ML19316A117

Attachment A:Structural Analysis of Worn Surveillance Specimen Holder Tubes

ML19316A117
Person / Time
Site:	Oconee
Issue date:	04/13/1976
From:	DUKE POWER CO.
To:
References
NUDOCS 7911280641
Download: ML19316A117 (37)

Text

i s

i

\\

s ATTACIDIENT A STRUCTURAL ANALYSIS OF WORN SUTIVEILLANCE SPECIMEN HOLDER TUBES m

C

$~'

(.

^

cl (-

j c'

/

/

C.-L.

."( Y3 f

a 61

'l ci_

Lj - / 3 -- 0 (

\\

lii:...g) w."a trilGh6 M%gigpixijp)LL jsLL'Us I

QL.

l o

0 q911~8 7

q

)7

/

u

l*

For a preliminary look at the SHT in a damaged condition, we may ~ use the approach given by Gorman (1).

We treat the damaged section as a beam simply supported at each end and assume the damage is located in the center.

If we take a worst case approach to the dimage, we can approximate it as shown:

?s

'I2 d, : s.372 o

~

d, 1d.

A i

de

• I 018..

gt

\\ I

\\

l

~Q s

e We consider the damage to be severe enough to leave only four ligaments with a width of 1/2 inch, a thickness of.120 inch and l

a length of 1/2 inch.

j The moment of inertia of this reduced section is then given by:

l' 6 tx2 dx =.3457 inches" I=4

'uy where t is the wall thickness =.120 inch.

The cross sectional area of the remaining ligaments is then:

A = 4 x t x 1/2" =.24 inches,

2 Treating the damaged beam as being symmetric about the discontinuous section, the first natural frequency is given as (28 )2 EI 1

yl I

l " 2r(L)2 1^1 9

where a is a constant dependent upon the reduced bending rigidity l

and the mass per unit length.

Using these constants (1) Gorman, D.

I., Free Vibration Analysis of Beams and Shafts, John Wiley and Sons, N.Y.,

N.Y.,

1975.

2 EI 5

8A b

22 22

=.67 2nd

&=

.66 a-g

=

8 ^1 11 1

and u = 1-y =.993 where y is the normalized length of the damage, we can choose 8 from a linear interpolation of the tables listed 7

in Gorman.

The subscript I refers to the undamaged section and 2, the damaged section.

We choose B to be 1.5628 (which should y

be accurate to a few percent) based on an overall beam length of 68 inches.

We can now predict the ratio of the first natural frequencies as 2

EI f (damaged)

(281(damaged) 2 2xL

~

1 1

p A 11

=

f (good)

(281(good) 2 21Lj gy 1

I1 8 ^1

{

1 f a (damaged)b y

~

i I81(g000 j I I.5628

=.9898

=

i.5708 /

1 This is approximately a 15 difference in natural frequencies.

The change in amplitude can be found from the solution of the system of equations resulting from the matching of the boundary conditions at the interface of the discontinuity.

The amplitude at midspan is now given as

'r=B2 cos )

+

D cosh 4 2

=B2+D2

= 3.073 - 2.017 = 1.056 Forming the ratio between this value at that of the undamaged beam we have

~~ ~

~~

r(damaged) 1.056

= 1.056 r(good) 1 This represents an approximate amplitude increase of 5.6%.

Additionally we can simply calculate a stress ratio due to bending.

MC /.

)

a(damaged)

T (i

'4/

T ; :c)i a(good)

MC For a moment that remains constant (nearly the case if the damage length is small), this reduces to o(damaged)

I(good)

Cfdamaged) a(good)

I(damaged)

C(good)

The moment of inertia of a good tube is j

w(Dj - D{}

I=

l 64 i

j D, = outside diameter = 3.5 inch j

Dg = inside diameter = 3.5

.24 inch I = 1.822 Therefore, the bending stresses will rise by a factor of a(damaged) 1.822 1.372 4.13.

=

a(good)

.3457 1.75 From Topical Report 10,039 the measured stress in this area was found to be 300 psi.

In the damaged section we would expect the stress to now rise to 1239 psi.

