ML20052E226
| ML20052E226 | |
| Person / Time | |
|---|---|
| Site: | Millstone  |
| Issue date: | 04/30/1982 |
| From: | NORTHEAST NUCLEAR ENERGY CO. |
| To: | |
| Shared Package | |
| ML20052E224 | List: |
| References | |
| TASK-A-01, TASK-A-1, TASK-OR NUDOCS 8205100211 | |
| Download: ML20052E226 (47) | |
Text
o G
Docket No. 50-245 Attachment No. 1 Millstone Nuclear Power Station, Unit No. 1 Isolation Condenser System Water Hammer Events i
i April, 1982 8205100211 820427 PDR ADOCK 05000;t45 P
PDR l
1 Originally Transmitted to the NRC Staff on September 23, 1981 l
Responses to NRC Questions Concerning August 10, 1981 Transient at Millstone Unit No. 1 Question No. 1:
Why didn't the turbine by pass valve to the main condenser throttle sufficiently to prevent low reactor pressure and auto closure of MSIVs; i.e. reactor scram without by-pass?
Response
Following a typical scram, the reactor pressure control system throttles the turbine control valves closed to maintain reactor pressure at the set point as reactor power decreases. When the turbine is tripped, the turbine bypass valves open and assume this pressure control function.
Af ter the operator verifies that the pressure control system is controlling reactor pressure, he procedurally bypasses the 880 psig isolation by placing the reactor mode switch to " Start Up".
After the August 10, 1981 scram, the operator tripped the turbine very quickly (i.e. six seconds af ter initiation of reactor scram) and the re-actor pressure control system demanded that greater than 5 turbine bypass valves open.
(Enough to bypass 55% of rated reactor steam flow). While the turbine bypass valves were opening, reactor power and steam generation were decreasing and a mismatch between bypass valve position and steam flow was created long enough to allow pressure to sag and trip the 880 psig pressure switches.
Investigation reveals that the 880 psig pressure switches are set at 890-895 psig as required. This allows only about 80 psig between the normal steam pressure at the turbine inlet and the 880 trip setpoint.
Further investigation and testing by G.E. after a subsequent unrelated l
reactor scram and isolation on September 14, 1981 revealed that the l
bypass valves control system was not properly adjusted. Although it was l
initially believed that the prompt tripping of the turbine after the scram caused the problem of poor pressure control, it is now known that the hydraulic controller settings were too slow to allow proper response.
The Moog valve (electric to hydraulic transducer) settings were found to stroke the bypass valves open in 7 seconds and closed in 7.5 seconds.
l These settings would open the bypass valves fast enough to avoid a I
pressure increase on a turbit 9 trip, but they would not close fast enough to prevent the 880 psig 3.v'
- ica from being reached.
i l
i The Moog valve has at e ec-and tested as recommended by G.E. to 5 seconds to open ano e sm.
to close the bypass valves.
These settings correspond to the otaginal plant startup test data.
While MSIV closure complicated the scram recovery, it must be reminded i
that Millstone Unit No. 1 is designed to handle MSIV closure from 100%
power. In fact, this has occurred twice at Millstone Unit No. I with no l
adverse consequences.
1 Additionally, closure of the MSIVs is not relevant to the water hammer event at Millstone Unit No. 1.
Although MSIV closure caused high reactor pressure which in turn initiated the isolation condenser (IC) system, high reactor water levels due to the locking up of the feedwater regulator valves led to the water hammer event.
Initiation of the IC is not required to cause a water hammer in that system since both steem side containment isolation valves are normally open. The Decembei l',
1979 water hammer event occurred without initiation of the IC.
Note:
A Technical Specification change request to lower the 800 psig setpoint to 825 psig, which would be more consistent with other BWRs, is presently being evaluated.
l l
1 i
5 i
l
[
\\
l t
a
Question No. 2:
Explain or justify electrical signal response changes to the feedwater valve controller to prevent valve lock-out of the type experienced during the August 10, 1981 post scram transient - relate to other previous scrams that did not result in feedwater valve lock-out in the open position.
Response
Af ter a reactor scram, reactor water level decreases due to collapse of bubbles to approximately 0" as can be seen on the attached recorder print-out (Figure 3).
The feedwater system sees the need to rapidly supply the reactor vessel with water. As the water level rises, the regulator valves should tend to close to maintain water level at approxi-mately the 30" level. However, when the feedwater regulator valve controller receives indication that the valves should go from full open to closed, current from the controller can drop as low as 4.9 mA in the automatic mode.
Below 6 mA the controller thinks loss of power is occurring and tells the regulator valves to lock in the as is position.
Once the valves locked up, water continued to be fed into the reactor vessel allowing the water level to go above the IC supply line nozzle. A time interval of only about 30 to 40 seconds existed between when the regulator valves should have closed and when the water level went above the IC supply line. This small time interval can be appreciated by looking at the attached recorder printout of reactor vessel level v.s.
time. The instantaneous dip in reactor vessel level to approximately 0" and then the instantaneous rise to greater than 60" appears as a straight line on the printout. This represents approximately a 60-second period.
Water level went as high as 149", which is above the IC supply line (124") but below the main steam line nozzle (169"). The reactor water clean-up system was placed into service to lower the vessel water level.
Approximately forty (40) minutes elapsed before the water level dropped below 60".
It is reminded that the clean-up system is competing with the CRD make-up at approximately 60 gpm.
Feedwater regulator valves have locked up in the past whenever the l
current from the controller dropped below 6 mA.
However, control room operators detected the locked up valves and were able to prevent any adverse consequences. After the August 10, 1981 transient, the 6 mA l
setpoint was lowered to 3 mA to preclude the regulator valves from locking up in the future. Additionally, the need for a new controller is presently being evaluated by NNECO.
Note:
The attached Figure 3 represents reactor vessel water level as indicated on the RPS Yirway monitor. The 30" level on the RPS Yarway is approxin.ately 155" above che top of the active l
core.
Print-outs from the wide-range GEMAC monitor were used l
to determine that water level went as high as 149" (219" above the top of the active core). Additionally, the 124" and 169" numbers represent the bottom of the IC supply line and main steam line nozzles using the wide-range CEMAC scale l
as a reference. The IC supply nozzle centerline is 201" above l
the top of the active core; whereas, the main steam line nozzle centerline is 249" above the top of the active core.
I i
-4 Question No. 3 Describe the upset conditions for each of the previous isolation condenser water hammer incidents at Millstone 1 including the August 10, 1981 incident.
Response
Isolation condenser water hammer events at Millstone Unit No. I have occurred on March 11, 1978, December 19, 1979, and August 10, 1981.
Information concerning the first two water hammer events can be found in Reportable Occurrence R0-78-7/3L and R0-79-36/3L. Additional information will be provided later.
Note:
Additional information regarding all four (4) water hammer events, including the February 12, 1976 event, was provided at the November 3, 1981 meeting and is included in Attachments No. 2 and 3.
-5 Question No. 4 Discuss the decision process that resulted in operator action to put the isolation condenser into service after the first water hammer incident on August 10, 1981. How was the condition of the isolation condenser assessed?
Response
The operator was not attempting to put the IC back into service but merely trying to realign the containment isolation valves in accordance with the technical specifications. The operator had no knowledge of any water hammer event having already occurred. Realignment of isolation valves was initiated approximately 30 - 40 minutes after the reactor
- s cram.
i IC-3, which is the normally closed valve on the condensate return line, was indicated as closed on the control board.
IC-1 and IC-2 valves, which are the two steam side isolation valves, were both opened (required to be normally opened by technical specifications). Once these two valves are opened, a direct downwardly sloped path exists to the reactor vessel. Therefore any water still present in the steam line should have drained to the reactor vessel. Approximately 5 minutes after IC-1 and IC-2 were opened, IC-4 which is the inboard isolation valve on the condensate return line was opened. At this time, rumbling noises were heard in the control room, which the operator believed were being caused by water hammering in the IC system. Subsequently, the operator closed valves IC-1, IC-2 and IC-4 and the rumbling noises ceased.
The above is consistent with the isolation condenser shell temperature v.s. time recorder printout. This printout illustrates that the IC was automatically initiated and then automatically isolated.
IC can be verified as never having gone into service since the shell temperature never exceeded 125 F, which is well below boiling.
l 1
A c---
Question No. 5 What simulator is used for Millstone 1 operator training relative to isolation condenser water hammer?
Response
The Dresden Unit Nos. 2 & 3 simulator, which does include an isolation condenser, is used for Millstone 1 operator training. However, water hammer events are not simulated. The purpose of the simulator is normally to train the operator to assure that enough water is present in the reactor vessel, not too much.
