ML19332A522
| ML19332A522 | |
| Person / Time | |
|---|---|
| Site: | Crane  |
| Issue date: | 09/10/1980 |
| From: | Hukill H METROPOLITAN EDISON CO. |
| To: | Novak T Office of Nuclear Reactor Regulation |
| References | |
| TLL-447, NUDOCS 8009160329 | |
| Download: ML19332A522 (10) | |
Text
i Metropolitan Edison Compa iy Post Office Box 480 g^g Middletown, Pennsylvan:a 17057 717 944-4041 Writer's Direct D:al Number September 10, 1980 TLL 447 Mr. T. M. Novak Assistant Director of Operating Reactors Division of Licensing U. S. Nuclear Regulatory Commission Washington, D.C.
20555
Dear Sir:
Three Mile Island Nuclear Station, Unit 1 (TMI-1)
Operating License No. DPR-50 Docket No. DPR-289 TMI-1 Fuel Assembly Holddown Springs This letter and attach = cats are in response to concerns regarding TMI-l fuel assembly holddown springs per your letter of July 1, 1980.
Based upon reviews of spent fuel pool and Cycle 5 core verification tapes and on special video examinations of assemblies in the spent fuel pool, no failed or damaged holddown springs have been identified at TMI-1.
- Also, evaluations performed by B&W have shown that no significant safety concerns exist for IMI-l operation even with the unlikely occurrence of broken holddown springs. Under the most extreme thermal-hydraulic core conditions (5000F, 4-pump startup) for liftoff no TMI-l assemblies are calculated to lift.
Any piece of a broken spring that escaped from the upper end fitting and large enough to do significant damage to any RCS components would be detected by the Vibration and Loose Parts Monitoring System. Reactivity effects of hypothetically lifted assemblies would be minimal. No excessive vibrations would occur due to lifting of an assembly and lateral repositions would be restricted by adjacent assemblies (or core baffles). Any mechanical damage to a lifted assembly would therefore, be limited to minor wear phenomena.
It may be (mphasized that B&W investigations of the failed springs at Davis-Besse 1 allow the conclusion that the cause was not generic and therefore failures would not be expected to occur at TMI-1.
This is supported by the results of TMI-l spring examinations.
Sincerely, A
f 001
~
E'.' D. Hukill.
///
Director, TMI-l HDH:DGM:lma Attachments cc:
J. T. Collins D. DiIanni H. Silver R. W. Reid B. H. Grier B. J. Snyder Met tm'8 0 lf9" scr C0rrrar/.s a '. e"ter c' me Ge"ers c ::c UUt es Svs em r 16ug '
4 ATTACHMFliT RESPONSES TO HOLD-DOWN SPRING QUESTIONS TO LICENSEES 1.
(If the reactor is down for refueling and the reactor vessel head is of f) Examine all fuel assembly holddown springs in the core and in the spent fuel pool and report.the number and extent of
' damage on the springs and affected assembly components.
or (Alt.) (if the reactor is operating.) Review video tapes of the core from the last refueling and examine all assemblies in the spent fuel pools. Report the number and extent of damage on the springs and affected assembly components.
Response
1.
No failed or damaged holddown springs have been identified in any TMI-l fuel assemblies. This is based upon review of spent fuel pool and cycle 5 core verification video tapes and special video examination of 104 fuel assemblies in the spent fuel pool. There is a total of 208 assemblies in the pool and work will continue until all springs have been examined. Examinations were delayed due to fuel handling bridge repairs and Tech Spec restrictions on removal of centrol elements frem the assemblies. These delays have been resolved. A follosup report on results of the completed examination will be submitted if required.
Notication will also be made if any defective springs are found.
It may _e noted that 10 assemblies of the thrice-burned Batch 4 were examined and no spring defects observed.'
25 Batch 4 assemblies are included in the cycle 5 core. Based upon these examination results and B&W investigations of the causes of the failed springs in Davis Besse-1 there is little reason to expect that any of the spring heats in cycle 5 will be detective.
a f
l l
2 2.