The safety factor becomes, for the case of severe damage, SF = 13

= 10.9 239

I 4

This is in addition to any safety factor built in to the allowabic stress 1cvel.

The Euler buckling load is given 'as P

n EA/(L/r) 2 er where r = least radius of gyration and I = Ar2 or r

.3 7

Therefore, r for the damaged tube =

1.2

=

and L/R = 86.67.

Now P

= nr 26 x 102 x.24/(86.67)2 er P

= 8199 lbs er In this case n = 1 (pinned-pinned ends) which is the worst case.

This buckling load is much greater than the applied preload even if we assume a large effect due to differential thermal expansion.

The actual natural frequencies to be expected can be calculated, again using Gorman.

The first natural frequency of the lower (ogee) section can be approximated by 1

18 ) j EJ g 1

f w

2xL2 E = Young's Modulus = 26 x 106 psi I=v(D$-Dy) 1.822

=

64 g = 386 in/secz L = 68 in

$1 = 3.927

" " " steel * " water * " virtual Ib/in

2 (D-D[]=.3606lb/in ws"Ps wD2

" water = p i

.2174 lb/in

=

water v

xD2

.2506 2b/in

" virtual " Awater 4

=

(3.927)2 26 x 106 x 1.822 x 386 f

78.84 hert:

=

2x(68):

.8289 Accounting for the damaged section we have il damaged = 78.84 x.9898 = 78.04 hertz For the upper section we have L = 100 in steel =.3073 lb/in w

.1558 lb/in w

=

water ud2 2+a2

" virtual " Dwater 4

2.2198

=

b2-a2 4

4 w(D -D )

1 I=

= 1.1276 g

f

, (3.927)2

' 2 6 x 106 x 1.1276 x 386

= 15.94 Wu 2r(100)2 2.6829 Even with the heavy damage previously assumed, this frequency will remain virtually unchanged.

T'he difference in predicted natural frequencies (a factor of 5) also tends to substantiate the choice for S of 3.927 which is the value used for clamped-simply supported y

end conditions.

..- ~.-.-.

6-The analysis of the wear mechanism is extremely difficult.

On the surface we can say that the spacer disks wore through the C

SHT.

This is obvious and therefore the pushrods with disks must be removed.

The excitation mechanism or forcing function is not so readily explained.

One possibility is explored below.

Consider the fellowing mcdel for the pushrod assembly:

I 2

3 4

5 p

p s

h 2 7. 9 J 7s "

J2 "

38.375 '

3 7. 3 75 ~

3 9,3 rs a i

Rod D =.5 inch I = wD4

. 0 03 06 8 in'+

a TT l

E = 26 x 108 psi g = 386 in/sec2 8

3 steel =.283 lb/in p

p

= 45 lb/ft3 water steel =.0556 lb/in w

virtual =.0051 lb/in w

total =.0607 lb/in w

w EIg P

fy=

y, (2)

(simple-simple supports) 2L2-w Per where P

= Euler buckling load er w EI (1st mode, simple-simple ends)

=

L2 and the effect of the spacers is neglected.

Depending on what spacers are against the tube wall at a (2) Timoshenko, S.,

Young, D.

H., Weaver, W., Jr., Vibration Problems in Engineering, John Wiley and Sons, N.Y.,

N.Y., 1974.

j.__

-1 given time, the natural frequencies can vary tremendously.

This, coupled with the variable preload and the pre-bent (installed) condition make any natural frequency calculations subject to large errors.

However, for a quick analysis we construct the following table:

I I

P Span I (P=0(H:)

l (P=200 lbs)(Hz) er (1bs) 1 45.3 40.6 1009 2

34.6 29.7 769 3

35.9 31.1 800; 4

24.7 19.7 549 5

22.8 17.8 509 1+2 9.85 2.91 219 2+3 8.81 0

196 3+4 7.38 0

164 g

4+5 5.93 0

132 l

1+2+3 4.24 0

94 2+3+4 3.45 0

77 l

3+4+5 3.00 0

67 1+2+3+4 2.12 0

47 2+3+4+5 1.79 0

40 I

1+2+3+4+5 1.25 0

28 The mechanisms which may excite these various modes are summarized in the following tables.

Thermal shield radial deflection peaks

H_z, SHT Span number with nearby resonant frequencies P=0 P=200 9.5 1+2,2+3 12 15 17.5 5

20 4

23 5,4 29 2

31 3

37.5 3

1 4

,o As we have seen, the latural frequencies vary with the end load and spacer contact.