In the past, water hammer events have been detected due to the noise in the control room and were quickly mitigated by closing the containment penetration isolation valves for the IC system.
)
_7_
Question No. 6
)
Explain failure mechanisms and corrective actions.taken for the following systems:
(a) -SRV acoustic monitors (b) Scram Discharge Volume (SDV) continuous UT monitors.
' Response:
(a) The reasons that the acoustic monitors did not indicate that the SRVs had opened during the 8/10/81 transient have been determined.
A time delay of approximately 5-6 seconds was designed into'the acoustic monitoring system..It is believed that some of the SRVs may have been opened for less time than this. Additionally, during the post-transient testing of the acoustic monitor which was manually 4
-activated during the 8/10/81 transient, the amount of vibration produced (8g) was less than the setpoint (9g). Therefore, it is believed that this particular acoustic monitor, as well as possibly others, did not indicate the opening of the SRV due to the setpoint being set too high.
The gains on the acoustic monitoring equipment have been increa' sed and setpoints have been lowered. All SRVs were opened during power l
ascension after start-up from the 8/10/81 reactor trip and all acoustic monitors worked properly.
It is noted that all SRV temper-ature monitors functioned properly during the 8/10/81 transient and that the plant was in compliance with related technical specifications at all times.
(b) The SDV continuous UT monitors were installed at Millstone Unit No. 1 in response to I & E Bulletin 80-17'as interim modifications until " final fix" modifications could be implemented.
In our letter dated March 20, 1981, we informed the NRC Staff that plant modifications which comply with the BWR Owners' Group Evaluation Criteria and the
~
l NRC's Generic Safety Evaluation Report are scheduled to be completed during the next refueling outage.
The intended function of these interim continuous UT monitors is to-detect water accumulation in the SDV before the ability to scram is,
compromised. The continuous UT monitors installed at Millstone Unit No. I are capable of performing _their intended function satis-factorily. The monitors work properly during normal operating conditions when any water in the SDV is primarily stagnant. During a scram, the water in the SDV is more turbulent and the continuous UT monitors are less reliable. However, it must be remembered that I
these monitors do not need to' function during a' scram. Their l
function is to alert the operators to take. corrective actions if too much water accumulates in the SDV.during normal operation, not to verify that a scram has occurred.
i.
Since the August 10, 1981 transient, the continuous UT monitors have been calibrated and tested. All tests were satisfactory.
~_, -
-g-Question No. 7 Review design adequacy of isolation condenser system anchor bolts consid-ering the reported looseness observed on eighteen (18) anchor bolts following the August 10, 1981 water hammer event.
Response
Subsequent to the water hammer event, an inspection of the isolation condenser supply and return systems was completed. This inspection covered 100% of all accessible pipe supports and piping on supply and return sides of the isolation condenser. The inspection of the pipe supports considered the following conditions.
1.
Misalignment 2.
Extraordinary Movement 3.
Broken Members 4.
Loose Members 5.
Bent Members 6.
Cracks in Steel and/or Concrete 7.
Corrosion 8.
Evidence of Leakage 9.
Loose Fasteners 10.
Concrete & Mounting Integrity The piping was inspected for any signs of gross physical deformation, excessive movement, or visible loss of pressure boundary integrity.
The inspection revealed four (4) base plates with loose anchor bolts.
The bolts were noted to have been pulled out by 1/16 inch on two of the four base plates and all were subsequently retorqued. There was no structural or other damage observed.
Evaluation:
1.
Pipe Supports The inspection report indicated that two nase plates had experienced slippage of the anchor bolts which was measurable.
In order to evaluate the condition of those base plates, the actual embedment length was verified after retorquing. This information was incorpor-ated into the analysis of those supports which was developed as an on going part of the work effort of 79-14.
The results show a safety factor greater than four for all anchor bolts on those supports.
The noted cracking in the grout is of no major consequence as the grout's sole function is to carry and distribute the compressive load to the floor slab. The cracking will in no way compromise the integrity or function of the supports.
It is believed that the observed displacement of the anchor bolts is due to slippage in the concrete slab due to relaxation of the preload. The displacement was sufficient to allow the wedges on the Hilti type anchor bolt to rescat and withstand their design load in the displaced position.
J
_9_
This is substantiated by our analysis which verifies a safety factor greater than four in the displaced position. We feel'that this conclusively demonstrates the structural adequacy of the pipe l
supports to withstand their design load.
2.
Piping The nature of the water hammer loading on the piping system is such that only axial forces are created. A review of the water hammer analysis indicates that the major force is created by a pressure wave which is reflected from the collision of the water slug with the standing condensate. The piping geometry is such that the pressure wave is concentrated in the area of the long straight pipe run at elevation 90'9".
Tt.is pressure wave is expected to attenuate with travel away from the condensate / steam interface. This has been shown repeatedly by evidence of damage during previous water hammer events.
The stresses in the isolation condenser piping generated in the water hammer event are primarily axial in nature and thus negligible bending moments are generate *. The water hammer analysis indicates all piping stresses due to de.Jweight, pressure, and water hammer are below 1.2 S.
i stresses below 1.2 S ere is no concern of H
H fatigue or need to further evaluate piping integrity. The pipe attachments which see high loads in a water hammer event are axial welded attachments. The attachments have large reinforcing pads and thus represent a minimal stress concentration and a minimal fatigue effect on the pipe. The nature and magnitude of the water hammer induced stresses are such that no ND8 of the welds is required.
To further demonstrate why NDE is not required one must look at metallurgical evaluations previously performed. A sample was taken from the isolation condenser line where a major crack existed, detected during this past refueling outage. The evaluation showed that the entire crack was environmental in nature (IGSCC) with no evidence of mechanical load contribution. We may therefore conclude l
that the water hammer events contribute little toward either fatigue crack initiation or crack growth.
There has never been evidence of isolation condenser pipe damage or motion within the drywell. There was an additional water hammer l
restraint added within the drywell to provide further assurance that water hammer induced motion or damage would not occur.
It should be further noted that all supply piping weldments within the drywell were replaced during the past refueling outage. For all of the above stated reasons, it was determined that an inspection of piping within the drywell as a result of this event was not required.
It is believed that the evaluation completed demonstrates the integrity of the isolation condenser piping.
_1o_
Question No. 8 Compare the number of SRVs that opened during the August 10, 1981 transient to the DBE turbine trip without by pass transient.
Response
Turbine trip without by pass is one of the transients routinely reanalyzed with each reload. Turbine trip without by-pass is almost identical to the generation load reject without by pass transient, reported in " Supple-mental Reload Licensing Submittal for Millstone Unit 1 Reload 7," Y1003J01A09, dated June, 1980.
The peak pressure calculated in the above reference for generator load reject without by pass is 1193 psig at end of cycle (EOC) and 1183 psig at 1800 MWD /t before EOC, both at the steamline elevation. Peak vessel pressures are slightly higher, but are not relevant to SRV setpoints.
Slightly lower pressures would be expected earlier in the cycle.
The attached Figures 7 and 8 show the transient results, which indicate that all six SRVs would be expected to lift.
Recognizing the following, it is reasonable that 5 out of 6 valves actually operated:
(i) The transient calculation uses conservative input assumptions, including +1% over the nominal pressure setpoint.
(ii) The transient is very rapid. Figure 7 shows the pressure rise turns around within about 2 seconds of the 4-SRV group opening.
(iii)Setpoints are typically within + 1% of nominal, which provides for approximately 23 psi range between a low biased valve and a high biased valve.
In summary, calculations show that the setpoint of the third SRV group (4 valves at 1125 psig) should be attained during such a transient, and that the pressure rise should be rapidly terminated. The August 10, 1981 transient data shows that 5 out of 6 valves in fact operated, terminating the peak pressure at 1128 psig.
Therefore, the August 10, 1981 transient is consistent with the calculations and it is presumed that the sixth valve had a setpoint slightly above 1128 psig, but still within the 1%
allowable limit of 1139 psig.
_11 Question No. 9 Justify or support omission of a feedwater pump trip per NNECO's letters
- dated April 23, 1979 and January 16,.1981.
Response
. We are presently evaluating several plant modifications which will preclude the occurrence of future water hammer events'.
One of these potential plant modifications is a feedwater pump trip on high reactor vessel water level. Another potential _ modification is feedwater isolation.
Complicating these evaluations is the fact that the feedwater and condensate -
system at Millstone Unit No. 1 is safety-related. Thus, stopping feed-water flow on high reactor vessel water level is a restriction on.the operation of a safety-related system.
Note:
Proposed plant modifications can be found in Attachment No. 4.
I I
i i
i
Question No. 10
Response
a).
Rated Thermal Power 2011 MW (th) b).
Feedwater Valve Size 12" c).