Provide a discussion of the safety significance of operating with one or more broken springs in the core. Your discussion should include, but not necessarily be limited to the following:
O j
l t
7
- tv -
T+% m
-'9--P'1"='-
- M'-
3 2.s.
Assume the holddown spring is broken, provide an estimate of the flow conditions under which the assemblies would be levitated.
(Provide the value of the force required to lift the assembly, the flow conditions under which that force would be supplied, the number of coolant pumps that would be in operation under such condi-tions, and 'the schedule of reactor operations under which such conditions might have been achieved.) Contrarily, demonstrate the margin between the assembly weight and the calculated maximum applied lift-off force, if there is such margin.
Response
2.a.
A break occurring in a holddown spring will result in a decrease in holddown force which is a function of the location of the break and the degree to which the coils become misaligned. To quantify this decrease, tests were run on springs cut at typical break locations. The test fixture was an upper end fitting complete with guide tube nuts, a holddown spider and a simulation of the upper grid plate pads.
Two springs were prepared for the test: one cut at the transition point from the dead coil to active coil, the second was cut at a location k coil from the transition point towards the mid-coil of i
the spring, A series of six tests were run using the two springs.
Each was tested with: 1) no breaks,
- 2) one break at the upper location, and 3) breaks at both upper and lower locations. The retained holddown force (at 100% power, BOL) exceeded 300 pounds for three of the four break configurations. Only the configuration with two breaks,-each h coil from the transition, resulted in a significant loss. The retained holddown force for this extreme configuration was 64 pounds.
The broken springs observed to date are characterized by breaks within k coil of the top or bottom transittan point. Two springs have contained breaks at both top and bottom. The single exception to this pattern occurred at Davis Besse. Here, one spring was broken at the lower transition and again approximately 1/3 coil towards the spring mid-coil. The broken piece was wedged between the dead coil and the upper coils. From the test results, it i.
estimated that this spring would have retained approximately 100 pounds holddown force without the wedged piece (which should in-crease holddown).
Thus, the tests encompass the observed spring breaks. Based upon these results, it can be concluded that a broken spring is likely to retain from 64 to 500,ounds holddown force. The break most frequently observed would provide a retained holddown force near the upper end of this range.
4 It is significant to note that each of the broken holddown springs observed to date has held the spring spider against the remaining plugs. This pinned condition is only possible due to some retained preload to the spring.
Springs in this condition are~ expected to develop a minimum of 100 pounds retained holddown force when extrapolated to operating condi-tions due to additional preload from the reactor internals.
'The flow of coolant water through the core during normal operation produces large hydraulic forces on the fuel assemblies.
The actual forces imposed during operation will depend on the total flow through the core and the distribution of coolant flow to the various assemblies.
The total mass flow is a function of the coolant temperature and the number of reactor coolant pumps (1-4) in operation. The flow distribution is affected by (1) the power distribution, (2) the assembly geometry (i.e., control rod, orifice rod, BPR, open guide tube), and (3) the location within the core (peripheral / interior).
Counteracting these large hydraulic forces are the fuel assembly weight (approximately 1510 lbs in air), the supplemental force supplied by the preloaded holddown spring and frictional forces exerted by the reactor internals and adjacent fuel assemblies.
The holddown spring is sized to provide a minimum force under the most adverse conditions (coolant temperature, irradiation exposure, dimension tolerances, etc.), without consideration of frictional-forces. The force required to lift a fuel assembly is assumed, for this evaluation, to be equal to the weight of the fuel assembly in water (that is, it is assumed that there is no holddown force available from the holddown spring or from frictional forces).
Based on a nominal system flow rate of 114% of design (where the design rate is 352,000 GPM) the maximum net lif t force en any assembly is +55' pounds in a core configuration corresponding to that of TMI-1 (i.e., 69 control red assemblies and no burnable poison rod assemblies). That is, the net vertical (upward)' force on the fuel assembly, taking credit for only the wet weight (no spring force),. is 55 pounds..The net force on fuel assemblies in control rod locations varies'between -129' pounds to -220 pounds indicating that all'of these'cssemblies have significant margin to lift even if no credit is taken for the spring's holddown force.
