURING THE TMI ACCIDENT As you requested, this is an interim report on the status of the analyses that the RES staff and our contractors have done concerning the generation and behavior of Hydrogen and Oxygen during the TMI accident.
I have separated the discussion along the following lines:
- What do we know now?
- What did we believe at the time of tne accident?
As you probably know, we do not have definite answers to many of the questions. We have started some research work and plan a more com-prehensive program to answer as many questions as we can concerning the hydrogen generation and behavior in the TMI accident.
A.
What do we know now?
1.
How much hydrogen was generated in the accident?
We believe that nearly all of the hydrogen present was generated in zircaloy-stater reactions and not from radiolysis.
Preliminary calculations show that 30-40% of the zircaloy in the core was oxidized.
Further calculations will be made as part of the NRC investigation of the TMI accident.
2.
How much hydrogen escaped out the relief valve to containment and how much remained in the primary system?
We have n] answer to this question yet; a study is still under way.
3.
How much of the " hydrogen bubble" in the reactor vessel was actually hydrogen (or other noncondensible gases) and how much was steam?
vff4
Roger Mattson JUN 131979 We have no answer to this question yet.
Preliminary cal-culations show that some if not most of the compressible volume in TMI could have been steam.
4.
If the compressible volume was a hydrogen gas bubble, how was it removed from the primary system?
We have no analyses currently under way on this question.
However, rough calculations during the period of April 1-4 indicated that it was unlikely that the rate of removal of hydrogen in solution via the letdown system and pressurizer spray system could account for the re-ported rate of decrease in hydrogen bubble volume.
5.
How much oxygen was formed by radiolysis?
The consensus now is that there was essentially no net production of oxygen gas during the event because the coolant was rich in hydrogen.
6.
What were the probable concentrations of hydrogen gas and oxygen gas in the primary system?
We believe that there was essentially no oxygen gas con-centration in the primary system.
There was clearly some hydrogen in solution, and there may have been some hydrogen gas in a bubble or in bubbles. We do not have a good estimate yet of the hydrogen gas concentration in the primary system.
Further calculations will be made as part of the NRC investigation of the TMI accident.
B.
What did we believe at the time of the accident?
1.
What was the hydrogen concentration in the primary system?
At the beginning of our involvement in the TMI events on Friday, March 30, we were told that measurements at the site indicated there was a noncondensible hydrogen gas bubble with a volume of 1000-1500 cu ft. at 1000 psi and 280 F.
To my knowledge this assertation was not seriously questioned during the period March 30 - April 4.
Subse-quently, (about April 5) RES staff questioned whether the compressible volume could in fact have included a significant amount of steam.
7 NAM 8
JUN 13 id79 Roger Mattson 2.
What was the oxygen gas concentration and rate of increase of oxygen concentration in the primerf system?
Information from RES contractors and consultants on Merch 30 and April 1 indicated that the assumed 1000 cu. ft. hydrogen bubble could have included oxygen generated fr ca radiolysis.
There were various estimates of the oxygen gas concentration in the hydrogen bubble, ranging from essentially none to a conservative estimate of oxygen gas concentration in the hydrogen bubble that was increasing at the rate of l'.' per day after reactor scram. The/c was some difficulty in calculating the oxygen concentration because the rate of recombination of oxygen with hydrogen in solution was not known. Later in the day on April 1, the consensus of advice to RES was that there was probably no free oxygen in the primary system.
I am enclosing a number of memos and references as background infor-mation.
Thomas E. Murley,dr-Mafety Research imctor Division of React
Enclosures:
As stated cc:
S. Levine L. S. Tong C. N. Kelber L. C. Shao
H.atson/ Rib 6/8/79 H IN TMI REACTOR VESSEL 2
Q - Can a H /02 mixture at elevated temperature and pressure detonate?
2 A - Yes it can.
SAND-74-0382 Chapter 10 presents a variety of experiments indicating detonation at temperatures up to 500 C (930*F). However.
detonation will not occur if water vapor is present in quantitieh = 60%
Q - Can H /02 mixture ignite spontaneously?