Feedwater Valve Manufacturer Copes-Vulcan d).
Steam Turbine By-Pass (1% Rated Power) 100%
e).
Number of By-Pass Valves 10 f).
By-Pass Controller Type Electro-Pressure Regulator (EPR) g).
By-Pass Controller Manufacturer GE h).
Number of Isolation Condensers 1
i).
Number of Isolation Condensers Required for Operation 1aboveg0% power j).
IC Capacity 206 x 10 Btu /hr.
k).
Number of Steam to IC Lines 1
1).
Size IC Line (Diameter) 14" from vessel, goes to 16", and then goes to 12" at T.
m).
Length Steam Line From RV to IC See Figures 5 & 6 n).
Height of Centerline of IC above top of Fuel See Figures 5 & 6 o).
Horizontal Distance: RV Centerline to IC Centerline See Figures 5 & 6 p).
Normal Water Level Above Fuel (Height) 155" (510" from RV bottom) q).
Centerline of IC Steam Line Above Fuel 201" r).
Number of IC kater Hammer Events Since Initial Criticality 3
s).
Dates of IC Water Hammer Events March 11, 1978 December 19, 1979 August 10, 1981 Note: An additional water hammer event occurred on February 12, 1976.
t r
i 0 l,1l0 lg 20 l 30 { y \\,5p l 60 j 70
, 80,. p ' },00 l i : i
!g4 6
.i 4
i 6
)
j
[
!d
.}l g '.
r
,!' '.12 MNf p
. gi i
p, b
i i
e i-
...a 4
j i.:,. -: ; ',- ;}.U%I0 4
- l'-
.] -
- I !9 d
.h, w i~
- p
.l
,e s-(
-l:, : W j-ij #
j q"'
l m
.-:: !,:i s
n NJ\\. l h.. o,i, l k-
~
l.. !!
Uj M $ R liji
!!i.
r a
s o, m a;
- a. n Ii -
l'll i;l:
.h
. ja l, I '"
i
'i..
so p*
l., !ll. 4:
- I I
.s s
- l:. r".
lip!
jt j W' '.y-i, 1-
. l
,i mi y.+
i,,.
,t!.
p:
-~
i
-.E.h e iiU,:ll
'i! h
$! !!i.
n
.i
.d5-6
-3, o
- ,a. ',
- )
73, ". 80 p i' S 0i: 10) 4, 1
, 30 ; ; 40 ;
e,!
n.o. n,.
n:.
. f.
!j I.
l.
- ! -j!
.I [.[.
'.8 I
- '. n.
io ~..
a 1
5
.l.
e,1 6!,
- p.11'1PM' i'
- i
- d,.,,. s4..
in lll!., ;.rl ";[ n:..,[
- l
];;
.. -.p
'.ti
..,.;e j'
ll f
.c a
p, ii A>
- .I i
I
- !: iii; ;,i; y l
j.
8 IH;
.i
- lt d
l
- i.,
,q
- p ;
p'
- ilj.
,1 p
i,
.l oi
.i i,.
.c.
l i
g.
i:
I l ". U.
- l. '
,e o
i g
l;!!
i' ll]'
'l l
. I!'
.16:
S t.
t-a, i
,e
!!Q f; :'7jl
- [ 3 o # 4 w, i
@lit t
.i t i
.5 Y e
y.,, ::,:,.
- n..o:
n.
.!!q a.
.e l
.ti:
r i
[l6).!-
I:m. ' " 20!! "30 il 40 '.. ~5
'.o:/
Se 1.0) io o.
2 80 70 0
- po.
i i.i...-
i ;..
. l ;.
i:,
I.
o is o,.
i-n.
gll;,i.
- p..
, +
- (t so
..I 7,o f.;,,
i.-
l.,
ji o
{i 4,
.i
- 4..
.:. j ;;;.jigo pg,,
y ?.'
a-e i;;; tg y int ij :.
l
[
n
... a. p......
,ii..,
.a!
i ip.
i. :d :' p..:L:
- . i i;.
c a.,
n:.3
,,'i...
- .p.
- :;1:.,:;....
n;
.i.
i;.
. ! i i J._
~-
.... m..,n i;
op 1~
!.,p n
- i' "l-1
- :
'F
. :;l
- ,"o i;!
i:4 n a.d i.
'ip i
- ,L p
i,i p.
o w
q
..g p
ai w q; m i
i i
I n
a
,o i
i.
m.
.b,, j ;-
gi
. ; h., @
?:
i'. 1':
i."
o.
)
p i.
.;i! n:
.p p
- 9 4
I, j.;
.o..n..
i.
[' t,.
-.'t,,,o.i 1,j-
- ,i i!; ; ;'
L, ' '.
'I l;: !, - id, i: i 1il -l0.i1,lH!h.,-dt:
- L il;l hi!
5:!.it i
..e i m.t.:
u,i; qp
-i
..i,,,,. :,..i; m.
i r::
i
. :i p.
- i, 0
.p0;i j20'.".c:'40.
- 5 )i.! 60
.,7
,n80.]. 'i. Im r.
.s 0:L 10)
- o I
o i.a co o, l
o"
!jl
.[
.il, i,.
p.;il t,
I i
't' o
i l
U l
S
-i i
,,q i,
elf g
{'
,.;,1, f l
.ig : n,In r.. n i
j i
i I
i,
)
vil i. :p:
i o
- l i.
l i.
i m
i, e
1:
s.!
-l.
"l:
a.;
,q" i :j jl i.:l
' 01 l!l t,,'
f U-
. @[.,,
.l..
i t
t
--'"V I
i i
e-
' a,
- lt, l. di; l ',
i.l-il in.n !
...i i9 n.o i.
'!+
- pl lli,, ' ' ii;
.:1 l'! Hi 4 PE:
[,
i'
'i '.f ' '
- ':. iii:,'
11 ;t:! q"
.;U j'!i jil,.ij :lli ilo t t M f'6/6\\
^
. '.. I!,i i. <
(o ip
..8 l
l p
pu
...,l, !jl' ll:
E.'
- p.,i li ii.. 4.
,"i :P
,-,n
. ! g' i..'
oi! :,I: l tl i ;1
- 'i.
!i i n.
- n.
}.
ll.
1 l,
'l'l
.L
.i.
!;i n.;;
.l..i,r, l1,
- i.
. 9 :. ip;
..i i:
m h..D l ', i l-
,: I i.I; I
'!j! 'l' ;i(!!ij lll )
d.i g
'. 3, 6..,
.!l;l.,.t;.l "
11 j
i n,:
blhl,i d;!
i.
ni.In.-
. e: l
.I
. g-i i,.
c.
F' li !
i.
e I UwN (J
Q
- .? liii D::
i'
.:i
- i' l'
j,;p,U M
l1, :;j:
'.'l ;i[h l't l;,
h Ib :!!,
l g,,,,,
ilp!!
i ::'
,: ii-c
..i.
.J; i'
u:!
.s s i
j.,.
il,.ll;,:t:.
g;
- o. e l;
. p. i i.ii
.ii, i
i
,..il.1 i
i o-i f;,: O!!
.30;;
40:
- l' i;' '[N
.:!lJ
, o i
L,ic i'i' '9 llhth.-
1, "50 16)'"
- 7) '!
.i.i, ' :l,
,il, on '! J,ii;
- i :t' ll80: L 50'i 10) n:10 ~.j 20'
- i sin lii; i>! I
,e!
.,! ll:1-i o.
p.l.-
!jj; l
- l" l
.p h:
,1, i i
- at ;ill.;it i
i
- li.lll
.I;.,;I..
.l.
i?
4'
- ::5'PM i i
l n-t i si 1
a n
,ii,
..I;,..
..i.
,i..
g.
j;j.!l l!g
'p..,.
"3 o u.
,p'"'
il!
- . !J i
,a m
'i i!
L1 I i.
, 3,,
- ,- l
.I' llll ii;,
u-Uj" j l
jil; l
,'j.--l L " ~ ';j '
$, !i.
,l.,.
ilp '
'. j!;
! li;:
i d: du c.
l'l
!P l
- ].
j in u 4:.
.ei, A
1,::
jlil i'l! G -,N %....'
y, F I,!l
.i n
y
'I i,.! si:I
!i!
li
,i, i
.::.i;.
,i!i
(
O,i ;h.
q.
,1 l.
p,p !P,l.'. '; W 1
, ! 1,,i ih,
,.i 30: : 40 ll l-i is.
0, :j,ID. ';!T. :20 M'P.b
.;.p 1 :'i ;5
'6) 70. '
80 ',J 90 i 10) h! >!i!
i.I
- 1
~
,. ii,
.l, i. -l
'l :
n i;
I i.n. '
..I
., ' gj.