- Of the core locations not occupied by control rods 68 have a net positive lift force with no spring force considered. Recalling i
that all broken' springs observed to date have retained at least i
100 pounds holddown force it can be concluded that no fuel assembly-
'I lif t would be predicted for normal operation with broken holddown springs.
For TMI-1, with a measured nominal flow rate of 110% of
' design, the margin is significantly increased.
5 i
The hydraulic forces on tre fuel assembly generally increase with decreasing temperature. The phenomenon is due to the increased fluid density at the reduced temperature. Therefore, the most severe lift condition is the lowert temperature at which four reactor coolant pumps are in operation. The holddown spring is sized to accommodate l
this limiting condition--the fourth pump startup.
The maximum net lift force at the fourth pump startup temperature of 500 F (114% flow) is +129 pounds. For this condition all control rod locations maintain positive holddown without the benefit of the spring force. Lift forces on assemblies in control rod locations vary from l
-47 pounds to -148 pounds. Assuming a minimum retained holddown force l
of 100 pounds for a broken spring only 44 assemblies would be predicted l
to lift for this extreme temperature condition; however, lifting tider this l
transient condition is not a significant concern and will not cause significant fuel assembly wear or damage. For TMI-1, with a measured nominal flow rate of 110% of design, the number of assemblies predicted to lift for this extreme condition is zero.
l Due to the increase in holddown requirements with decreasing temperature, l
transients which cause an overcooling of the primary system are the most l
limiting with respect to fuel assembly lift.
Such transients will, in general, be terminated before reaching a condition analyzed for the fourth reactor coolant pump startup. However, if the primary coolant temperature 0
l were to go below 500 F and all four reactor coolant pumps were inadvertentl'/
l left on, the required holddown force would continue to increase at a rate i
of approximately 120 lbs for each 100 F the primary coolant temperature drops below 5000F. Without the force from holddown springs, a significant l
number of fuel assemblies would be expected to lift under this condition; l
however, lifting under these transient conditions is not a significant l
concern and will not cause significant fuel assembly wear or fuel damage.
l l
Operation with less than four reactor coolant pumps is expected to pro-l duce no fuel assembly lif t regardless of the spring holddown force l
available. The maximum net lift force (with no credit for spring force) f for three pump operation at 100 F is - 75 pounds, indicating significant l
margin to lift. This temperature was chosen for the evaluation to con-servatively accomodate all possible three pump operation. Due to the demonstrated conservatism shown for 3 pump operation no evaluation of 1 or 2 pump operation is required; fuel assembly lift will not occur for these pump operating conditions regardless of the spring force available.
l l
t
6 f
2.b.
Have any loese assembly parts (i.e., broken springs, pieces of cladding) been observed anywhere in the primary system? Describe your methods for i
loose part detection. Are there installed noise detectors capable of
[.
detection of broken springs, pieces of cladding, or vibrating assemblies?
l
Response
l-l 2.b.
No loose fuel asrembly parts have been observed in the TMI-1 primary system.
TMI-1 has a B&W Vibration and Loose Parts Monitoring System (VLPM).
This system was installed during the last refueling outage replacing l
the original Rockwell International VLPM system utilized for the first four cycles of operation. The VLPM has both recording and r.udio capa-bilities.
The system has twelve (12) channels dedicated to loose parts detection:
two (2) on the upper reactor vessel shroud; two (2) each at the "A" and "B" OTSG upper tubesheets; one (1) each at the "A" and "B" OTSG lower tubesheets; one (1) each on the #5 and #13 incore instrumentation guide piping; one (1) each on the main feedwater pipes. Detection of a loose part would trigger an alarm in the control room. The operator would then check the VLPM monitor located in the relay room to determine the specific channel that detected the loose part.