2 A - Yes at a temperature of approximately 550 C (1020*F) or higher (see SAND p.10-28).
However flamability limits are affected by water vapor content above 25% concentration. At 60% water vapor the mixture will not burn (ref. KAPL report).
Q - What is the effect of increasing temperature on the limit of detonation?
A - The lower limit of H2 concentration is reduced and the upper limit is increased.
The range is broadened over which ignition can occur.
For H -Air 017'C range is 9.4 - 71.5% (SAND p.10-9) 2 0 400*C range is 6.3 - 81.5% (and Figure 10.4).
Not a major effect over this range - but ignition can occur over a broader range.
The above is for H2-Air mixtures (i.e. H N ) if the N 2 2 2 2 is considered a diluent. Mixture behavior with water vapor would be different if N2 diluent is replaced by steam diluent', as discussed above.
Q - What is the effect of increased pressure on detonability?
A - The lower limit does not seem to change much but the upper limit (for H /Af r) 2 increases from 68 to 74% as the pressure is increased to 140 atm (2060 psi)
(See SAND Figure 10.5, p.10-11).
Data are not available for the affect of water vapor concentration with increased pressure.
Q - How much steam must be present in an H /02 mixture to avoid detonation?
2 A - There is no effect of water vapor below 7%; however, as the amount of water vapor increases above this value, the probability of an explosion decreases.
Flammability limits will not be affected until you have concentrations above 25%
water vapor. At 60% water vapor, the probability of burning H2 and O i
t s zero.
The effect of pressure has yet to be determined (ref. XAPL report on TMI-2 Incident).
Q - How do diluents (N2 or steam) in H 02 2 mixture affect peak pressures?
A - Addition of diluent narrows the range of flamability or detonability (see ternary diagram, p.10-20), reduces flame velocities (see Figure 10.8,
- p. 10-14) and reduces peak pressure (see Table 10.11 part 2, p. 10-17).
The effect of steam is discussed in the previous Q/A.
7/4240
[gne na cq%
UNITED STATES NUCLEAR REGULATORY COMMISSION
=
f j
WASHINGTON, D. c. 20555 i,Nt/'!
APR 2 51979 MEMORANDUM FOR: Thomas E. Murley, Director Division of Reactor Safety Research Charles N. Kelber, Assistant Director for Advanced Reactor Safety Research FRC" Louis N. Rib, Special Assistant Advanced Reactor Safety Research SUSJECT:
SUM".ARY OF PARTICIPATION IN TMI-2 INCIDENT--EVALUATION OF HYDROGEN GAS IN PRIMARY CLOLANT SYSTEM AND CONTAINMENT BUILDING ATMO5PHERE Following the TMI-2 incident on 3/28/79. Lenny Rib and John Larkins of ARSR with the support of Roy Ferson (NMSS) and Frank Witt (05D) provided support to RES in:
(1) the evaluation of the quantities of H2 that could be disolved in the primary coolant water following a Zr-H O 2
reaction in the core, (2) means oy which the H,> in the primary coolant system might be reduced and (3) estimates of the hydrogen reduction in the containment atmosphere. Support in the evaluation of items (2) and (3) was obtained from engineers and scientists at KAPL and ARSR contractors at LASL and Sandia.
The evaluations are sumarized in the following attachments:
1.
Sumary of Hydrogen Evolution (from saturated solutions) Calculations for TMI-2. Attachment 1 2.
Report on Hydrogen Solubility Data Ease and Estimated Hydrogen Releases at TMI-2.
3.
Upper Limit Estimates of Hydrogen Formation. Attachment 3 4.
Radiation and Physical Chemistry of LWR Primary Coolant.
h 5.
Hydrogen Reduction Methods - Sumary. Attachment 5 p( y
+
6 e
o.
W4241
APR 2 51979 Thomas E. Murley, Charles Kelber 6.
KAPL Sumary of Efforts on TMI-2.
Report in progress, (will be supplied at a later date.)
Louis N. Rib, Special Ass'stant Advanced Reactor Safety Re,-
ch cc:
- 5. Levine, RE5 R. Budnit:, RES R. Benaro, SD G. Arlotto, SD L. Person, NM55 F. Witt 50 J. Larkins, RSF.