.i. ;;;- n., -.,.,,; l,..
- l
,.i 4
.m i:n
. i ti; : j,
. p..
!l n
p..
- i..4 pg i.
l l
f ':
i
,i e
n .
.,!. ' h ; l. *. +;
'O
.~
n*
'@ 'l,4 l,U U.
i i
'n d,i:E,y' 'li'
';i
.b,
4 d!! 'f;! l;!! "
o n
i 1,,$...
.e]..,f.
E 1 i<.M m
9,i 4
4.,
i ::
I
- i.
. i
.O L:
(
.i ;
.p.b Qh :52%fp> _
m.m
I f
th;,.
, r"f,'j"'p.;!!
t i
I
<jp i
l t
I
.'., l i.4pg_.
.i i
.i l
l.
1 b;0..
4,0 l ;. 5,0 ;.,
h
.O i
10 20 i, 60 I
i
.i.
l j
jl i
~
il
' i.
-l l
,,j i
i
~!
,, l' e
3pM-8 I
! l 3
.N I
i l
~ \\1 ~~
~
h
(
p r Ml-l.f i
y i
jy
-a 2?MH I
/ : 9.;ll.'
i i
l i
- ,9
@.. e l'
i -
6 g
9 f
., 7q iFM-0 10 20 l ; 4,$.
50,, ',60 I
i.
i i.
, g.i.
i,.,
el' j-
- !i ;;il 3
1:
"-12 AM-i 4
i l
i.
j ~
f d
q P~
!l l
l g
~-
- WAM"i i
i d
l l
l i..- i
[
i.
m i-I i.i '
j:.ij~
I
.i' 8
.i j.10, AM.
li'-
j
~
O' 't:
10
. 20
'd 50 'l j;60
,3l,
- !..I i
g
,i C l ;'
- e i:.
9
[
! li ii i
g i
i
. ' ' g A. M'- j
?
[.__.____.___..
u I.
a i
i
,v.'
.- r l
j 4'
i L[
i t
a i
2 g Ag
.3 i
e L _ _._ _ _..
I._>
e i
a=
g
. l,.
l o
e i
- l'
. o 1
j I 7,AMq.g j..
x ___
,d 10 20 t
50 60 l.
i i
A c'
l_... _. _.
'l i
=
1 D*
W r* m a s
. o
(
).
C/
,, 4,,,
qii....o a.,
- ,!! d!! :Hi,io['ql!!!l..:
.:;i tI d - l 1 e r:l i.; 'P8 l.f!!!;
,.,.. sa
. i. i l ~,.
.,,i.,,..,.. ).
30
~an
.. n.p.
- i. i, -
i inn.,
,.,...,i f.
's l,t i 3
- I
<llil'ij 1,
,j;;
.f f
1
..I 'l,-, p' 8;
,. ' 2, h ;O M,,,j.
illi 'lh.vi !I l lji lb-4 i
ij i
,i',I i;t el !,. a, i,: ::;-
l
., '(' ;[
!!!, "? f,'
5 M-[.-
-i"
'., !!i!
d d0 70 80
,'90' 10)
- 20'i"30 Ir,.l D i
':l '
.<, ::i; t
-J
.4)'
)
,i
!e
,!I U h h 5[
[;h! i'i
, ' ::9 d l'
1 I-O !h, i,j a]M U Q:i4;i s Oh O;:j'fi i!!$ Wid $. '
}
AG c
ly hi i,![ V
'E O < ; ll;. gM dp! '! : :]' 'il l'i I;il 's 'l ', h.,; * ;ili i
d i l-
,1.-
en
- N ;,:'
..,,. Ih,!!; i.;
-,l:.
' til y;,'
e
,t p.
- ,- J i.;
- o...Q
- F g
~'
l' d
il e.. !
i N,
'I,l' ll l
,:l
!d k,!,h
- i ll!
s.
m, 10 n
s o
3 i,.. -
i.
g.
3 n.,
,p i.
W.:.
-4 m, '!
l h
i I
N
-2 I
i, g
l,.
Ib
?' )./[
i l
i
.;l is c.#
li..
l
' ' :idc Aa - p :. p!'.
l c
i:
l
.e g,A ac:
s, y
l y
c.;
i...flF; 0;1
' 6)' l7)1,;80.i.901 10)
! 20:: i,30 : 'q 51)
- I I
l,-
- ic
,u, i;.,
i.e,,;.t; 1,;1 !.
i-c' t-i
~
- i. ;.
t, i,,
.[
il
,l i;
..p '.
,l u Nl, );.l
'], l [
.j 4 Q -]'
'. *;;l
' ul'
,o ' il U:
-i
,j.. :!'.
b
.i.;.
t
- l '
l1
',,- l' !
f.'
8 g,
s
- f i.,
.! -l,
- ll'
- :
e i.
i p:
.i i..
I'j.,.l i.e p
"l
. ;a;,I! !,
- t
'l!- ';.
,-l.
i.
1 i.
i
- d.
pt:
n:
.l -
-,, 'i i,:
- e.... : 'i
.,i:,'
.,. I,
,p g
I i e im i:
e i
i
.i
, s.:.
. c, i
,?
s.
i..
s g
a I'
'I g,
8t; g
gg-j 9
D
'i:
-l.
l i;,
'll.
, i;.,
hl ;;['
I ' L j.'* : !,-
l i;
l
-ll' e
i, o-
.. m i
'!m l.
1,,
8 l '-
,,ll!g!i j
l!
IP]L ;;
l 1,
.i ip-p
!M i
- g-s' i
s.; i.,i i;;l,
i
, l!
l
'6 l
l I
t I
l l
l 1
i l
1 i
L
~
4
~
- n. e
~. '
.a
~
h h *~ d
!o Ehtl o
1 i
e:
P-P
- g=.8 3;
e 4:
2.
Pgj
- $ 32 5
2 1
~
S' t1 2 WLg 3
g m
- 5: 5 at 3s
!E L I 5
IJI
!-55 E! 55 f* r!
oO A g3 5 '-
- 3 C c
4 e13
" i. !! !!!s
= m"t y 34
- l{
v 2
a i, 2 g4 Y
u I:
k iG u
9 9-i.
i I
+
f,,,'s.,
M *C s' '
f' r-0
,, ; i e, e
G,\\ 4
+)N,.9
", 'i,, ?.
9,%
Ug i
i ' ~- (
gg l' /4 59 s
g e
J c9 g
g i
\\
i:
u t.i x
w-so
,8 I
'g
.9 f,2 x
1
.3 cc
~
l 1
~
bl
, N 'I33 %
.s i
5 N
go y
.$f
~
L t{u f
'g
,t O/
t.1$
[
~
s;r
~
5l
<g
.ie
/
T s
g
_3 i
.4
/ '
. i a,
,i g.
<ec
/ Q }A
.. f h M83 W
,a w a..
ah i.-
====o Lt j D I u.,,) z 4g 3
j 3
y i
1
~
l.
ti
i N
o
\\
11 TC 29 11 TO 3*7/
gdR X-19 AO'
'X= -53.59' v
m no.es-v. i c. s s-o 365 25-32.15' Z = -3 2.15'
/
go s
~ s.=
9 i.6
/1
,u
.is-s m.
.sc o
,95
,yy i.s&
A 31
'O. Co.
16 IJO' S@
1 3.
,i ss,ap'b
?h.C
'* c '
\\
ko0 8~
o j
C *
2:
,,go s.c 4,.
- r,
o
'c.,,,
'g
,f 2,
22 7',
6
,A i
23
. e,.
I '5
,f'
~
2'\\,3 g S
9 g
m s
/
<a
<e.
o r.
2.cla e.1 16
~
T+
l'gg Q
n 9
hb,9
/
.m,'
.sA z.ud s.w' l.53' 3.09' TEMP. IS 5Wr; A. 5.92'/lCO'
' HOT MODULUS C: ELASTIC \\T'l SEE FA6.lSdS.
- 1S 25,55G 000 P.S. I E-2 3(o 2 -IC-69 f 70 MILLSTONE-4 7
~ REACTOR ISOLA.TtCN CONDENSEi SUPPtN.
(281)
I SO -E-2.
i i
l 1 NDITROP4 ftUX 1 VESTEL PI ES RISE IPSil 2 AVt SURFrCE HEAT FLt'?
2 9 VEIT WLvf I t tH C A.5 M 3 COF# Ittt i FLOW 3 F2 LIEF Vi tvr ftp.4_ ( 4 S M.
150.
g, 4
9 BrfH35 v>'_ vt9L 04 5
5 6
i N
3
't 1
8 100.
200.
l 3
W E
E 5
I
\\
l ia 50.
100.
Q 3
3
[
2 2
O.
nA.
u iu 3
3.
