ML19199A297
| ML19199A297 | |
| Person / Time | |
|---|---|
| Site: | Crane  |
| Issue date: | 04/05/1979 |
| From: | Murley T NRC OFFICE OF NUCLEAR REGULATORY RESEARCH (RES) |
| To: | Levine S NRC OFFICE OF NUCLEAR REGULATORY RESEARCH (RES) |
| Shared Package | |
| ML19199A301 | List: |
| References | |
| NUDOCS 7904110084 | |
| Download: ML19199A297 (1) | |
Text
.-
s t
I
(
e APR 5 1979 MEMORANDUM FOR: Saul Levine, Director Office of Nuclear Regulatory Research FROM:
Thomas E. Murley, Director Division of Reactor Safety Research Office of Nuclear Regulatory Research
SUBJECT:
SEMISCALETESTSFORTHREEMILEISLAkD(THI) INCIDENT Enclosed are preliminary results from two Semiscale tests which were performed to evaluate a plan of action for securing TMI. These tests evaluated a proposed contingency plan to vent a gas bubble from
~
the upper plenum of the reactor vessel. The results from these tests will be contained in a final EGAG Idaho Report.
odgMd Signed by T. E. Murley Thomas E. Murley, Director Division of Reactor Safety Research Office of huclear Regulatory Researth
Enclosures:
As stated Di s+.ribtiti on Tubject ACRS Chron PDR Cir I&E Operations Center Branch R/F ARosztoczy 10 D:
TEMurley VStello LSTong RJMattson CEJohnson RTedesco WDLanning SFaWe D
WDLanning R/F QL WM g
b}
N
,qp,,h,:,3,,
Rg kk,d, J
ER.W:
.reies,,
WD_L
- bslAWSerki z,,,,,
..CEJ.ohnson.,,
Lh.
,,g, rEMufr ey,,,
--..s,.
.. 04//.H.9
.0M.819...
04Lf[7.9..
04I..F9 04./>/79
.m 7904 1008'(
C.
EXPERIMENTAL SEnUEt!CE AND SICrlIFICANT EVEllTS A description of the initial conditions for the TNI plant. response test and a table of significant events follows.
Also includad 15 a description of the significant sentscale system configuration or operating conditions that may not be typical of the THI plant.
Initial Test Conditionr a.
Nitrogen bubble inttially established at a level of 43 in above hot lag pipe upper invert (about 0.5 ft ).
The elevation o f the 3
bubble is higher than expected for the full scale plant.
b.
No secondary side ater was added to the steam generater.
This lack of water would cause a lesser amount of energy to be trans ferred h *the primary fluid.
k Pressurizer steam done was established at 30% of the pressurfier vo l urie.
c.
d.
Leak rate of the system was estabitsbed since the flow out the simulated pressurizer relief valve now was of the sane regnitude as possible leak rate.
System was heated to 410'K utilizing the core and an initial pressisre e.
of 7.24 kPa was established.
[ O 4}; Q n,
p,,,,
.4 h f 'l q(
~ "
P, ] ['] ]
],] C,'[. b
.d g/pn :iriq! ].;l/; ci U
a ut d Y %.j,; '3 ! i
Initial Test Condt tions (Contd.)
f.
Purps were coasted down and an initial pcwer of 7.64 kW was established.
This power was about 2 kW above the scaled value to allow for annroy losses to the structure.
,e)
TABLE OF SIGNIFICANT EVENTS Time Event 0
Experiment initiated (pressurizer steam bubbic venting started Initial clad temperature - 420*K (295"F)
Initial core outlet water tenperature - 411*K (200*F)
Initial pressurizer pressure - 7.6 MPs (1100 pst)
Inf tial inert gas / water interface at 1.35 m above (53.15 in.)
cold leg centerline Initial heat flux 4.1 kW/m2 (1300 Stu/hr-ft2) 0 - 1000 see Nitrogen bubble in upper head expands downward, pressurfrer water level rises as steam is formed and vented.
Pressure falls, temperatures rise slig.itly.