K. Steyer, SD R. Minogae, SD V. Stello, DOR R. Mattson, 055 R. Scroggins, RES
[
umtso suits NUCLEAR REGULATORY COMMIT $10N g
g
}
n sm cro.o.c. m.
%,..../
EM)RANDUM FOR: 1hcnas E. Murley. Director Division of Reactor Safety Research FRCE:
Louis N. Rib, Special Assistant Advanced Reactor Safety Research
SUBJECT:
SUMMRY OF HYDROGEN EVOLUTION CALCULATION FOR TMI The attached table sumarized the calculations of hydrogen gas at several pressure 3nd temperature conditions corresponding to a systerr. cooldown, controlled pressure reduction and also for a rapi d depres s uri zation.
6 Y Louis N. Rib Special Assistant Advanced Rear. tor Safety Research cc:
S. Levine R. Budni tz C. Kelber
.~. r k
6 0
.~
V W(.
\\
l 6
f 7/4243
~
H Evolution fecrn System Temperature & Pressure Changes
- 2 3
P Hg (ft ) e3cived Comment psig *F) psih*T Case P
T F
a t' P T ' ft 22 1
1000,280 0,212 27,000 depressurization accident 2
1000,280 1000,212 98 stepwise cocidan 3
1000,280 1000,140 105 continued cooldown 4
1000,140 350,140 400 pressure redsction 5
1000,280 350,280 760 pressure reduction
- calculations based on H2 saturation concentrations checked against recent experiment in Provo Utah, and obtained from J. Chem. Eng. Data, 5_ 10 (1960),
D. M. Hi mme lbl an, Ch p t. o f Chem. En g., thi v. o f Te xas ( Aus ti n ).
9 774244
6 Ups1TED STATis k
NUCLEAR REGULATORY CCNMISs10N
.f L
}
WASHIMTON, D. c. 20E35 S
I April 11, 1979 g..C.
MEMORANDUM FOR: Charles H. Kelber, Assist' ant Director for Advanced Reactor Safety Research Division of Reactor Safety Research THRU:
/M/LouisN. Rib,SpecialAssistant Advanced Reactor Safety Research Division of Reactor Safety Research FR0m LeRoy S. Person Fuel Reprocessing and Recycle Branch Division of Fuel Cycle and Material Safety SU5]ECT:
REPORT ON HYDROGEN SOLUBILITY DATA BASE AND ESTIMATED HYDROGEh RELEASES AT THREE MILE ISLAND In order to bound i.he limits of hydrogen solubility and release, a parametric appraisal was performed to help detemine possible conditions of the reactor coolant at Three Mile Island-II.
Three references were examined for solubility data for hydrogen; I. S. Andersen's " Correlation of Solubility Data for Hydrogen and Nitrogen in Water," WAPD/TM-633, D. M. Hinnelblau's, " Solubility of Inert Gases
- - - -in Water O' C to Near the Critical Point of Water," and the Reactor Handbook, 1, pp. 852, 1960. Comparison was also done of experimental data from Roger Billings at Provo-Utah.
Data from the " Correlation of Solubility Data for Hydrogen and Nitrogen in Water" was very similar to that obtained frm the Solubility of Inert gases in water i.e., the difference in the Henry's law constant for each in the temperature range of interest (140' F to 280* F) was less than one percent.
Information obtained frm Volume 1 of the Reactor Handbook showed a discrepancy between what was reported there and the two above mentioned references (approximately 10 percent difference).
These calculations were compared with experimental values obtained from Provo-Utah and the differences were approximately 20 percent for Hinnelblau and Andersen's Data. There was not enough infomation from the P. actor Handbook to compare with the experimental infonnation.
h c
p
~
1
.O
/
/
Charles N. Kelber April 11. 1979 It is believed that the measurements from Provo-Utah were not taken at eouilibrium conditions (the approximate situation at Three Mile Island-II) and this is the reason for the discrepancy of 's percent in the calculations. Therefore, the information' developed by Himmelblau and Andersen was subsequently used in the calculations. A table is attached showing values obtained for the various conditions analyzed.
d
.P-LeRoy S. Person Fuel Reprocessing and Recycle Branch Division of Material Safety and Safeguards
Enclosure:
As stated cc:
R. E. Cunningham J. B. Martin L. C. Rouse R. A. Scrano e
9
0
~'o 9*
=
3 3
~??