I I
0.
O.
4.
8.
12.
16.
O.
4.
8.
12.
16.
l TIE ISEC)
TIME (SEC1 M
Ho 1
I O
8
=
C o*
I 1 LEVELilWH-REF-5EP-SMIRT I VOID REW71Y13 W I
i 2 VESSEL SIEAMrLOW 2p+tfFTt0CTIVITT 3 TURRINE S TEAMFLOW
/7 Srfue RFTr. T I V I T T l
200*
g*
4 FEED @TEF FLOW y
4 TOTrCIUDIVITY-~
5
/
~
2 100-u 0-
,~
~~
N t
a A
l I #
3
_3
. o 3
0.
=
.l.
3 L
M
. W i
.g @,
.t.
4 2,
.t.
4 l
0.
4.
8.
12.
16.
O.
0.4 0.8 1.2 1.6 TIE (SEC1 TIME ISECI l
Figure 3a.
Generator Load Rejection, Without Bypass, EOC8 s
.t
)
.N
l I I NEUTRON FLUX 1 VESSEL PfES RISE (PSil 2 AVE StNfCE HEAT FLUX 2 SAFEIT VI LVE FLOW ( 3.
SMs ISO-3 C0dE INtfT FLOW 300*
3 ret;1{F_VFLVE FLOW _(g Sjtya 4
4 BTPRSS Vft.vE FLOW S
S 6
1 3
9 3
b is 100.
b'
\\
u.
E
'o 50.
100.
S
~
h b
3 3
N f
12 0.b-
_1 1
1 M.
u l
u 3
O.
m.
O.
4.
8.
12.
16.
O.
4.
8.
12.
16.
TIME ISEC)
TIME (SEC)
N 80 so o
V o
I LEVEL (INCH-MF-SEP-SKIRT h10f0 MfrTIVITT
$o 2 VESSEL SlEAMFLOW
/
2 DOPPLER f-EACTIVITY 3 TlitBINF ' TEAMFLOW-I-
V 3 SCRAM RFITTIVITY FEEDWATErTE0W f
u TaTr( MK TIVITT 100*
M O*
2 W
K N
N z
y f
)
- ig 6
2 l
L_V 3
i O.
=
-1.
y 3
g W
-100.
2.
8-0.
4.
8.
12.
16.
O.
0.4 0.8 1.2 1.6 TIME (SECl T!PE (SEC)
Figure 3b.
Generator Load Rejection, Without Bypass, EOC8-1800 mwd /t
Docket No. 50-245 l
Attachment No. 2 Millstone Nuclear Power Station, Unit No. 1 Isolation Condenser System Water Hammer Events 1
i April, 1982
v 3L
...-.a
.-~ -
x -
-- u Original:
Novtaber 2, 1981 Revised:
March 15, 1982 U
RESPONSES TO ADDITIONAL NRC i
QUESTIONS CONCERNING THE AUGUST 10, 1981 TRANSIENT AT MILLSTONE UNIT NO. 1
References:
(1) NNECO Responses to NRC Questions Concerning August 10, 1981 Transient at. Millstone Unit No. 1, transmitted to the NRC on September 23, 1981.
(Attachment No. 1) s (2) CE Report NEDE 25400, " Millstone Point I Isola-f i
tion Condenser and Shutdown Cooling System Piping
^
Cracks."
(Attachment No. 5)
(3)
"The Application of Tearing Modulus Stability f'
Concept to Nuclear Piping," Fracture Proof Design
[
Corporation, EPRI Project T-ll8-9.
Question No. 1:
Basis for new feedwater controller?.
Response
Investigation indicates' that the existing feedwater level controller.can.
~
be upgraded precluding the need for a new controller. Minor modifica--
tions - recommended by 'GE have 'aiready_ been made to the controller to preclude inadvertent " lock up" of the feedwaterfregulating valves with-the controller in the automatic mode. These modifications were dis-cussed in response to Question No. '2 in Reference (1). Additionally, a-design change to preclude, inadvertent-" lock up" in the manual mode is feasible-and-is-being-pursued. Therefore, a new'feedwater controller is-no: longer. deemed necessary.
t Question No. 2:
I Displacement of anchor bolts, you believe, resulted. fro:m slippage in the concrete slab due to relaxation of the preload. Explain relaxation mechanistically.
Response
When an anchor bolt is installed, Lit is. torqued to provide a tensile preload approximately equal to ' the design' load. ~This tensile preload is transferred to the concrete as a compressive load by means of two -
l sliding wedges. The wedges are most likely in some form of point con-l tact with the concrete or aggregate. The compressive stresses which re-sult from this point contact are believed to exceed the crushing strength
-of the concrete. As the concrete or aggregate crushes, the area of con-
[
tact. grows until the resultant compressive stress is less than the crush-ing strength.
4 l
h_. -
This localized crushing allows the wedges to move and hence lowers the initial preload.
It should be noted that upon application of load, a small displacement will cause the anchor to reseat and carry its de-sign load. This behavior, which is characteristic of expansion anchor bolts in general, explains what was experienced at Millstone Unit No.
1 after the water hammer event.
Question No. 3:
Amplify your response relative to axial piping stresses, minimal stress concentrations and fatigue ef'fects at reinforcing pads.
Response
The water hammer event is created by the collapse of a steam volume.
This collapse creates a partial vacuum which draws and accelerates a slug of water into the steam space.
In the case of the Millstone isolation condenser water hammer, this slug of water accelerates until it collides with the standing condensate in the condenser. This im-pact creates a pressure wave which then travels through the piping system.
The loading on the piping system is created by the reaction of this pressure wave at elbows or other interruptions in its path of travel.
This type of load and the fact that all axial mction is restrained out of the isolation condenser system by means of snubbers, will limit all stresses to axial (i.e. - no deniection will occur thus no bending stresses will be present).
The entire isolation condenser system has been analyzed for the water hammer event. The analysis results indicate that all resultant stresses were below 1.2S, the code allowable for occasional loads combined with H
sustained loads.
If one combines the low stresses with the large reinforcing pads on the axial restraints one can see that the stress concentration will be minimal. With the low stress and minimum stress concentraton there is
(
no need to evaluate fatigue effects.
I Question No. 4:
Discuss the expected consequences if o Main steam isolation had occurred coincident with scram from 100% power o Main steam isolation had occurred or would have occurred after automatic trubine trip and explain the basis for operator trip procedure following reactor scram (any improvements from previous water events).
Response
Assuming the locking up of the feedwater regulator valves as a given, the transient would not have been significantly different from the one experienced f or either of the above cases.
In both cases, the iso-condenser would not have been available.
In the first case, pressure would initially rise rapidly due to the residual heat in the core and be controlled by the SRV's.
The pressure rise would probably be sustained for a longer time due to the effects of the feedwater system pumping up vessel pressure.
The second case is a less limiting case and should not differ signifi-cantly from the transient analysis in the reload submittal.
It is of interest that Millstone Unit No. I has on two occasions under-gone a MSIV closure scram from 100% power. These resulted in no SRV actuations and pressure was controlled by manual operation of the iso-condenser. The actual transient is much less severe than indicated in the transient analysis.
Additionally, since discovery of incorrect Moog valve settings, we no longer believe that the early manual tripping of the turbine is related to MSIV closure.
Ques *
- o. 5:
There are four reported isolation condenser piping water hammer events, February 12, 1976, March 1978, December 19, 1979 and August 10, 1981.
Please compare the event sequences.
Response
(1) February 12, 1976 (RO-76-4/10): Isolation Condenser (I.C.) Tube Rupture o Scram on Generator Load Reject.
1 o MSIV closure on 880 psig isolation.
o I.C.
tube rupture.
o Reactor water level maintained below I.C.
o Opened MSIV - Surge brought level above I.C. steam line.
o Water hammer occurred (unknown at the time),
o I.C. was isolated to control leakage from ruptured tube.
o Damage due to water hammer was limited to damaged insulation due to pipe motion.
(2) March 11, 1978 (RO-78-7/3L) o While shutting down for a refuel outage, SRV testing was being performed - A SRV stuck open and the reactor was manually scrammed.
o About 3 hours3.472222e-5 days <br />8.333333e-4 hours <br />4.960317e-6 weeks <br />1.1415e-6 months <br /> later, while undergoing a LNP due to switch-yard problems, the vessel was overfilled to the IC steam nozzle.
1 o The IC experienced a water hammer event and was isolated manually.
o No significant damage except to insulation.
(3) December 19, 1979 (RO-79-36/3L) o Reactor scram due to generator load reject.
o MSIV's closed on 880 psig isolation.
o Reactor water level was maintained at +55 inches (about 15 inches below the IC steam nozzle).
o The MSIVs were reopened.