1000 see Pressurtzer water level reaches vent connection Clad temperature - 435*K (323*F)
Core outlet water temperature - 420*K (295'r)
Pressurtzer pressure - 4.7 MPa (680 pst)
Gas / water interface - 0.36 m (14.17 in.) shove cnid Icy centeri f ne.
1000 - 3800 sec Nitrogen bubble continues +.u.. ward expans fon, pushing water out pressurizer vent connection.
Temperatures centinne increasing, pressures fall.
3800 see Nitrogen / water interface reaches top of hot leg opening 0.25 m (9.84 in.) above cold leg centerifne Ciad temperature - 442*K (336'F)
Core outlet fluid temperature - 427'K (300"F)
Pressurizer pressure - 2.6 HPa (380 psi) l'} O!"/
TALLE Of W. iHIFICANT EVENTS (Contd.)
ilme Event 3800 - 5000 see Nitrogen expands into loop pfptng.
Stear generator drains through co1J 1eg and downcomer forcino cool water into lower part of core.
Clad temperatures generally decrease as well as pressure.
5000 see Steam generator tubes empty completely.
Peak clad temperature - 456'K
( 361 *F)
Core outlet fluid temperature - 444'K (319'f)
Pressure - 2.2 MPa (319 psi) 5000 - 7000 see Clad temperatures increase as core flow stagnates, fluid temperatures rise as pressure continues to fall.
7000 see Fluid temperature reaches saturation and bult boilina begins in core.
Core outlet fluid temperature - 452*t.
(364"r)
Peak clad temperature - 456*K (361"F)
Pressure - 0.8 HPa (116 psi) 7000 - 9000 see Low void fraction fluid rises in core, clad terperature decrease, fluid temperatures remain at saturation as pressure falls.
9000 see Test Shutdown.
Hitrogen has expanded into steam genera tor but pressurirer appears to still be filled with ifquid or very low void fraction fluid.
Pressure - 0.34 HPa (49 psia)
Clad temperature - 4?6'K (307'r)
Fluid temperature - 423*K (302*F)
Core bulk boiling occurred when saturation condit' ions were reached.
Max clad temperature - 456*K (361*F) 13 ~ C28
In several instances the Semiscale syster configuration or conditions were not typical of the THI Plant.
operating instances include:
The nost significant (1)
Potential for structures to provide excessive cooling of th e
fl ui d.
(2)
Lack of sinulation of the FWR vent valves.
(3)
Core elevation effects.
(4)
Possible atypicality of power operated relief valve flo w
due to et'fects of sfre.
(5)
Use of charging punps to account for leakage from purp seal (6) s.
Lack of steam generator secondary water.
The st uctures in the Semiscale Mod-3 system have excessiv e surfac e area which will cause atypical energy transfer during the cour truperature transient.
se of a flutil During the simulation of a pressurf 7et relief transient the structures will absorb excess energy which would tend to increase the depressurization rate and prnvide cool er water to the core region.
An attempt was made to provide more typical fluid condition by increasing the core power level by about 35% abov s
e the scala value which was arrived at by determining the rate of energy trans fe r to the vessel structure and increasing the core power appropriately.
The Seniscale system cannot provide a particularly good si mulation of vent valve actuation during a pressurfrer relief drain t est because of elevation differences between the hot and cold legs flowaver, this inability does not strongly influence the Seniscale test resuit s since the influence of the vent valves on the Till plant cold l eg behavfor is limited due to the cold leg geonetry.
No adverse effects on tes t results are expected in Semiscale pressurizer relief valve drain testin g because of the lack of a vent valve simulation.
1L 3 ~ 0 ;3 ()
A particulerfy e: 1gni fic n nt dif fer ence lirtwecn the 'e-!**~7-
'lIf-t and the T711 Plant is related to the location of the top of the enre relative to the vessel norries.
In the THI plant the top of the core SS appreminately at the elevetten o f the vees.1 ness 1.e.
In eh. s ea t e c = 1 =
systam, however. the top of the core is aprrou tena t ely 105 cri beneath
'0".!*1"U.* U L'*J,"*"",1**,.r*f '",'0*;; "e**e~4 7Te 0 5 % "",*!*.;,I! l*,
the TN! plant than in the Semiscale Mod-3 system.