Although the frequency predictions of the pushrod are certainly not precise, it does seem apparent that they are of a low order of magnitude and at least in the region of possible forcing mechanisms.

The radial displacements of the thermal shield are quite small (only on the order of several mils peak to peak) and with the reasonably well damped system expected inside the SHT, the aligning of the pushrod. natural frequencies with external forcing functions may not be sufficient to cause excessive wear.

However, there is a small leakage path through the SHT to the upper head region.

The expected pressure drop from inlet to outlet conditions, it is felt, is sufficient to cause a localized velocity between the spacer disk and SHT wall of somewhere in the neighborhood of 20-30 feet per second.

This flow over the spacer disk should be sufficient to cause vortices to be shed with frequencies of at least several hundred hertz.

This could cause the spacer disk to vibrate with very small amplitudes.

These potential " grinders" may now be moved around the inside of the SHT by some as yet undefined coupling between the thermal shield motions and the overall pushrod.

There may be a hydroelastic coupling between the SHT walls and the pushrod or perhaps the specimen capsule itself transmits the SHT support motions to the pushrod.

It does seem likely, however, that the coupled problem of vortex shedding from the disks and overall pushrod motions caused the excessive wear.

)

ATTACIDfENT B ADDITIONAL DETAILED RESULTS FROM OCONEE 1 HFT INTERNALb VIBRATION MEASUREMENTS i

r i

l

3.4 RESULTS M?D DISCUSSION FROM THERMAL SHIELD MEASURDETS 3.4.1 Thennal Shield Accelerometers l

3 4.1.1 Location of Themal Shield Acceleremeters. Six biaxial accel-erometers were installed arcund the periphery of the themal shield. Three j

of the six accelerometers were located in the 11-X quadrant. These accel-erometers were denoted as TS2, TS3, and TS4.

Each of the other three quad-rants contained one accelerometer. The X-Y quadrant contained accelerometer TSS, the Y-Z quadrant contained accelercmeter TS6, and the 2-tf quadrant con-tained accelerometer TSI. The radial direction for each axis was denoted with an R, and the tangential axis was denoted as a T.

These accelerometers perfomed satisfactorily thmughout the entire test program, except for TS2R and TS2T which failed during the first few days of the hot functional test program. Appendix F shows the location of each of the accelerometer axes.

The installation and a further detailed description of these transducers is contained in Volume I.

3.4.1.2 Results fmm the Thermal Shield Accelerometers.

Figure 3.4.1 shows the acceleration amplitude for each themal shield accelerometer axis i

for four-pump and no-pump operation at 530*F. This figure shows the amplitude 2

of the radial accelerometers at four pumps exceeds the amplitude at no pt=tp by more than 20 times. The crest factor for each of these accelerometer axes varies between 2 and 3,which indicates the signal contained no large transients.

I This figure also shcws that the radial acceleration of the themal shield is at least twice as much as the tangential acceleration.

.a.

Figures 3.4.3a through 3.4.3d show spectra response of accelerometer TS1R for four different ptrnp conditions at 530*F. TS1R was chosen as representative of the spectra response from the four other radial accel-erometers. These figures show that the themal shield responds at many fre-quencies in the 10 to 50 H: range.

In general, the frequency content for each a.

pump operating condition is similar, as shown in Figure 3.4.3 and Table 3.4.1.

a- :

3-18 u- :