The " swell" momentarily brought the level above the IC steam nozzle.
o IC water hammer occurred.
(4)
The August 10, 1981 water hammer event was discussed in detail on November 3, 1981.
Question No. 6:
Provide assurance that an undetected iso-condenser pipe crack would not result in pipe failure during water hammer conditions.
Response
Information provided in the March 6,1981 meetiag with the NRC in Bethesda and the GE report NEDE 25400 " Millstone Point 1 Isolation Condenser and Shutdown Cooling System Piping Cracks" (Reference (2))
provide the required assurance. This information can be summarized as follows. The crack samples showed no evidence of crack propagation other than IGSCC or TGSCC. This indicates that there was no crack propagation directly resulting from mechanical loadings such as water hammer. Work conducted by Fracture Proof Design Corporation on the Millstone Unit No.1 isolation condenser supply and return piping done under EPRI Project T-ll8-9 indicates that the subject systems are fracture proof.
(i.e., leak before break demonstrated by tearing modulus concepts) (Reference (3)).
The conclusion reached by materials evaluations presented in References (2) and (3) is that an undetected isolation condenser pipe crack would not result in pipe failure during water hammer conditions.
Question No. 7:
Discuss the tests completed to assure that UT devices perform satisfactorily on the scram discharge volume header.
1
. Response:
A special test procedure SP-81-1-12 was conducted to test the operability of the cor.tinuous UT monitoring system.
This test was performed to comply with Item 2 in 1E Bulletin #80-17, Supplement 4.
(i.e., Oper-ability Test of CMS).
The test utilized single rod scrams (during the scram timing test), select rod insert, and full scrams at various power levels.
All results were satisfactory.
Additionally, procedures exist for the performance of an annual cal-ibration and a quarterly functional test.
The functional test utilizes the test pipe set-up where water is introduced into a piece of pipe and the electronics are checked for proper operation.
Since the transducer is not removed from the header (a spare transducer is used), the installed transducer's operability is verified by reading the pipe thickness.
The annual calibration goes through the entire electronic alignment as well as a functional test.
This system was calibrated after the August 10, 1981 transient.
Question No. 8:
How was isolation condenser piping integrity confirmed following the water hammer event?
(Inside containment in particular).
Response
To evaluate piping integrity one must look at work which has been completed for the isolation condenser system.
1)
To begin with all weldments on the isolation condenser supply piping inside containment were replaced during the past refueling outage per NUREG-0313 requirements.
2)
The entire isolation condenser system has been analyzed for the water hammer transient.
This analysis revealed the need for additional pipe restraints to keep piping stresses below code allowables.
The modifications were completed during our last refueling outage. With the modifications completed this system meets all code requirements for occasional loads and all pipe supports have the required safety factor of 4 for expansion type anchor bolts per requirements of I&E Bulletin 79-02.
The fact that the piping outside containment behaved as predicted by our analysis is confirmation that the loads and stresses predicted were not exceeded.
This indicates that the piping inside containment was within the design parameters and hence there is no reason to expect degradation of piping system integrity.
The integrity was confirmed by a visual inspection of all accessible piping and supports as described in our response to Question No. 7 in Reference (1) and by verification of the ability to repressurize the IC system.
It should also be noted that during previous water hammers where damage was evidenced outside containment, no damage was noticed inside containment.
It must also be recognized that the length of piping inside containment is small relative to that outside containment.
y Question No. 9:
Provide Figures 7 and 8 referenced in your earlier answer to Question 8.
Response
These figures are attached.
Question No. 10:
Discuss the response expected for August 10, 1981 transient if the feedwater pump trip had been installed.
Response
Assuming the MSIVs isolated due to the 880 psig setpoint and the locking up of the feedwater regulator valves, a _high water level feedwater pump trip would have reduced this event to a routine reactor scram. The iso-condenser and SRV's would have been available to control reactor pressure until the MISVs were reopened.
Question No. 11:
In response to Question No. 1 (Reference (1)) you stated that with a turbine trip, bypass valves open and assume pressure control function.
Will these valves close to maintain reactor pressure above 880 psig?
If reactor. pressure is being controlled, why take the mode switch out of "run"?
Response
The reactor pressure control system will control pressure above the 880 psig setpoint until the heat loss of the reactor main steam piping and air ejectors exceeds the decay heat available from the core. However, the operator shifts the mode switch out of "run,"
after verifying that the pressure control system is functioning properly, to maintain the main condenser as a heat sink as pressure sinks below 880 psig and to allow the scram to be reset.
Question No. 12:
(Question No. 1, Second Paragraph, Reference (1))
Should the operator have placed the mode switch in " start-up" to preclude isolation?
Response
The mode switch should not be shifted out of "run" to preclude an isolation.
To do so would be bypassing a designed protective function.
It is conservative to leave the mode switch in "run" until it is verified that the pressure control system is functioning properly 1
and there is no steam line break. This assures conservation of water inventory.
I T
i
Question Moo 13:
(Question No. 1, Fourth Paragraph, Reference (1))
Did GE actually do turbine trip. tests with the new 5 second valve opening and closing
' times to show reactor pressure will not decay?
Response
-Trip. tests were not performed to verify the 5.second pressure control setting. These tests would cause unwarranted and costly plant trips.
The MSIV closure.following a turbine trip is in itself not a sig-nificant problem to warrant this action. Millstone Unit No. l'is designed for such an event.
The operability of the pressure control system will be verified 'during subsequent plant trips. If the system does not function properly, additional corrective action will be considered.
Question No. 14:
(Question No.
1, Last Paragraph, Reference (1))
Statement is not clear, all plants can handle MSIV closure from 100% without tech-nical specification violation.
Response
The point that was being made in Reference (1) is that isolation of the MSIVs is not relevant or significant to the water hammer event.
4
)
Question No. 15 I
(Question No. 2, Last Paragraph, Reference (1)) Are there any adverse consequences from lowering the setpoint for the feedwater controller loss of power from 6 mA to 3 mA?
Response
The change in the lockup setpoint is not significant in its degradation of the desired lockup on signal failure. The controller can go from 50 mA to 3 mA in a fraction of a second during the postulated failure.
The additional time to go from 6 mA to 3 mA is negligible.
Question No. 16:
1 (Question No. 3, Reference (1))
After each isolation condenser water hammer event, procedureal techniques were to have been improved to alert Millstone personnel to the consequences of water hammers.
}
What had you done before the August 10, 1981 event?
Response
The scram procedure was modified af ter the 1979 event to procedurally restrict water level to +50 inches after a scram.
Each iso-condenser water hammer event was reviewed with the operators as part of their training. Corrective actions taken after each of the water hammer events at Millstone Unit No. I were discussed in further detail on November 3, 1991.
t Question No. 17:
(Question No. 4, Reference (1)) Why do the technical specifications require the IC to be realigned af ter the IC has isolated? The isolation is protection against a pipo break in the IC system.
Ilow does the operator ascertain that the IC isolation should be opened?
Renponse:
The technical specifications do not specifically require the IC to realigned af ter the IC is isolated. The operators were merely atten.pHng to realign the containment isolation valves to the positions indicated la Table 3.7.1 of the technical specifications. The oper-ators had no reason to doubt the appropriateness of realigning the IC at the time they did.
The IC system is a heat sink in the event of a loss of normal pawer and gas turbine failure, and is, therefore required to be operable whenever reactor pressure is greater than 90 psi and 40% power. The operators knew the IC system was intact because it remained pressurized even af ter isolation, there were no reported steam leaks, and there was no rise in drywell pressure.
Question No. 18:
(Question No. 6, Reference (1))
Why weren't the acoustic monitors for the SRVs tested previous to this event.
It seems that a test during a power ascension would have shown improper operation.
Re,sponse All the acoustic monitors were tested during the June, 1981 startup and all worked properly. Ilowever, they were tested at pressures much lower than the nominal 1100 psi relief pressure.
It was later learned that the response of these monitors is sensitive to the pressure being relieved by the SRV.
Question No. 19:
Question No. 6, Reference (1))
llow did Millstone Unit No. 1 verify that the SDV UT monitors perform satisfactory? Did they perform a partial scram? Did instrumented volume indicate full and initiate scram?
Response
See Question No. 7.
Additionally, all instrumentation on the in-strument volume tank functioned as required; however, since the reactor had already scrammed, such instrumentation did not initiate an actual scram.
Question No 20:
(Question No. 7, Reference (1))
Jim Shea, who is reviewing inspection report from NNECO? Ilow was IE involved?
_9
Response
Not applicable.