Flow through the Seniscale pressurizer relief valve sirulation is not expected to completely duplicate, on a scaled basis, that which nicht occur in the TitI Plant.
Critical flow through small ori fires (a.g., the 0.030 in. diameter Semiscale flow area) has been shown to be di f ferent from that exper!*nced in 2-in. diameter pipes, so that vent flow /prestore relief characteristic:. might reasonably be expected to dif fer somewhat between the two facilities.
Because of considerable leakage of pump seals (and other miscellancous The small leaks) it was necessary to provide nakeup liquid to the syst em.
HP!S punp was run for brief periods at fixed intervals 1.hronohout the tee t to supply the additional liquid necessary to account for the pump s eal leakage rr.te.
Although the makeup rate was srall compared to tha t"tal discharge rate through the simulated pressurizer relicf vsive, t he amount of HPIS liquid injected into the systevn over the dur ntiori o r the test was a substantial amount, Thus, considerable addit ional u.
.I te the l',15 was added to the primary systen liquid inventory through uso n' As a result, the cort thermal response m6y have been le'.< 5evere pump.
than would have otherwise occurred.
13'040
-. - ~
The secondary side of the steam generator was dry, f.e., no motive fluid for heat transfer.
This condition mininf red the influence or secondary heat trans fer on the enurse of the PPV trans tent in Semist nie.
If the temperature on the secondary side of the steam generatnr is lowsr than the prinary systen temperature the subcoolina on the pump suction leg w+11 be increased, thereby increasing the subcooling to the core.
Conversely, if the secondary side temperature is higher than the primary side tenperature the pump suction density will be reduca<1 t hereby reducina the tore inlet subcooling.
For a PP.V transient the secondary side terperature should be less than or equal to the primary sida temperature.
Based on prevf ous experiments conducted in Semisenle, stear gen rater beat transfer is not expected to have a signi ficant influence on t he tir as s ur 17 er relief draf n tests.
13 011
D.
PRELIMINARY ASSESSMENT OF THE TH1 PLANT RESPONSE DURING VENTING Arealysis Simplified calculations were performad to evaluate expected response of Till to the PRV release mode of depressuriration.
Fi rs t, the hen tup rate of the core fluid was calculated to be 400'F/hr if no circulation occurred (heating the core liquid volume only) and 170'F/br if the total reactor vessel fluid volume were to be heated.
(Neither calculation included fuel or metal mass heat capacity).
From this it was concluded that makeup should be provided to assure core coverage as heat is removed by steaming.
About (30 gpm was calculated as the required rate to meet the steam generation needs.
The expansion of the (then assumed) 1500 ft J of gas to fill the hot leg, steam generator, and pressurf rer to reach tha point of gas venting and nore rapid depressurization would reach this point at about 300 psi.
At assumed 11guld relief rate of 600 opn and steam rate of 110,000 flhr, this was calculated to take about one hour, The Semiscale system depres.surized slower; reaching ahnut 350 psi in one hour with a smaller relative gas volume.
Integrating the high velocity " gas" relief showed that the hot liquid in the pressurizer flashed to steam and separated yielding a larger total volume of stenn to be relieved prior to the tire of liquid relief.
The data also showed that the pressuri.ter and surge line renef ned liquid full, thus not making that volume available for gas expansion.
These features are being added to a more cenplex model.
Initial indications are thet e reasonable description of the pressure transf ent and ~nlumn changa will result from this model, and it should be app 1feable to THI, n
q <>
E.
CONCLUS!0!15 (1)
The Seniscale system can be depressurized via the proposed method to a level at which the RilR pumps can be activated and used to renove residual heat from the core, (2)
In the Seniscale systen noncondensible gas did not vent easily or unifo' rmly with the proposed method.
The noncondensthic gas bubble entered the hot leg at approximately 3000 seconds.
(3)
Core unenvery in the Sentscale facility did not occur untf1 a fter a point in tlne at which the RHR pumps could have been activaterf ff des f red (thus preventing core uncovery),
(4)
The Seniscale results sugges t that if ECC fluid is in,lected into the systen at a rate comparable to that at which the s ytten it being vented, signi ficant beneff ts in the overall systen resronse and core cooling may be realized.
