ML19350B592
ML19350B592 | |
Person / Time | |
---|---|
Site: | Sequoyah, McGuire ![]() |
Issue date: | 01/09/1981 |
From: | Strehlow R STREHLOW, R.A. |
To: | |
Shared Package | |
ML19247D089 | List: |
References | |
NUDOCS 8103230084 | |
Download: ML19350B592 (9) | |
Text
s
\
ti i isir
\/ . . .:. t,>
. lC. A l3 * .1 l
- -s
.s e. .-al%.
Evaluation of the Glow Plug
. /-
Igniter Concept for use in the
/ . Sequoyah Nuclear Plant Prepared for Mr. James Milhoan, P. E.
Office of Policy Evaluation
=
Nuclear Regulatory Cormission Washington, D. C. 20555 Y
Prepared by Roger A. Strehlow Consultant January 9, 1981 n103230084
Overall Evaluation In my opinion, a well designed and maintained glow plug igniter system which is energized only for testing or during an event which has the potential of generating hydrogen is an effective way to protect the Sequoyah nuclear plants from the possibilityofbreachingthecontain[entvesselduetothe inadvertent combustion of accumulated hydrogen. Furthermore, it 4/ is my opinion that the implementation of this glow plug igniter technique will have no negative effects on overall safety in such a nuclear plant. I base this opinion on the following information that was supplied to me by the Nuclear Regulatory Commission:
Tennessee Valley Authority, Sequoyah Nuclear Plant Core Degradation Program, Volume I, Hydrogen Study, September 11, 1980 News Release No.80-159, USNRC, Septe=ber 11, 1980 News Release No.80-163, USNRC, Septe=ber 18, 1980.
Safety evaluation rep :t related to the operation Sequoyah Nuclear Plant, Units 1 and 2, Docket No. 50-327 and 50-328, Tend 6ssee Valley Authority, NUREG-00ll, Supplement No. 4, September,1980 Memorandum for: ACRS members , from: J. C. Mark, Subject Notes on hydrogen burn with igniters , December 4, 1980.
Memorandum to: ACRS members, from: 'H. Etherington, Subject Memorandum P. G. She= won to' ACRS members :
" Quantity of H at TMI-2 and source . "
2 Tennessee Valley Authority, Sequoyah Nuclear Plant Core Degradation Program, Volume 2, Report on the Safety
~ Evaluation of the Distributed Ignition System, December 15,-1980 Tennessee Valley Authority, Sequoyah Nuclear Plant, Research Program on Hydrogen Combustion and Control, Quarterly Progress Report, December 15, 1980 P
l
Draft copy of Supplement No. 4 to the Safety evaluation report by the Office of Nuclear Reactor Regulation, U. S. Nuclear Regulatory Commission in the matter of Tennessee Valley Authority Sequoyah Nuclear Plant Units 1 and 2, Docke t No 's . 50-327 and 50-328, undated.
Attendance at the ACES subecmmittee meeting held in Washington, D. C. on January 6, 1981.
A meeting with Mr. Tinkler and Mr. Butler of the NRC staf f on the morning of January 7, 1981.
Other open literature references which helped me form =y opinion will be referenced in the detailed supporting statement that follows. I also have opinions concerning the dynamics of a
-f .
combustion explosion in a Sequoyah type containment and new research and data accumulation efforts which would be necessary to strengthen and quantify the justification for using glow plug igniters as the only hydrogen control technique. These will also be disucssed below.
=
Effectiveness of the Glow Plugs.
The Singleton Lab., Fenwal and LLNL tests have shcwn the glow plugs that are being considered for the Sequoyah plant to be very r
effeu'ive igniters down to 5% hydrogen even in the presence of 15% dry steam. Thus, in a real accident we now knew that the igniters would initiate a partial burn at 5% H2 in the CV. Further-more, the Singleton lab tests - show that. in a small vessel even with 3.5% H2 present initially a five minute " burn" reduces the H 3 concentration to about 0.1%. This is very encouraging because it shows that a hot glow plug will act as an H2 scavenger even outside the flammability linit for pward propagation of about 4% H 2 . Furthermore, sparks of. the type 'that undoubtedly initiated the Three Idle Island burn are not effective at such a low hydrogen
- ~
concentration.