.S*

4 40 -

.30 Q AMPLITUDE

@RMS b

2 O

4 PUMPS, 5300F P<.20 O PUMPS, 5300F Ew d

n wo F

3

[

h

.10 i

l 7

R A

R R

9 R

E A A 0

0 PUMPS 4 0 4 0 4 0 4 0 4 0 4 0 4 0 4 0 4 0 4 0 L

SENSOR TS1R TS3R TS4R TS5R TS6R TSIT TS3T TS4T TS5T TS67 FIGURE 3.4.1 THERMAL SHIELD ACCELER ATIONS, RMS AND AMPLITUDE

.200 -

0 AMPLITUDE 4 PUMPS, 5300F O PUMPS, 530oF

.o z'o P

<.100 g

=

8 J

a a

l l

R-l l

h h

~

9

~

2

~

0

~

PUMPS 4 0 4 0 4 0 SENSORS LGCX LGCY LGCV FIGURE 3.4.2 LOWER GRID ACCELERATIONS, RMS AND AMPLITUDE

==

3-19 6

1.0 4 PUMPS, 5300F NORMALIZED TO 5.34 x 10 g fg, 4 2

.50 a

{

0 I

O I

1.0 PUMPS A1. A2 - 92, 5300F y

NORMALIZED TO 7.83 x 10~4 g /Hz 2

d so

]

u a

~

o

=

c

.i 0

I On b

i i

{

! 1.0 ' PUMPS A1 B2, 5300F l

I$

NORMALIZED TO

,,i.'

4.31 X 10'4 9 fg, f

2 a<

> 50 f

x c

l i

E L

['

0' I

I 1.0 PUMPS A1 - 81, 5300F a.

NORMAllZED TO 4.99 x 1d g /Hz 2

.50 j,

d II.

O t

1 i

0 to 20 30

.40 30 FREQUEllCY, Hf I

w.

8 e

3-20 e

_... _.,... - - -. + - + -

TABI.E 3.4.1.

DEPPAL SilIELD RADIAL KUI.ERATinNs Frequency Content 2

4 Peak g /II:. x 10 Test Condition Frecuency TSIR TS3R TS4R TSSR TS6R 4 Pu.gs 530*F 12.0 1.51 1.82 1.40 1.80 2.39

~

Tape 116 Set 174 15.0 4.66 7.00 3.91 8.11 7.71 17.5 1.30 0.20 2.33 3.85 4.68 20.0 1.16 1.61 0.56 0.59 8.08 23.5 5.34 11.0 6.41 12.9 15.6 27.5 1.44 1.77 2.91 6.31 2.02 29.0 2.60 2.41 2.91 1.56 1.84 31.0 1.71 2.62 1.56 1.65 37.0 4.38,

11.5 6.00 12.0 9.00 44.0 0.66 0.84 0.84 3.21 l~

49.0 3.90 4.40 1.75 0.46 I

Pumps Al-A2-B2 12.0 1.16 1.50 Tape 216 Set 187 15.0 4.45 5.62 19.5 0.93 t

20.0 0.51 0.46 21.0 0.40 23.0 5.48 15.5 31.0 1.78 1.93 35.0 1.13

~

37.0 7.88 2.25 38.0 3.63 13.4 48.0 2.05 49.5 1.71 4.07 Pumps Al-B1-B2 9.5 0.32 0.44 Tape 117 Set 181 12.0 0.96 3.22 15.5 3.00 8.90

~

18.5 0.37

~

20.0 0.38 0.60 21.0 0.25 23.5 0.31 15.4 29.5 1.34 1.61 4

3-21

~

Tablo 3.4.1 Thermal Shield Radial Accelerations (Cont'd)

Test Condition Frequency TSIR TS3R TS4R TSSR T56R Pumps Al-B1-B2 31.0 2.88 1.66

-(Cont'd) 36.5 3.32 37.5 0.43 2.84 38.5 17.7 3.38 Purps Al-B1 9.0 0.46 0.38 g

Tape 117 Set 180 12.0 1.03 0.63 15.5 2.19 7.40 17.5 8.90 20.0 0.45 0.62 23.5 4.25 12.1 i

29.0 3.97 3.16 32.0 1.51 1.61 37.0 3.63 5.63 38.0 4.93 13.4

~

l i

t Pungs Al-B2 530*F 9.5 0.36 0.32 Tape 117 Set 184 12.0 2.19 3.64 15.0 5.14 8.04 j

17.5 0.66 0.20 20.0 1.33 1.07

+

23.5 5.75 12.9.

29.0 3.97 2.68 31.0 2.74 1.82 36.5 4.38 1.98

~~

38.0 5.34 17.6.