(5)
The heater rods in the Sentscale test remained in a mode of good cooling during the proposed trans font and rod temperature riset unre minimal, (6)
Depressurization from 1050 psf a to 49 psfa was accompfIshed in the Semiscale test in approxinately 3 hours3.472222e-5 days <br />8.333333e-4 hours <br />4.960317e-6 weeks <br />1.1415e-6 months <br />.
Boiling in the cora did not occur until approxtnately 6000 seconds after the vent relief transient was initiated.
j['] ~()4J'l
_. : = :... r - _ _ - -- -.
w TO:
W. D. LANNING, MAC
.. RECEIVED
_-.e /
1979 APR 4
- 41121
~
u.s. re : F ' PrC'?. ANkk M AIL & hid5LLN BRAEH
..i v.
PRELIMINARY TEST RESULTS I.C 5! v ;
SECOND SEMISCALE RELIEF VALVE VENTING TEST FROH THREE MILE ISLAND CONDITIONS
?
I April 2.1979 t
[
4 t
- -- - ~......
2 - -
A.
SWEARY Rasults from the initial Semiscala Mod-3 test conducted at conditions sinflar to the Three Mile Islartd (THI) plant provided afgniffcant insight into the growth and movement of an upper plenum bubble during continuous pressurizar reltaf valva (PRV) cparation. On ApM1 1.1979 NRC management requestad a second test to be conducted at conditions stattar to the initisi test except for a larger gas bubble volume and injection of a scaled enount of coolant from the HPIS.
The test was conducted between 1 s16 a.m. and 4:35 a.n. on April 2.1979.
The test appears to be a s9ccess although some loss of data was expertenced as a result of a malfunction in the data acquisition system. A preifminary evaluation of results indicatas the test behavior was similar to the inittai Injection of coolant from the HPI$ maintained cooling in the core tes t.
Pressurizar relief flow was terminated and recharging of the systaa region.
with a h?gh HPIS flow was intuated een the systen pressura reached about 270 psfa. The system was returned to a stable state where the primary coolant punp could be used to circulate flow within the prfrary system about one hour after termination of pressurtzer relief flow.
e 0
e 2
13 ~ 'Mb
l_
-b7
-7__
- 1 8.
INITIAL TEST CONDITZ6*.S AND SEQUENCE OF EVENTS A description of the initial conditions for the second THI plant response test and a table of significant avant follows. A table of the initial operating conditions Tabla I, and a description of significant Sesiscale operating conditions are also included.
A description of the systes configuration that say not be typical are documented in the report on the initial Semiscale pressurizar ra11sf tast.
INITIAL TEST CONDITIONS a)
A Helium bubble was initially attablished at a laval of 61 Cs above 3
the hot leg pips upper invert (equivalent to 0.6 ft of Halium).
Helium was chosen for this test since its properties are relatively close to Hydrogen.
The bubble size was chosen to provida a larger bubble voltaa than used on the initial tast, b)
The diamatar of the orifice used to staulate the relief valve was 0.091 ca. This diameter wu slightly larger than the diameter used in the initial test (0.079 cm) and was adjusted based on results from the initial test to provide a steams flow more typical of the spect(f ed relief valve flow for the THI plant.
c)
The HP!S system was initiated at ruptura at a constant rate of 12.6 al/s (0.20 gpe). This rata was scaled to an average injection rate which was being considered for the THI plant over a pressure rangs of 300 to 1000 psi.
Additional injection was included to maka up for the normal leakage from the Semiscale systas.
d)
The steasa generator in the same loop as the pressurizer had a water level aquivalent to 2/3 the total tube elevation. The initial steam generator fluid temperature wu 416 K.
The second steam generator was run with a dry secondary sida due to difficulty in detarnining a consistant sat of initis1 conditions.
3 13 ~ 0 it.
?
- .+.--..-
~
.g..
- q.
e)
An crifice was included in the' intact loop hot leg to cos:pensato for -
the difference in size, and hence elevation of the top invert (Figure 1), '
of the intact and broken loop pipes. This elevation difference was belteved to have caused the intact and broken loops to behave differently
,.. y. y.
in the inttial test.
s..