It is important to note that glow-plug initiated burns will be much less dangerous than spark initiated burns. This is because between 4-8% hydrogen in air burns with a very lazy upward propagating flame which spreads at a maximum half angle of 1
about 20 degrees and extinguishes when it reaches the top of the vessel. This means two things: 1) the pressure rise will be minimal for such a burn, and 2) the hot product gases will be
, confined to this cone-shaped volume and subsequently will spread along the ceiling. In other words, the flame will not contact and therefore not heat most equipment that is in the containment vessel. . Also, if the rate of hydrogen generation were slow so that the fans produced a rather uniform hydrogen concentration, I
the burn would be almost continuous once the hydrogen content reached 4-5%.
I On the other hand, a fast leak which caused a localized higher concentration of hydrogen would also not be dangerous when ignited by the glow plug.
- This is ba,cause glow plugs strategically
~
placed ab'ove potential hydrogen sources would ignite a high hydrogen concentration pl 2 on contact and only a localized high temperature burn would occur. It is well-documented (Cubbage and
'~
Marshall,1972) that such a partial burn yields a pressure rise in a vessel which is proportional to the energy released by the localized burn (Joules) divided by the total volume of the vessel (m ). Thus, a small localized burn cannot cause a really large pressure rise.
In my opinion, properly located and functioning glow plug igniters would reduce'the probability of a burn leading to a La -
~ -
transition to detonation to virtually zero. This is because the very weak flames produced by a 4-5% hydrogen burn cannot generate significant pressure waves or significant flow velocities ahead of the flame. This means that the mechanisms that lead to flame acceleration do not exist under these conditions. In other words, the weak 4-5% hydrogen flames will remain weak irrespective ol' the environment that they-encounter.
Combustion dynamics in the Sequoyah concainment vessel.
'/ .
'The.Sequoyah containment contains three main ecmpartments:
- 1) the upper compartnent, 2) the lower compartment, and 3) the ice condenser. The upper and 1cwer compartments both have a rather low length-to-diameter (L/D) ratio and therefore if they
.could be treated as independent vessels they would be capable of
' supporting-only a simple overi pressure explosion. The containment at Three Mile Island was essentially of this type and that is what happened there when the hydrogen concentration reached about 8% .
In suchLa case, the flame prcpagates slowly enough such that the y
pressure is_ rel'atively uniform spatirlly in the vessel during the
+ burn and simply rises with time. (approximately as a cubic of i' time; see -Bradley and Mitcheson, '1978 a, b) . This is.true even if there is some: acceleration due to turbulence generation.. Under these conditions, there are ~ essentially no pressure waves generated.
Note 'JaatD at TMI the transit . time of a sound wave from top to
-b'ottom~ to top is L aboutDO .2 seconds and the. burn took about 10 seconds, i
- . 1Unfortunately,71n the Sequoyah configuration the upper and
- , ' lower
- ccmpartments are '. lot independent but are connected by the M M u.ft . . x.
\
, o
- ice condenser. In my opinion, this is a very dangerous configu-ration because it would generate pressure waves which could possibly lead to local over pressures that could breach the containment. This is because the ice condencer contains hundreds of tubes (the spaces between the baskets) which have a very large L/D and which could cause significant flame acceleration and possibly even transition to detonation. This mechanism has been adequately documented by Urtiew et al (1965, 1967) and could occur after primary ignition at or above 8% in either the lower or upper 1 .
compartment. The sequence, without detonation, is as follows:
ignition in one compartment causes a slow pressure rise and starts a flow through the condenser, pressurizing the second compartment.