Pumps Al-A2 530*F 9.5 0.16 Tape 117 Set 176 12.0 2.06 0.75 15.0 5.96 3.86 17.5 0.42 20.0 0.34 0.20 21.0 0.19 0.31 23.5 4.66 1.45 31.5 1.78 0'.91 36.5 2.57 1.02 38.0 2.81 7.40 40.5 0.38 0.42 3-22

~~

ser

FJgure 3.4.5 shows the displacement for various frequencies in a spectrum analysis froa 0 to 50 H:. This figure shows that the primary contribution to the displacement of the internals occurs at less than 20 H:.

Tabic 3.4.2 not only summari:es the results for an accelerumeter, TSIR, h it shows the displacemnt results for the other radial accelerometers.

This table also shews that the primary displacemnt of the thermal shield occurs for frequencies less than 20 H:.

It further shows that accelerometer TS3R

~

has the greater maximum amplitude of displacement. This value is approximately 3.85 mils.

Figure 3.4.4 shows the typical spectra response for the tan'gential accelerometers.

This figure shows the thermal shield vibrates at many frequencies in the range of 0 to 50 Hz.

Table 3.4.3 compares the response of the other four operating tangential accelerometers at various frequencies.

It shows that the energy is the same at many frequencies.

3.4.2 _ Lower Grid Accelerometer

+

3.4.2.1 Location of the Lower Grid Accelerometer. Three accelerometer axes denoted LCCX, LGCY, and LCCV were installed on the lower grid.

=

These t

transducer axes measured the planer and the vertical motion of the icwer grid. The location of this accelercmeter is shown in Appendix F.

Volume I further describes the location and installation.

3.4.2.2 Results from the Lower Grid Accelerometer.

Figure 3.4.2 shows the acceleration of the lower grid in three orthogonal directions.

This figure shows that the signal-to-noise ratio of these three accelerometer axes was approximately 5.

The crest factor for accelerometers LGCX and LGCY was approximately 2.

a Figures 3.4.6a, 3.4.6b, and 3.4.6c show the normalized abridged spectra for the three axes of this accelerometer. These figures show that the accel-erometer response was cor: fined to frequencies between 10 and 40 Hz.

Figure 3.4.6c shows one frequency response in the range of 40 to 50 H: on LGCV.

~

Table 3.4.4 contains the frequency response and amplitude of response for the lower grid accelerometer axes.

3-23 h

y

~.

O 1.0 N,

.Jg "i

2

,I 4 PUMPS, 5300F n

g NORMAllZED TO 3.56 x 10-5 2

[ y !.50 g fg, ai gg F,

=<

>U j

M3 50 0

10 20 30 40 50 f

FREQUENCY,Hz FIGURE 3.4.4 TS1T ACCELERATIONS FREQUENCY CONTENT

.i

, 10 E

2' O5 o 2.0 4 PUMPS, 5300F a

lLwO 2 11.0 E

b wb I

n n nin 11 l l

0 10 20 30 40 50 FREQUENCY,Hz FIGURE 3.4.5 TS1R DEFLECTIONS FREQUENCY CONTENT 1

e M

I L

l

.4 ss 3-24 i

t

TABLE 3.4.2.

THEPM\\L S!!IELD RiDLU. DEFLECTIC.'.'S Max. P-P Frequency

% of Cum.

P-P Sensor Defl. nils 11:

Total Defl. nils TS1R 5.5 12 30.0 30.0 1.65 15 27.5.

57.5 1.51 17.5 10.0 67.5 0.55 g

20 7.5 75.0 0.41 E9 23 12.5 87.5 0.69 27-31 7.5 95.0 0.41 37 5.0 100.0,

0.27

'IS3R 7.7 12 28.6 28.6 2.20 15.5 33.3 61.9 2.56 17.5 0

61.9 0

20 7.1 69.0 0.55 23 16.7 85.7 1.28

.=.