. y.. -
f)
The core power was held constant at 21.5 kW throughout the tast. This
is substantially above the 3.9 kW 1evel necessary to simulate tha known THI conditions. The additional 17.5 kW was included 2 nake up for ambient and structural heat losses which are in excess of the THI heat losses. The sectscala heat losses were determined expertmantally using initial test conditons.
It is expected this excess power would cause the Samiscale core rod and flutd temperatures to be conservatively 5 [I
.. 9. s___...,.,..
hf gh compared to the TMI plant.
.";_..m.
g)
The initial pressurizar steam dome was about 50% of the pressurtzer volume.
h)
The system was heated to 416 K utt11xing the core and an initial prassure -
of 7.24 HPs was established.
The pur.ps were coasted down aad the initial power was [stablished.
1)
I t
. I*
t 9
.f
.. t ?...
^
3 a
^
~ ',
.r 4
- 4. 1 n "(
J a
J
1
,mT Stata Ves2al hasarer Ossaraxar Scao.tn
)
6 ettera' Car-1 J1 1.3 b"rD.
Ob s G.'c.e p
_\\
Lt.
a g_________.
Brekan Lamp Hor La3 4;S3"I.D.)
Inta.st Lp Mdt La3 (2.62.b~r n)
/3 Figure 1 Schematic of the intact and broken loop hot legs.
9 4
5 J73 f
I TABLE OF SIGNTFICANT EVENTS Time Event i
Experiment inttf ated (pressurizar staan bubble venting starts (
Initial peak clad temperature - 442 K (336 F)
Initial core outlet water temperature - 436 K (325 F)
Initial pressurtzer pressure - 7.3 MPa (1058 psi)
Initial inert gas / water interface et 0.97 m above (36.4 in.)
cold les centerifns Initial enre power - 21.5 kW
- 200 s Half ue bubble in upper head expands downward, pressurizar water level rises as staan is formed and vented. Pressure falls, temperatures rise slightly.
40 s Hellum / water interfaca reaches top of hot lag opening 0.25 m (9.84 in.) abova cold leg cantarlina Clad temperatura - 442 K (336 F)
Core outlet fluid temperature - 435 K (325 F)
Pressurizar pressure - 6.0 HPa (870 pst)
- @3 - 700 s Helium continues to expand pushing water into pressurizar and filling intact and broken loop steun generators with gas.
100 s Pressurizer watar level reaches vant connection Clad temperature - 465 K (378 F)
Core out1st water temperature - 465 K (378 F)
Pressurizar pressure - 4.45 MPa (640 pst) 4 4
- ----~
A TABLE OF _SIGNIFICANT EVEMTS (Contd.)
e, fina Event
)d - 2200 s Helium bubble continues downward expansfon, pushing water out pressurizar vant connection.
Teeparatures continue increasing, pressures fall.
e
!00 - 2800 s Intact and bmkan loop steam ganarators drafn through. cold legs and downcomor forcing cool water into lower part of core.
Clad temperatures ganara11y decrassa as well as pressurg 100 - 4800 s Data acquisition system malfunctioned during this period.
All data was lost.
During this time the steam ganarators complete) dratnad but soca water ramalnad in the bottom of the hot leg piping.
10 0 s Data acquisition systate back on lina.
Recording of data continued.
100 - 5800 s Temperatures continue to decrease as system pressure decreases, 100 - 6700 s Clad temperatures increase as core flow stagnatas. flufd temperatures rise as pressura continues to fall, f00 s Test terminated when vent valve was closed. HPIS injection rag increased to 0.63 gpco causing system pressure increase.
100 s System pressur.s peaked at 1050 psia. At maximum system pressur the broken loop pump was slowly brought up to speed and initial conditions wara re-estabitshed.
Helium was forced out of broksi loop but collected in intact loop.
It appeared impossible to resturt and achiava full flow from both pursps sinca Helium wouli collect in one pump when the other pump was being startad, preventing the second pump from establishing a net positive suction head.
7 13 000
4 TABLE OF SIGN!rIC/JiT EVENTS (Contd.)
Time Event 12.000 s Test shutoown O
d 9
4 e
u 1.