Question No. 21:
i (Question No. 7, Reference (1))
Is just a visual check adequate to assess pressure boundary integrity? It appears that a more rigorous test is in order.
l
Response
Tb-visual check is indeed suitable in light of the extensive work l
coupleted to qualify the isolation condenser system for the water 2
hammer transient. The analysis of the water hammer transient reveals stress levels below the 1.2 S allowable for occasional loads. The H
examination completed on accessible piping and supports indicates that the analysis results are correct and thus postulating degradation to the pressure boundary is not appropriate.
Question No. 22:
)
(Question No. 7, Reference (1))
I am not convinced that inspection of the IC system piping inside containment is not necessary.
NNECO should provide a stronger argument. Why were all IC supply weld-ments inside containment replaced during last refueling outage?
Were these stressed during previous events?
Response
See Question No. 8 for justification for not inspecting inaccessible piping.
Regarding the replacement of weldments, as discussed in the March 6,1981 meeting with the NRC and in Reference (2),
the weldments were replaced due to the presence of stress corrosion cracking detected by inservice inspection. The weldments which were replaced were stressed by prior events but the results presented in Reference (2) clearly indicate that all cracking was strictly either IGSCC or TGSCC and that no evidence of crack extension due to pre-vious water hammer transients was detected.
Question No. 23:
i j
(Question No. 8, Third Paragraph, Reference (1)) What does the statement " peak vessel pressures are not relevant to SRV setpoint" mean?
Response
l SRV setpoints are not based upon peak vessel pressures. However, for every reload, SRV setpoints are verified to be adequate to assure that peak vessel pressures are within pressures allowed by the ASME Code.
Question No. 24:
(Question No. 8, Reference (1))
Provide Figures 7 and 8 that are i
referenced in your response to Question No.
8..
l l
[
Response
These figures were provided on November 3, 1981.
(Also, see Attachment l
No. 1) l Question No. 25:
(Question No. 9, Reference (1)) Would the feedwater pump trip have changed the course of events of the August 10, 1981 transient?
Response
See Question No. 10.
Question No. 26:
What new procedures / instructions or modifications does NNECO propose to preclude Millstone from having overfill events and IC e:er hammer events.
Response
Short-term and long-term corrective actions to preclude water hammer events at Millstone Unit No. I were discussed on November 3, 1981.
(Also, see Attachments No. 3 and 4)
Question No. 27:
What procedures are followed to reinstate an IC that has auto-isolated during an event? How is the operator assured that no break has occurred in the IC system?
Response
See Question No. 17.
I&C Procedure OP 307 is utilized to reinstate the I.C.
Question No. 28:
Provide isometric drawings of isolation condenser and associated piping. Previously submitted drawings did not contain all significant details (e.g., vents, valves).
Response
Requested n awings were provided on November 3, 1981.
Question No. 29:
Either commit to installing a feedwater pump trip on high vessel water level or provide justification for not installing the high water level trip.
If no high level trip is to be installed, provide data to demonstrate that the SRVs on the steamlines can accommodate two-phase discharge without damage.
Response
Such a commitment was discussed on November 3, 1981.
(Also, see Attachment No. 4)
Question No. 30:
In the absence of water hammer, will water flow through the isolation condenser steam lir.es cause isolation of the isolation condenser?
Response
It is expected that water flow through the isolation condenser steam lines at the appropriate velocity would cause isolation of the isolation condenser.
Question No. 31:
What method is used to verify that the vents on the isolation con-denser steamlines are open, i.e.,
no plugging due to crud or other foreign matter?
Response
Surveillance testing to verify the operability of these vent lines is periodically performed. During such testing, these vent lines are isolated then reopened. Subsequent to reopening, these lines are utilized to equalize the pressure between the isolation condenser steam line and the main steam line.
If these vent lines were plugged, equalization could not occur. Therefore, periodic surveillance testing will detect any plugging of these lines.
l l
Docket No. 50-245 Attachment No. 3 Millstone Nuclear Power Station, Unit No. 1 Isolation Condenser System i
Water llammer Events
?
April, 1982
l s'
l
)
MEETING UITH THE URC i
TO DISCUSS THE AUGUST 10, 1981 TRANSIENT NOVEMBER 3, 1981 AGENDA 1.)
Millstone Unit No. 1 Plant Design 2.)
Sequence of Events of August 10, 1981 Transient i
3.)
Short-term Corrective Actions After the August 10, 1981 Transient t
4.)
Discuss Previous Water Hammer Events l
5.)
Corrective Actions Following Previous Events i
6.)
Isolation Condenser System Qualifications 7.)
Long-term Corrective Actions 8.)
Responses to NRC Questions l
i
)
i i
f I
I 8
e l
e 8.-.,-
,~,.--
-. =-
.---v-n.
Attendees Name Affiliation P. A. Blasioli NUSCO Licensing J. T. Shedlosky NRC J. J. Shea NRC R. N. Artwell NUSCO I&C T. F. Starr NUSCO Piping Systems R. J. lierbert NNECO Unit No. 1 Supt.
R. Frahm NRC A. W. Serkiz NRC W. D. Romberg NNECO Operations M. Ross NNECO Operations K. Murphy NNECO Operations W. C. Mihal NUSCO Reactor Engineering R. J. Palmieri NNECO Engineering D. Lipinski NRC
M ILLSTOIVE
/
PLAN T DE,3/GN O
B WR - 3 2011 mwr / seo mwe O
COM M ER CI AL l970 D
iso - CONDENSER
(/)
a FWCl /N PLACE Of HPct O
GAS TUR8INE /N Pl. ACE OF JEcoND D/ESEL D
MARK 1
CONTAINMEN T D
4 ADS VALVE 3 PL US 2 SRV (2 STAGE TARGET ROCK) c
SEQUENCE OF EVENTS ON AUG u.sr to, 1961 l.
REACTOR
,5 CRAM DUE TO S UR VE/LL A NCE TEST /NG Of MAIN STEAM H/GH RA D/A T/oM SIGNAL.
2.
NORMAL REACTOR SCRAM.
5.
MANUA LL Y TR/ PPEO 70R8/ME 4
ONE /~EEDWATER pump WAS SECVRED PER PROCEDURE 5
DUE TO SLOW RESPONSE OF REACTOR PRESSURE CONYROL SY. STEM, STEAM L/NE PRESSWRd DECREASED BELOW 880 PS/G, WH/CH CLOSED THE MS/ V's.
6.
FEEDWATER REGULATOR VALVE WAS LOCKED UP AND FEEDWATER CONT /NUED To BE PUMPED /NTO THE REACTOR.
WATER LEVEL /NCREASED ABot/E /SOLAT/ON CONDENSER
(/C)
STEAM SUPPLY L./ME.
7 PRESSURE /NCREASED TO loesPS/G AND THE /C WAS AuroM ArtcALLY /N/T/ATED.
A WATER SLUG THEN ENTERED 7~HE
/ C SUPPL Y / /NE.
THE LEAK DETECT /ON bP DETECTORS A UTOMAT/ CALL Y
/SOLA7"ED THE
/C.
SEQUENCE OF EVENTS ON AUGUST to,1981 (Cokr) 8.
WITH MSIVh CLOSED AND / C /SOLATED, PRESSURE
/NCREASED AND F/VE S R V 's L./ fred. ONCE THE FivE SR V ',s RESEATED PRESSURE WAS MANUALLY CONTROLLED BY ONE SR Y.
9.
3ECOND FEEDWATER PUMP W4S SECURED, WITH INATER LEVEL A80VE /c SUPPLY L/NE BUT BELOW 61AIN I
S TEAM L/NE.
l
/O.
NI S/ V's OPENED ll.
REA CTOR CLEANUP SYSTEM PUT /NTO OPERAT/oM 7~o REDUCE REACTOR WA TER LEVEL.
12.
IC VAL VES REAL/6NED AfrER dEUEV/NG WATER HAD A DEG UATE L Y DRA/NED FROM SWPPLY L/NE.
- NQWEVER, WATER WAs sr/LL. PRESENT /N 7~HE /C S YSTEM AND A WATER HAMN1ER EVENT OCC yRRED l
B
]
4 e
G g/
~
lEN<
/
l ng-O
- /
,2 r;, ) / l i;
,jd's H-V lM h' 8
-g e3 l
m O
u Y
s 8-1F' a
c
<s 9
1 Y
_ 7 7
.::m T
1 1
N 14
+3I 6
c e
c e
s n
m
/
~
~
l os I
l h
Y lnM 2
t
)
1 gl l
ae i
s M
(
l ryt e
a r eo cf 3
_p 5
a1Zim n
. x.
l r
e u
3
=
g yaiXm[j g s 4
a<
3 t
t n
n e
e v
v h
1 l
lI
~
SHORT TERM FIXES i
I.