Also, the alarm starts continuous audio tape and chart recordings at the VLPM monitor. Audio monitoring capability is available both in the control room and at the VLPM console. An X-Y plotter and a spectrum analyzer are available for on-line diagnostics.
Maximum sensitivity of the system is 0.01 g within a 0.1 g range. Present settings will detect an impact of less than 0.5 ft. lbs within three feet of a sensor.
For expected flow conditions in the RCS the system would detect a loose part with a mass as low as 0.25 lbs. Any portion of a broken spring smaller than this would not be expected to cause signifi-cant damage to any RCS components. The largest spring piece that could escape the upper end fitting has been estimated as one full coil or a mass of approximately 0.6 lbs.
The twelve sensor channels of the VLPM are monitored at least once per shift by operations personnel. Each channel is listened to for unusual noises characteristic of loose parts.
It has been estimated that the VLPM system could probably detect fuel assembly vibrations of a magnitude sufficient to cause damage to the assembly. However, since the actual motion that an assembly can ex-perience is-limited by adjacent assemblies (or core baf fles) to ampli-tudes much smaller than that of the damage threshhold, there is only a small possibility that such vibrations would be detected.
- .- :- - +
. m _ g.:_ w..
g,.
7 2.c.
Have there been any excore or in-core neutron detector indications of levitated assemblies?
Describe the expected reactivity effects that would result from lift-off or reseating of assemblies with 'roken holddown springs.
What efforts are being utilized to detec
- oose assemblies by either nuclear or mechanical monitoring devices?
Response - There have been no observed excore or in-core neutron detector indications of levitated fuel assemblies at TMI-1.
2.c.
Normal steady state operation with lifted fuel assemblies does not represent a safety concern.
If a lifted assembly were to reseat during operation a small increase in core reactivity would occur due to the relative motion between the fuel assembly and a partially inserted control rod.
Conservative calculations have predicted that a fuel assembly lifting 1.5 inches (the maximum possible) would change the core reactivitys.002% Ak/k at hot full power ands.006%
Ak/k at hot zero power. The limiting reactivity insertion would occur if the fuel assemblies in all control rod locations were lifted the maximum distance. As discussed in the response to question 2(a) assemblies in control rod locations retain positive holddown during normal operation even with no spring force. Thus, this limiting reactivity insertion is a hypothetical event. For this condition a maximum reactivity insertion of onlys0.1% ak/k at HFP is predicted.
The resulting transient would, at worst, be characterized by a small, rapid i'ncrease in neutron power tripping the plant on high flux in the first few seconds of the transient. The transient would also result in a small increase in reactor coolant system pressure with no change in core inlet temperature for approximately 10 seconds
(
(one loop transit time). Thus, even this hypothetical reactivity insertion does not significantly affect the steady state and transient safety analysis; the potential reactivity insertion from a small number of spring failures, it lif ting were to occur, is shown to be of no consequence.
2.d.
Have there been any observed indications of lateral repositioning of loose assemblies? Describe the methods used to detect lateral assembly motion. Describe the degree of lateral repositioning that is physically (dimensionally) possible af ter lif t-off.
What are the postulated worst-case effects of a laterally displaced assembly?
2.e.
(i)
Describe the degree of "Ucrst-case" =echanical damage that would be expected as a result of movement of a " loose" assembly (one with a broken spring) against adjacent assemblies, core baffle, or other core components.
~
8 2.e.
(ii)
Discuss the results of flow tests or other experiments that have provided measurements of axial or lateral vibratory motion of an assembly af ter lif t-of f or that would otherwise support the response to question 2e(i).
Response - There have been no observed indications of lateral repositioning of fuel assemblies at TMI-1.
2.d.& e. As discussed in response to question 2(a), fuel assemblies with broken holddown springs would not be predicted to lift-off during normal operation. Furthermore, there have been no indications that any of these assemblies did lift-off at Davis Besse-1.
Three fuel assemblies containing broken holdown springs were visually examined.
No evidence of lif t or of wear from lif t or lateral displacement was found; no fuel assembly damage of any kind was fo"nd.