J.. o
,9 LIST of INITIAL CONDITIONS TABLE I CONDITIONS FOR SEMISCALE TEST 3MIf PRIMARY SYSTEM 7.29 HPa (1058. psia)
Pressure Tenparature - 416. K (289.4 F)
PRE 55UR17ER Pressure - 7.29 HPa (1058. psfa)
Temperatura - 560 K (548.6 F)
Top 400 K (260.6 F)
Sottom Leve1 - 50%
Heaters - Turned off after pressurf ter conditions established STEAM GENERATORS Intact Loop Level - 0. (Dry Secondrey)
Broken Loop Laval - 1/3 - 2/3 full Temperature - 416 K (290 F)
CORE POUER 3.9 kW Decay Heat 17.6 kW Heat Losses 21.5 kW Total l..,.rry' a
g
9 TABLE I - LIST OF IllITIAL CONDITIONS (Contd.)
UPPER PLERUM / UPPER HEAD 4
3 Hel1 m Volume - 1.67x10 1 567 os (0.59 1 0.02 ft )
3 3
W;ter volume above hot leg nozzles - 3400 cm (0.16 ft )
NIGH PP.E55 JRC It!JECTION SUMP Locaf/fon - Intact Loop Cold Lag
~12.6 nl/s (0.20 gpm)
Rate (constant)
CORE ftUID TEMPEPfJURE DISTRIBU_ TION 436 K (325.4 F)
TDP 430 K (314.6 F)
MIDDLE 420 K (296.6 F)
BOTTOM LOOP COOLAMT PUMP 5 Standby conditions (power off) 9 10
~
(4 C.
CONCLUS10M
- 1) Tha' additional gas volume and the enlarsonant of the size of the scaled reltaf valve area influenced systs response somewhat but did not change the ganara1 conclusion that tt.6 systos coald be depressurized to a level where the RHR system could be activated without significantly shocking the system.
- 2) The usa of the HPIS pump during the early part of the experiment caused cold watar to be supplied to the core. This additional cob **ter maintained high density fluid in the core and contributed to
..-2f aad i
effectiva cooling of the rods.
3)
Based on a compartson of subcooled flow rates the sinulated relfaf valve flow in Sesstscale is 70 to 90 parcent of the subcooled flow calculated for the THI plant valyss. Tht lower flow may be due to the presence of noncondensable gases or to orifice size.
The possibility of reducing.
the subccolad reltaf valve flow wheit noncondansable gases are present should be considered in calculations performed for the THI valves.
4)
The need for higher cora powers to make up for additional'systaa heat losses resulted in higher rud temperaturas and fluid temperatures in Semiscale than would be expected in the THI plent.
These higher temperatures would influence the depressurization rata late in the tast when the pressure reached saturation levels.
5)
After cocplation of the system depressurization, attespts wara nada to repressurize the system and re-estabitsh initial operating conditions with both Samtscale pumps operating.
It was dfscovered that when either one of the punps was started gas was forced through the loops f ato the other loop pump and the remaining pump could not be startad and maintafnad at full flow because a nat positive suction head could not be established.
Therefore it appears that once Helium is present in tha seatscale loop piping it is unlikely that the gas can be removed from the system by operating both pumps sinultannously.
g-N4 11
?.
3. -.-,..
~
4 6)
In semiscale Test 1HII. differences between the intact and broken loop pfpe sizes allowed tha expanding Helfus bubbla to reach the top invert of the larger intact loop pf pa carly in the transtant. This produca8 prafarential dispersfon of the gas into the intact loop, which was fait to be atypical of a PWR.
In Ta-t 3H!2 an ort flee was ad.iad to the intact loop hot leg nozzla to produca squal top invert stavations in the two hot legs.
This geometry change produced more untform gas disparston into the hot legs and an incraased venting of the gas out tha pressurizar which is believed to be more typical of the TH!
plana.
7)
Semiscala results are definitaly influenced by such scaling distertions as geometric size, one-dtmansionality, structural heat transfer area.
and slavation influances. Caution should be exercised in the inter.
pratation and extrapolation of these results -to any other size facility.
4 6
.