4 a.
a
,"3 9
~
~-
=
=
4 g$*E
=
4 aa
=
.
4 1
o a=
S
=
00 2
4
'o
- gg t
~~
~
_S
=
=
~
=
h
.S.
a g
~
g3 m
8
==
0
~
a 8
~
~
o
>=
o.=
M 5
Y
?
e b
o S
2.
t
=
~
~
7 m
w l
5 z:
a s
o gg ei 5
=-
~
E
\\
S-3 an.=
M s~
N.
h Q
o i
N a
9 a-
?
2
'O~
D N
'l 6
~
f g.
4
~
-)].1
==
o d
D
- b. Ed
..]
., _2 1
,, ~
g.
a.s _ 3 3
88
--2~-
-,6
.gj I j =t*
--223
- ta 24 A
o 4
8 e
W
. =
O gBB@B@I 2
S
=
a a
~~
-E f'
2 2
~
'o s
5 :g=s
~
aa
~
s
=
J 4
N N
O a
3
=
E SS J
e S
8 E
~
=
g W 21
'o g.
5 gs f
4
~
-5 b
==
s
=
~
g~
2
~%
.A a
=
W e o
e
$ g"g*.~.
W
=
3 pn e
>Oq,;Pjd g
- w s u O
eE Ce, e5 O
- d e
g {3 22 _
- ~
6
- 23
-e:
- 2...
I*2 233 223
??"~a t
3 a
a.
o e
UNtTED ETATts p
NUCLEAR REGULATORY COMMISSION f
n AsMiscTon. o. c. nosas APR 11 1979 MEMORAND'E. FOR:
C. N. Kelber, Assistant Director for Advanced Reactor Safety Research Division of Reactor Safety Research L. N. Rib, Special Assistant THRU:
for Advanced Reactor Safety Research Division of Reactor Safety Research F R0". :
J. T. Larkins Experimental Gas-Cooled Reactor Safety Research Branch SU5 JECT:
UPPER LIMIT ESTIMATES OF HYDROGEN FOPF.ATION Using available information and data (I & E reports, plant status reports, etc.) an upper limit estimate has been made of the amount of hydrogen generated via the zirconium-water reaction and from radiolysis.
Also, using available data and expert comments an estimate is included on the amount of hydrogen that could be absorbed by the zirconium cladding.
A hydrogen explosion was reported to have occurred as the primary system was being depressurized to allow operation of the Residual Heat Removal (RHR) system. An estimate of 70,000 scf of hydrogen was assumed to have been burned. This woulo have represented a hydrogen concentration of 3.33*. in containment (2.1 million cubic feet) therefore, the burning or explosion must have been localized or the estimate of the amount of gas involved is too low.
On March 31, 1979 it was estimated that the containment building contained 1.7% hydrogen and that there was 3 hydrogen bubble in the reactor vessel of approximately 1,000 ft. at 1,000 psi. Based on available solubility infomation these three sources would provide a total hydrogen inventory of approximately 113,000 scf of hydrogen. This total inventory 5.3t (exceeds flaw. ability limit of 4%) gen concentration of a in containment would have given a hydro. Assuming that all of the hydrogen generated (including that which was burnt) was generated from the zirconium-water reaction this would calculate to 42-44% of the available zirconium being oxidized.
(
a g t.
Q W4249 O
g\\
b
~
O e
~
APR 11579
~2-C. N. Ke Iber
)
On April 4,1979 it was reported that the con imately in the reactor vessel (~10,000 scf) and with a saturatedvolume a coolant (27,000 scf) and waste gas tank (1938 ft.3 200 ft, inventcry of approximately 56t hydrogen (~ 5,400 scf) there was a hydrog of approximately 90,600 scf.would still have exceeded the set flam is The amount of hydrogen being fomed frce radiolytic dec:cposition t) probably very small, however, using a normal decay curve (B&
f LASL using the COGAP code calculated and 1% of (449 scf) per available fission products in the coolant another 2.5 lbs.T gallons) day in the coolant.