The flame then gets into the ice condenser at some location and accelerates- in this turbulent -flow causing large turbulent jets to enter the second compartment. Once the flame reaches the=second compartment, . it is already pre-pressurized and the burning velocity is now so large that' combustion in this compartment produces
\ pressures that are up to a factor of 2-4 above the calculated r
maximum adiabatic constant volume pressure (Heinrich, 1974).
There is another mere recently discovered combustion. dynamics possibility.- Knystantas et al (1979) have shown that large scale -
eddy folding of hot combustion products into an already turbulent i
' jet of' reactants can produce shockless initiation of detonation.
Here the mechanism is that radicals in the product gases trigger combustion reactions in the mixing volume and as the system explodes the pressure increase augments the combustion process.
This coupled ~ augmentation eventually culminates in a detonation wave. For hydrocarbon-air mixtures, the critical eddy size is t . . . _ . __
~
large, about three meters in diameter. Note that at the exit of the ice condensus, conditions would be right for the formation of such a large mixing region. The required eddy size for hydrogen-air is not known but it would probably be smaller than the critical size for a hydrocarbon-air mixture.
Thus,.in my opinicn, the combustion dynamics of an explosien in which a continuous flame is able to propagate (i.e., in a mixture containing greater than 8% H3 ) is a very dangercus situ-
-/ ation and would have the potential to breach the containment vessel. We know that glow plugs have been shown to yield partial burns when the flame is lazy and not dangerous. This, coupled
-with the vulnerability of the facility to a dynamic ccmbustion explosion, is one more point in favor of using glow plug igniters to-protect the containment vessel from the adverse censequences of an accidental spark-ignited burn..
Research Needs, c
Glow plug testing should be continued. Specifically, I agree with the LLNL recommendations for further work that was presented at the AC-3S subcc=mittee meeting of. January 6, 1981. I would also like to see some continuous burn tests at ccncentration less than 4% to' determine how rapidly a glow plug will scavenge hydrogen at these low concentrations.- In these tests, the effect of fan-induced flow across the plug should also be investigated.
, Even,though I feel that the gicw plugs will virtually elini-nate the possibility of detonation in the containment vessel, I stilll feel that some work .on' detonation limits should .be performed.
-4 .
,. . ~
I do not believe the 131 figure that is in the reports. I feel that the limit is much lower, possibly 12%. At any rate, this uncertainty can be relativel, aceily answered by a few rather simple tests that should be performed.
-t ,
W 4
m M
- y, \ y_ - - -
~
References Bradley, D. and Mitcheson, A. (19 7 8a) , Comb and Flame 32, pp. 221-23d.
Bradley, D. and Mitcheson, A. (1978b) , Comb and Flame 32, pp. 237-255..
./ Cubbage, P. A. and Marshall, M. R. (1971) , " Pressures generated 4 in combustion chambers by the ignition of air gas mixtures",
I . Chem E. Syraposium Series #33, Inst. of Chemical Engineers, London, pp. 24-31.
4 Heinrich, H.-J. (1974), " Bum Ablanf von Gas explosionen in mit
" Rohleitungen verbundenen Behadtern", BAM Berichte #28, Berlin ( August 1974) .
Knystantas,-R.; Lee, J. H.; Guirao, C.; Freuklach, M.; and Wagner, H. G., " Direct Initiation of detonation by a hot-turbulent gas jet",17th ; Symposium (international) ' on -
Combustion', . The Combustion Institute , .oittsburgh; Pa. ,
(pages unknown).
Utriew, P.A. and Oppenheim, A. K. (1967), "Detonaticn Initiation by Shock Merging",11th Symposium (International) on Combustion, The Combustion Institute, Pittsburgh, Pa. ,
pp. 665-676. ^~
Utriew, P. A.,JLaderman, A. J. and Opcenheim, A. K. (1965),
" Dynamics of the Generation of Pressure Waves by Accelerating _'
, Flames", 10th Symposium ~(International) on Combustion, The l Combustion . Institute, -Pittsburgh, Pa. , pp. 797-804 r
t e
. . I-f
. _ _ , e- _ .