A fuel assembly suddenly experiencing a loss of holddown could move upward a maximum of 1.5 inches, with a corresponding impact energy level of less than 50 ft-lbs.
This level of impact ir far below the energy necessary to damage the fuel assemblies. For example, LOCA analysis has shown that the fuel assembly can withstand impact energies in the range of 500 ft-lbs.
Thus, gross impact of fuel assemblies can be climinated as a cause for concerg but there is the possibility of lower level vibrations which could cause some wear.
Also, there is the possibility of spacer grid mismatch due to lifting of one assembly while its neighbor remains seated. The fuel assembly can lift up to 1.5 inches at beginning of life whereas 1.2 inches lift will result in the spacer grids outside strips no longer matching up.
Long term operation under this condition would, at worst, result in damage to some peripheral fuel rods. There is no possiblity of damage resulting in noninsertion of control rods since the guide tubes are protected by two rows of fuel rods.
Horizontal vibration of the fuel assembly while in the lif ted condition may be more pronounced at the lower end fitting since it may not be held tightly by tha grid pads. Lateral motion in which two adjacent assemblies contact at the lower end fitting is possible and could cause wear on the lower end fitting. However, the lower end fitting has thick cross sections which can withstand significant wear without loss of function. Peripheral assemblies might contact the core baffle plates but again wear would not be a significant problem. The lower end fit-ting of a fuel assembly which is postulated to lift th inches can raise up onto the chamfered lead-in surfaces of the guide blocks such that 0.4 inches of lateral repositioning could theoretically occur. However, lateral repositioning is nominally limited to the clearances between the lif ted assembly and adjacent seated assemblies or baffle plates which are 0.05 inches and 0.1 inches respectively.
9 The upper end fitting will remain closely aligned by the upper grid pads at all times. Lateral vibracion would not be expected to increase.
For this reason upper end fitting wear or control component wear would not be expected to be any greater than the low levels experienced during normal operation.
There have been several tests run o determine the flow required to cause fuel assembly lift. These tests also provide an indication of assembly vibration levels in the lif ted condition. They were run in the Control Rod Drive Line Test facility (Alliance Research Center),
i which is a single fuel assembly test loop simulating reactor flow, temperature and pressure. A displacement transducer was used in determining fuel assembly lift. During these tests, the holddown spring remains uncompressed since the maximum loop flow is incapable of lifting the assembly with the spring compressed. The flow is increased in small increments until the assembly lifts at which point the flow is then varied to determine the lift velocity as accurately as possible. There has been no indication of vertical oscillation of the assembly during these tests. Also, the fuel assemblies were examined after each test and no evidence of impact to wear has been found.
These results indicate that severe vibration will not result for a lifted
- assembly, l
l
c-
_.a_.-
10 3.
Provide a description of the cause of the failures and corrective action to reduce the likelihood cf future failures at your facility.
According to B&W, the cause of holddown spring failures at the Davis-Besse Unit 1 plant was an improper material condition characterized by a coarse outer grain structure. Coarse grain structure is indi-cative of less fatigue resistance. The coarse grain material pre-cipitated fatigue crack initiation. The mechanism of failure was then fatigue propagation followed by the secondary effects of stress corrosion cracking and final fracture.
Corrective action at Davis Besse consisted of replacing the springs from this heat of material. Replacement springs were made to a current specification which controls grain size to obtain uniform fine grains to provide increased fatigue resistance.
Future actions to reduce the likelihood of failures at TML will include:
1)
Assurance that future reload fuel assemblies contain holddown springs fabricated according to B&W's revised specification whfch calls for improved material and process control. This material will be more reciscant to fatigue and stress corrosion cracking.
2)
Establishment of a fuel surveillance program to inspect assemblies during each outage for indications of spring failures or draage to other assembly components.
3)
In the longer term, and if available, require the use of the B&W improved end fitcing on all B&W reload assemblies. The design of this improved end fitting is presently scheduled to be completed in early 1982.
m