in the containment is approximately 50 scf per day.
t in of hydrogen from radiolysis would be 1,755 scf (.08t volum 0.45 (molecules of hydrogen evolved 100 ev of radiation abso containment) per day.
rring.
not take into account the amount of hydrogen recombination occu The amount of hydrogen recombining to form water could b 10.
and reduce the amount of hydrogen generated by a factor of d one Two hydrogen recombiners were made operational on April
- 79. Using the unit reportly started processing gas on April 3, 19 t ring the efficiency given that for a 4% hydrogen concentration en e cess flow rate recombiner that the exiting gas was.1% hydrog imately one week to drop the concentration by 25t and 17 days to by 50t.
From an assemblage of various references and conversa been absorbed by concluded that the amount of hydrogen that could haveAssuming 10 the zirconium cladding was small.
ly absorbed in the zirconium cladding, one would calculate on1%
~815 scf of hydrogcn absorbed, which is less than hydrogen generated.
e 774250 a
APR 117979
-3 C. M. Kelber Lastly, on the use of chemicals for gettering dissolved and gaseous hydrogen Sandia has used unsaturated hydrocarbons (dimer ether) in the weapons program for controlling hydrogen b investigated somewhat, however, some further work would be needed.
closed system.
Sandia has proposed its use in LWR safety for gettering hydrogen from core-melt accident and it appears to have a good potential for use in I anticipate receiving more preventing hydrogen build-up in a LOCA.
information on this subject.
fn John T. Larkins Experimental Gas-Cooled Reactor Safety Research Branch 774251 l
[
UNITED STATES NUCLEAR REGULATORY COMMISslON
~'
3 I
C o
nASHINGTON. D. C. 20688 l
..)..
APR 2 51979 MEMORANDUM FOR: Charles N. Kelber, Assistant Director for Advanced Reactor Safety Research vklouisN. Rib,SpecialAssistant THRU:
Advanced Reactor Safety Research FROM:
Frank J. Witt, Nuclear Engineer Fuel Proces:, Systems Standards Branch Office of Standards Development SUBJECT-RADIATION AND PHYSICAL CHEMISTRY OF LWR PRIMARY COOLANT 1.
Introduction A thorough review of the radiation and physical chemistry of LWR primary coolant is urgently needed since an understanding of this field may effect the design or primary coolant chemistry treatment of LWRs. The present safety philosophy used in the preparation of Safety Analysis Reports does not consider the uncovering of the core via steam and gas bubble formation since the redundant Emergency Core Cooling Systems (ECCS) are designed to prevent this happe.'.g in the event of a loss of coolant accident (LOCA). H wever, operator errors can lead to bubble fomation in the reactor vessel.
Since the Three Mile Island accident scenario could happen again, it is prudent to thoroughly evaluate the radiation and physical chemistry of LWR primary coolant so that safeguards can be established to prevent or minimize the hydrogen problem in potential accidents.
It is extremely important that this th> rough evaluation consider the complete reactor system, containment
.d auxilliary buildings to understand what is going on).
2.
Scope a.
Perfom complete systems evaluation on reactor coolant system, containment building and auxilliary building including material balances on:
(a) hydrogen, (b) oxygen, (c) nitrogen, (d) tritium, (e) fission gases, (f) particulate fission products.
(g) particulate activated corrosion prcducts, and (h) dissolved activated corrosion products.
7 O
\\
e y
np' m asz
APR 2 51979 Charles Kelber
-2 b.
Evaluate radiolysis and recombination release and sink rates of the following in reactor coolant system, containment building and auxilliary building:
(1)
HOva H, H 0. H+, OH", 110 2
2 27 2
2 + H* + H O OK" + H 2
OH" + H 02 2 --+ HO*2
- H O 2
H'+H0
-+ OH" + H O 23 2
H+ + O
--+ H O*2 2
overali reaction 2H O @ 2H2+O2 2
The radiolysis product in neutral water is the solvated electron (the H atom in water might conceivably take on a nr might lose a proton, proton to assume the acid form H7 leaving behind a solvated electron, which would be the basic form of H.