29 4.8 9,0.5 0.37

^~

31 4,8 95.3 0.37 7*

38 4.8 100.1*

0.37 TS4R 6.6 12 33.3 33.3 2.20 15.5 25.0 58.3 1.65 17.5 13.9 72.2 0.92 20 0

72.2 0

23 16.7 88.9 1.10 29 5.6 94.5 0.37 38 5.6 100.1*

0.37 TSSR 7.5 12

.27.2 27.2 2.04 15.5 31.8 59.0 2.38 17.5 13.6 72.6 1.02 20 0

72.6 0

23 13.6 86.2 1.02 27.5 6.8 93.0 0.51 31 2.3 95.3 0.17 38.5 4.5 99.8*

0.34

• Not exactly 100*, due to round off error 3-25

Table 3.4.2 Themal Shield Radial Def1cctions (Cont'd)

Max. P-P Frcquency

% of Sensor' Defl. mils H:

Total Cum.

P-P TS6R 7.0 12 31.2 31.2 2.18

_De,fl. mils.

15 22.9 54.1 1.60 17.5 14.6 68.7 1.02 20 0

68.7 0

23 20.8 88.8 1.46 27-31 6.25 95.05 0.44 37 4.17 99.22*

0.29 "Not exactly 100% due to round-off error 4

f a

I e

a f

A e

N ll j

F' j

s

.2 1.,

P

.- a 3-26

~

me me N W w

J

}

TABLE 3.4.3.

DIEPMit SilIELD TNJETIAL ACCELERATIONS J

~

Frequency Content 3

2 4

Peak Value g /Hz x 10 Test Condition Frequency TSIT IS3r TS4T TS5T TS6T

_]

4 Pumps 530 F 11.0 0.279 0.303 0.549 0.300 0.266 Tape Set 12.0 0.279 0.276 0.250 0.323 0.328

}

15.5 0.268 0.380 0.1 70 0.332 0.363 17.5 0.356 1.21 0.280 O.684 0.085

~}

20.0 0.147 0.110 0.356 0.223 0.283 23.5 0.308 0.132 0.333 0.258 0.213 27.5 0.180 0.318 0.333 0.177 0.239 29.0 0.235 0.275 30.5 0.106 0.060 0;072

_]

33.0 0.0 76 0.106 0.088 0.101 34.0 0.110 0.076

0. 11 3' O.104 0.106 36.0 0.238 0.144 0.177 37.0 0.150 0.182 0.185 0.213

~

39.5 0.125 0.163 0.144 0.142 0.137 44.0 0.044 0.062 0.064 0.146 49.0 0.125 0.038 0.102 e

e O

e W

e h

l l

t I

3-27 i

='

(

l.

3.4~.5.4 Response of the Themal Shield. Tests have been conducted to I

evaluate the response of the themal shield @). These results have been compared to the mathematical model developed to analy:e the themal shield

~

in the Oconee I reactor system.

Results of both the test and the analysis are summarized in Table 3.4.6.

As can be seen in this table, good agreement is obtained between the test and the predicted frequencies and mode shapes.

However, for the second ring mode, the experiment and the analytical model did not agree. The frequencies differed by approximately 33 percent.

Neither the mathematical model nor the test accounted for the effects of water surrounding the internal structures.

Tests have been conducted which show that the natural frequency of a mechanical stmcture is alt'ered as a result of the surrcunding fluid as much as 50 percent (10)

Other effects

~

influence the behavior of the reactor's internal stmeture; for example, damping. Three methods were used to determine the mode shapes that existed l

during operation and they are explained in Appendix C.

Consistent mode shapes were found using these three methods. However, at certain frequencies, no consistent mode shapes were discernible as a result of insufficient infomation.

Results of the modal analysis are shown in Appendix C.

l i

The 12 Hz frequency appears to be a combination of a cantilever 6

beam mode and a second ring mode. The cantilever beam mode is seen on the lower grid as a horizontal motion. The orientatien and amplitude en the themal shield seem to vary continuously. This randomness may be caused by the coolant flow or interaction between the lower grid and reactor vessel guide lugs.

The 15 Hz frequency appears to be a beam mode from themal shield accelerometer infomation. However, the lower grid does not exhibit this frequency. 'Ihis suggests that the themal shield acts as a fixed-fixed i'

,j beam, with the lower edge restrained by the themal shield bolts.

g 1

At 20 Hz the themal shield seems to be vibrating as a cartilever

~ '

beam.