CORRECTED SETT/NG OF REACTOR PRESSURE CONTROL SYSTEM.
(MOO & VALVE -
S 3Eco^/DS) l 2.
MoDif/ED THE FEEDWATER CONTROLLER TO PREVENb LOCK up /N A vroMAric.
(6 ma - 5ma L ock up) 3.
IMPROVED RESPONSE OF FEEDW4TER RE6utAT/NG VAL VE POJlTroNERS
(//. S SEC.- SSEC.)
4.
/NSPECTED AND ANALYBED
/C foR DAMAGE.
FoWND No DAMAGE AND ONLY A Loss of PRELOA6 ON SOME OF THE ANCM:A 80LT3 RETORQuSD ANC//OR BOLTS.
5 CHANG ED THE /C PROCEDURE To REQu/RE THE S TEAM TRAPS To BE /N SERV /CE.
l 6.
/ NCREASED EMPHASIS ON WATER l EVEL CO/VTRol DUR/M6 S/MWLATOR YRA/N/N6.
7.
RE SET ACOUSTIC MON /roRS (SRV) 70 /MPRO VE PER FORMANCE.
5
~
PREVIOUS WATER HAMM E R E VENTS a FEB.12,1976 (RO 4 /10)
IC TU6E RUPTURE l.
SCRAM ON GENERATOR LOAD REJECT 2.
MSI V cLosqRE ON 880 /SoLAT/oM t
5 IC TuSE RUPTldRE l
4.
/ OCK UP Of FRV-LEVEL 3roPPED BELow /C.
5.
OPENEO MSI y -
3 URGE BROWG//T LEVB 480VE
/C S7~f4 M L/NE.
1 6
WA7 ER HAMMER OCURRED (UNk/VoWM ATTHE 7/ME)e 7
/C WAS ISoLAniD 70 CONTROL LEAKAGE FROM 1
RUPTURED TUBE.
8.
DAMASE Duf Yo WATEK //AMMER W46 L/M/TED TO DAMAGED /WSULAT/ON DUE 7'o P/M MO770N.
l
+
PR EVIOUS WATER HAMMER EVENTS a MARCH Il I978 (Ro 7/SL) o l.
WHILE S HUTTING DOWN FOR A REFuGL OUTA GE, SR(
W/IS BEING PERFORMED.
AN SRV stuck OPEN AND THE REACTOR WAS MAVuALLY SCRAMMED.
2.
A80ur
.7 MoWRS LATER WN/LE UNDEAGo/N6 AN LNP DUE 70 SW/rcM YARD SwircN/N6 pro 8LEMs Ti(E VESSEL WAS OVERS /LL ED To 7WE /C STEAM No 2fLE.
3.
THE
/C MAD 4 WATER HAMMER EVENr AND t-WAS
/SQLA TED MANWALL X i
4.
NO SIGNIFICANT DAMAGE EXCEPT To /NSULAT/0N s
PREVIOU.S WATER HAMMER EVENTS O DECEMBER l9, I 979 (Ro 36/3L-)
I.
REACTOR SCRAM DuE TO GENER470R LO40 REJEG 2.
M SI V's CLOSED ON 880 /6044T/CW.
l 3.
REAcrox WATER LEVEL WAS N/A/NTA/NED + 37/NCHE3
( ABour /S" /MCHES BELOW THE /C STEAM NO22LE).
4.
THE MS/v's WERE REOPENED.
THE "SWEU."
MOMENTAR/L Y BR006HT T~NE L.EVEL Af80VE rME
/C
.STE4M No itLE.
S.
/C W4 TER MAMMER o cc u/rp ro d.
NO 3/6N/Ftc4Nr DAkr46E -
ScME /_oosEN/N6 of
)9NCHOR BOL 7S.
e
Docket No. 50-245 Attachment No. 4 Millstone Nuclear Power Station, Unit No. 1 Isolation Cor. denser System Water llammer Events April, 1982
PROPOSED DESIGN CHANGES TO PREVENT REACTOR OVERFILL FOLLOWING PLANT TRIP The implementation of several design changes, which we believe will climinate repetition of water hammer in the isolation condenser piping due to reactor vessel overfill following plant trip, are being considered. The proposed changes and expected benefits, as well.as the schedule for evaluating and implementing these changes, are previded.
1.
High Reactor Level Feed Pump Trip We intend to implement a feedwater pump trip upon high reactor level. The setpoint for the trip has not yet been calculated; however, it will be above the normal reactor vessel level maintained during power operation (approximately 510 inches above vessel bottom), and below the bottom of the inlet nozzle to the isolation condenser (549 inches above vessel bottom).
Should reactor level drop to the FWCI initiation setpoint (434 1/16 in. above vessel bottom), the pump trip signal will clear, and reactor feed pump (s) will automatically restart. A FWCI actuation signal will override the feed pump trip signal.
We feel that this design modification will prevent water hammer in the isolation condenser, and allow sufficient time for the operator to re-establish reactor vessel level control following a trip, without compromising the availability of the FWCI system.
An option considered in lieu of tripping the feed pumps, was automatic closure of motor operated isolation valves at the dis-charge of the reactor feed pumps. This option was rejected due to the long cycle times for the valves (approximately 55 sec. from full open to full close), and their fail-as-is con-figuration.
It was felt that tripping and restarting the reactor feed pumps is more reliable, and will result in smaller level oscillations, than cycling the isolation valves.
2.
Reactor Level Setpoint Setdown Following Reactor Trip Millstone Unit No. 1 operates with normal reactor water level at approximately 510 inches above the bottom of vessel. After a reactor trip, reactor level decreases rapidly due to collapse of voids. This causes the feedwater control valves to go to the full open position. Thus when void collapse terminates, the reactor vessel fills at a rapid rate and overshoots its programmed level.
By automatically lowering the programmed level setpoint followira reactor trip, the feedwater control valves will go to their full open position later, and start closing down sooner, thus reducing the tendency to overfill the vessel. We have not finalized the value of the lowered reactor level setpoint, but expect that a value in the vicinity of 490 inches above the vessel bottom will be acceptable.
This design change is still being evaluated to determine its actual anticipated value in preventing reactor vessel overfill following a scram. The results of our ongoing evaluation will determine whether this design change will be implemented.
1 3.
Change In Reactor Water Cleanup System Isolation Setpoint In the Millstone Unit No. I design, the RWCU system is isolated on a low reactor water 1cvel setpoint of 482 1/16 inches above the vessel bottom. Following a reactor trip, the reactor level initially shrinks (due to collapsing voids), to approximateif 480 inches above the vessel bottom, thus resulting in RWCU system isolation.
We propose lowering the RWCU system isolation setpoint below 480 inches, but no lower than 434 1/16 inches above the vessel bottom. This will allow the RWCU system to be availabic shen the reactor vessel begins to refill. The operator will thus have a letdown path available to prevent reactor vessel over-fill during the low feedwater demand conditions shortly after shutdown.
Isolation of the RWCU system on low reactor water level is a safety related function, since it is required to ensure con-tainment isolation after an accident. Prior to making any modifications to the existing RWCU isolation setpoint, we will ensure that the proposed change does not result in unacceptable radiological consequences following an accident.
4.
Low Flow Feedwater Controller Improvements As part of our effort to comply with NUREG-0619 limitations on feedwater nozzle crack growth, we intend to implement certain improvements to the low flow feedwater control system during the 1984 refueling outage. These improvements include provisions to allow remote actuation of the low flow feedwater control system from the control room. Provisions to maintaih constant reactor vessel water icvel automatically, with the low flow feedwater control system in operation, will also be provided.
In addition to minimizing thermal stresses on the feedwater nozzles, these modifications will allow precise reactor level control between the time the low flow feedwater control system is put into operation, and shutdown cooling operation is initiated, thus preventing reactor vessel overfill.
O PROPOSED SCllEDULE FOR IMPLEMENTATION OF Dr. SIGN CllANCES I.
Items 1 and 3 Subtask
' Completion Date Complete preliminary Engineering May 1, 1982.
Finalize design changes June 30, 1982 Purchase material
- September 30, 1982 Implement design changes
- November 1, 1982
- Note:
These dates assume that the material necessary-to implement the necessary changes can be obtianed in time for the September 1982 refueling outage.
Should this prove impossible, we will submit a revised schedule.
II.
Items 2 and 4 Subtask Completion Date Complete preliminary engineering May 1, 1982 Finalize design modifications November 1, 1982 Purchase materini January 15, 1984 Implement design changes **
1984 refueling outage
- Note:
This date has not yet been established. We expect that it will occur in the first half of 1984.
e M
Docket No. 50-245 Attachment No. 5 Millstone Nuclear Power Station, Unit No. 1 Isolation Condenser System Water llammer Events April, 1982