(2) Recombination C. I. Hochinadil, Oak Ridge National Laboratory, has a U. S. patent (abcut 1960) covering the re:ombination of to form water provided that there is ar. excess of H7+O2 dTssolved H in an aqueous solution.
In effect the followingeku111briumequationwillproceedtotheleft:
2H O q*
,12+O2 2
However, if the aqueous solution is boiling, the H., on the right side of the above equation will be stripped out of solution into the gaseous phase resulting in a depletion of H concentration on the right side of the above equation.
p If tne H, concentration is depleted below a minimum concentr8 tion, recombination (equilibrium r 1ction to the left) will not occur and decomposition tequilibrium reaction reverses to go to the right) will result.
This is extremely important at the present time at Three Mile Island. Because of the combination of feed and bleed and degassification, the dissolved hydrogen may be reduced below the minimum required (about 2 cc H2 STP/Kg e
e
??4253
Charles Kelber APR 2 51973 coolant) to suppress radiolytic decomposition and the very undesirable reversal of the equilibrium reaction, H O s*. H2+O2 2
will occur leading to the possible femation of another Hydrogen gas bubble in the reactor vessel. Therefore, the dissolved hydrogen concentration must not be permitted to decrease below 2 cc H STP/Kg water.
If necessary, smallquantities(2-15c!H STP/Kg water) of H should beaddedtothereactorcoo!antviathecharginhsystem to levels exceeding 2 cc H STP/Kg water to suppress 2
decomposition.
(3) H0 #dditi "
22 Radiolysis of dilute solutions of H.30, injected into results in a primary coolant containing dissolvec M 2 rapid, completely reproducible rate in equimolar quantities, the overall reaction being H2+H022@HO 2
"The high rate of this reaction indicates that it is probably a chain process. Such a result is racher unusual in chemical kinetics.
It suggests the idea that peroxide undergoes two reactions with radicals, of which one stops the chain and the other carries it on. The reactions which carry on this chain are not difficult to propound.
Because of the similarity of OH to C1', we think of the well known Nernst chain reaction between H2 and C1 and propose that our chain is carried on by the steps:2 2 H O + H+
OH + H 2
H+ + H 0 # N 0 + OH~
22 2
Hydrogen peroxide also shows a chain-stopping reaction that is evidently not possible with C1, since in the caseofCl.,thereactionchainsaremubhlongerandthe rate does hot noticeably decrease with increasing C12 concentration. The reaction OH + H 0
~'"N0+HO{
22 2
T/4254
Charles Kelber APR 2 51979 has all the properties demanded by our chain stopper and in addition had been proposed many years before (Haber, F. and Weiss, Joseph, Proc. Roy. Soc A147, 332 (1934) in connection with other reactions with H 0. The only difficulty now is to dispose of the H0,2 which evidently g
did not react with H 0 to carry on th$ chain, otherwise itsformationwouldko!stopthechain."
(Quotation from "The Radiation Chemistry of Water and Aqueous Solutions,"
Augustine 0. Allen, Senior Chemist, Brookhaven National Laboratory, P. Van Nostrand Company, Inc.,1961, page 77.)
The above radiolysis theory was provan in practice by additions to numerous operational commercial nuclear reactor power plants as reported in EPRI report No. NP-692 " Effects of Hydrogen Peroxide Additions on Shutdown Chemistry Transients at Pressurized Water Reactors,"
Nuclear Water and Waste Technology, S. T. Sawaochka, et.
al., April 1978.
Hydrogen Peroxide was added successfully to cornercial nuclear power olants to study the effect on activated core crud deposits and the resultar c effect on shutdown primary coolant component radiation levels. The following plants were involved in this R & D program:
Turkey Point.3 Indian Point 2 Prairie Island 1 & 2 Point Beach 1 & 2 Fort Calheun bb!ne Yankee H. B. Robinson Kewaunee (4) Nitrogen Radiolysis This radiolysis reaction should be evaluated in detail since it is conceivable that it may have been the most effective mechanism for the collapse of the hydrogen bubble at Three Mile Island. The dissolved nitrogen introduced into the primary coolant may come from two sources:
(a) accumulators - The accumulators on pressure vessels filled with borated water and pressurized with nitrogen gas.
During normal operation each secumulator is isolated from the Reactor Coolant System by two check valves in series. Should the reactor coolant pressure fall below the accumulator pressure, the check valves open and borated water is forced into the reactor coolant system.
rj74255
Charles Kelber APR 2 51979 (b) Boric acid tank - The presence of dissolved nitrogen in injected makeup water could lead to the fomation of NH 0H by radiolysis with dissolved Hg 4
in the primary coolant by the following reactions.
- 2H O @ 2NH OH 3H2+N2 2
3.
Procosed Procram Detemine first hand at Three Mile Island what experimental a.
tests are being conducted to support reactor coolant technology.
Experimental work is being perfomed at Idaho by EG&G as well It at Provo, Utah, by Billings Energy Research Corporation.
is vital that first hand infomation is obtained from K.yne Lanning (Research coordinator of coolant technolegy testing at Harrisburg). Lanning just relieved Bixby (DOE-Idaho) of this responsibility.
Obtain complete primary coolant, containment building and b.
auxilliary building H, N, 0, NH, T, fission product gas, and fission product p$rtibulake an$lysks to perfom essential material balances. Must be done first hand, can't use second and third infomation. It is possible that this infomation is not presently available and additional analysis would be required from laboratories that have perfomed radiochemical analysis (Bettis, Battelle Columbus, Oak Ridge, Idaho National Labs, Savannah River, B L W, etc.).
Work directly with KAPL personnel who have developed applicable c.
KAPL has provided theoretical complex computer programs.
computer analysis of primary coolant radiolysis and recombination KAPL and Los for several urgent Three Mile Island problems.
Alamos Scientific Laboratory have computer codes that are directly applicable to the complex radiolysis and recombination One reactions that are taking place at Three Mile Island.
computer code evaluates the solution of fifteen simultaneous differential equations which has been verified by experimental First hand NRC contact with KAPL personnel is essential data.
to define the problem precisely and to provide latest data This will provide an orderly flow of from Three Mile Island.
information with the minimization of side-tracking or snafo.
7/4256
Charles Kelber APR 2 51979 d.
Consult with Dr. Jay Young, Manufacturing Chemists Association.
~
Washington, DC, who is an expert in practical radiation water and physical chemistry problems. At his present position at Kanufacturing Chemists Association he is also very knowledgable in the explosive and flarrability ranges, and catalysts and conditions that initiate hydrogen-oxygen explosions and deflagration.
e.
Evaluate TMI trends in primary coolant pH, Li, Tritium, dissolved gas, H,
0.,, N, etc., and NH from start of accider.t.
Preliminary analfsis'of hamples indicat$ an increase in pH and which may be accounted for by radiolysis of dissolved N2 H to form NH '
2 3
2Nh N2 + 3H2 3
Must get addition primary coolant samples for proper physical chemical analysis.
4.
Conclusions a.
Task force should be set up to evaluate radiation and physical Chemistry of LWR primary coolant. This task force may recomend upon conclusion of evaluation that LWR radiation and physical chemistry should be factored in the safety models of LOCA Computer programs and should be developed to account for Zirconium-Water reactions and radiolysis.
b.
Imediately establish direct NRC contact with KAPL, and Los Alamos to perform needed Zirconium-water reaction and radiolysis material balances in reactor coolant system, containment building and auxiliary building (Systems approach very important).
Review Safety Analysis Reports cnd Safety Evaluations Reports c.
to recomend where additional information and discussion of reactor coolant radiation and physical chemistry of LWR's is necessary.
d.
One individual should coordinate all activities relating to primary coolant physical chemical and radiation analyses as well as any laboratory testing relating to the above.
WY g Frank J. Witt, Nuclear Engineer Fuel Process Systems Standards Branch Office of Standards Development
?M257 t
0 4
UNITEo STATIE f
- d >.
NUCLEAR REGULATORY CCMMISslON was cfou. o. c. msa j
s, u APR 2 5 B73 NOTE TO: Saul Levine e
Robert Budnitz Thomas Murley FROM-Louis N. Rib SU5 JECT: HYDROGEN REDUCTION METH005 -
SUMMARY
The following infomation was developed and informally transmitted on 4/8/79 for the purpose of making it available on a timely basis.
For the record, it is surcarized below, l.
H radiclysis/ recombination 2
A group of engineers and scientists at KAPL have been providing assistance to us on the TMI-2 incident with the aid of Frank Witt (NRC/SD) who worked at KAPL for a number of years.
Frank and I have posed several questions on which we requested KAPL analyses.
The following preliminary infomation has been received:
(1) Assuming an opening in the Primary Coolant System, leaving the reactor vessel full of water up to the pipe elevation in the Reactor vessel, no dissolved H in the water and decay power 7
at 10 MW(th); what radiolysis may occur?
Response
15 scf/ day H2+O2 generated.
Note: With a minimum concentration of H in the water, hydrolysis is unlikely, hochinad$1(ORNL)hasalong standing pstent in thig area.
5 (2)
Consider primary goolant water spilled in containment ( 10 gal) at 100*F, 10 microcuries/ce; How much radiolysis may occur?
Response: Negligible amount, especially with H in containment 7
atmosphere.
(Other variations of quantity of water and temperature are being considered.)
f n
y\\ e &. p
- oms u
9
/.R0
/.
- 4. *
. APR 2 51979 (3) Containment atmosphere made up of air, water vapor (condensed steam) at 100*F and a high radiation field (30,000 R/hr); what will occur?
Response
Recombination in gas phase of H and 0 to produce wateratanestimatedrateof2bOOscffday.
(4) Same containment condition with a 60 cfm recombiner operating.
Response
1680 scf/ day H removal.
2 2.
H recombination 2
An initial suggestion was made (4/1/79) that the presence or addition of N to the water might promote the fomation of armonia (NH, NH 0H) 4 whickissoluble,asamechanismforremovingH. Thegradual release of H fromsolution,bycautiouslyredu$in bleeding the gas out of the system was suggested (g pressure, then 2
4/4/79) as possibly tue best of the available approaches and the addition of concentrated hydrogen peroxide solution as the next best approach. The background on the use of hydrogen peroxide was mainly obtained from an EPRI report (EPRI-NP-692, Effects of Peroxide Additions on Shutdown Chemistry Transients at PWRs, April 1978).
(1) Provo, Utah experiments I contacted Tom Murley at his home about 11 a.m. on 4/8/79.
We discussed the Provo, Utah, H, solubility tests including the unsuccessful H 0 addition test. Tom was going to discuss 3
atestatrelative$yhightemperature(280*F)withHO 2
2 Thei$ea addition in the presence of a Platinum catalyst.
being that the radiation field in the primary coolant system would be approximated by the platinum catalyst in the Provo test.
(2) Hydrogen Peroxide Addition (H2+H02 2 pr duces water)
I contacted Milt Levenson at the TMI site on 4/8/79.
I infomed him of this EPRI report. He was generally familiar with the technicue but not this specific report which sumarized actual reactor plant experience with this technique. He noted the report number. Milt indicated that this approach was limited in use now because they were short coolant storage space and whatever fluid volume was added, a like amount would have to be removed and stored.
In the meanwhile, they were trying to 774259
-3*
APR 2 51979 detemine the quantity of H, disolved in the primary system coolant by lowering the system pressure by 25 psi increments and looking for a pressure increase resulting from H #U*i"9 out of solution.
2 e
(3) Removal of H by organic absorbers 2
Another hydrogen removal suggestion was offered by Harry Gray (Cal. Tech.).
C. N. Kelber first spoke to Dr. Gray on 4/6/79.
Dr. Gray called back on 4/9/79 with the suggestion that the addition of a high boiling point olifin with a catalyst (such as cobalt cyanide) would absorb the hydrogen at a good rate.
If we wished to pursue this suggestion further, he recermended Jack Halprin at the University of Chicago as an expert on the kinetics of this reaction. 52ndia also suggested an unsaturated hydrocarbon as a hydrogen getter.
Future efforts in hydrogen removal approaches should review the feasibility of this option.
Louis N. Rib, Special Assistant Advanced Reactor Safety Research 7742GO f