The themal shield accelerometers show no preferred orientation,

~

but the lower grid shows a definite orientation in the Y direction.

No explanation has been detemined for this anomaly. The excitation results from the pump blade-passing frequency.

3-40 M

M

.I

The 23 II: frequency is strong on the lower grid, thennal shield, and on the bolts, but it does not seem to correspond to any expected mode shape.

For frequencies above 23 Fl: only one mode shape is presently defined.

'Ihis is the third ring mode at 37.5 H:.

TABLE 3.4.6 EER*t\\L SHIELD NATURAL FREOLTNCIES, IN AIR Calculated Measured Mode Shape Frequencv,11:

Freauencf, it:

n = 1 (cantilever beam) 22 n = 2 (second ring) 45 30.8 n = 3 (third ring) 50 49.9 n = 4 (fourth ring) 68 70.4 n = 5 (fifth ring) 96 95.3 3.5 RESULTS AND DISCUSSION OF INCORE N0ZZLES AND GUIDE RJBES s

3.5.1 Incore No::le Sensors t

3.5.1.1 Lecation of Incore No::le Senscrs. Two incore instrunent 6

4 no::les were instn=ented with four strain gages each. The no::le 34 gages j

were 34N1, 34N2, 34N3, and 34N4, while no::le 38 gages were 38N1, 38N2, 3SN3, and 38N4. The strain gages were located as shown in Appendix F.

All gages performed satisfactorily throughout the tests.

Four static pressure taps were installed on no::le 33. These taps, 33P2, 33P3, and 33P4, were arranged to provide differential pressure from which the lower head flow velocity could be calculated (11)

Volume I describes the installation of the taps.

3.5.1.2 Results from Incore No :le Sensors.

Figure 3.5.1 shows the measured stress amplitude on no::le 34 is less than 200 psi for varying nunber of pumps. Results for no::le 38 are similar. The signal-to-noise ratio varies from 1.5 to 2.6.

The crest factor varies from 2 to 3.

i 3-41

=w*w g'

3.6.3 Discussion of Results from Plenum Cylinder and Upner Grid Sensors The maximtn measured stress amplitude in the plenum cylinder is 160 psi.

These gages were located to measure the maximtm stress in the plenum cylinder, hhile the locations measured may not be the exact location of maxinum stress, it is likely that the maximum stress in the cylinder does not approach the fatigue limit of 13,500 psi.

The upper grid and plenum cylinder sensors do not exhibit the same ficquencies.

This was expected since a lateral acceleration of the upper

~

grid does not necessarily cause significant stresses in the p1'enum cylinder.

The 23.5 H: shown on the upper grid may be a forced vibration.

The pressure variations in the external lines display a 23.5 H: frequency, as do the thennal shield accelerometers. The pressure variations may cause the accelerations of the upper grid and/or the thermal shield.

Also, it is possibic that the movement of one structure ir mechanically transferred to the other through the support.

3. 7 SURVEILIEG SPECDEI HOLDER TUBE SENSORS 3.7.1 Surveillance Specimen Holder Tube Strain Gaces 3.7.1.1 Location of Surveillance Specimen Holder %jtrain Gaces.

Four strain gages were attached around the circumference of the surveillance specimen holder tube.

These gages, SLT1, SHI2, SHr3, and SHr4, were spaced every 90.egrees at one elevation. The location of these gages is given in Figure 1.4.1 and their installation is given in Volume I.

Despite the failure of all four surveillance holder tube strain gages during the i

second test, reliable data at 530*F and four pump operation were obtained, M a except from Sfr4.

EU 9

5 4.4

,.J 3-54

~

l

4 3.7.1.2 Results from Surveillance Snecimen !! older Tube Strain Caces.

C Figure 3.7.1 shows the measured stress amplitude at 530*F for fcur-pump 7

and no-pump creration. The signal-to-noise ratio varies from '1.2 to 2.0.

The crest factor varies from 1.0 to 2.6.

The maximum measured stress amplitude is 235 psi.

The spectral plots of data from all the strain gages show only one discernible peak at 17.5 H:. The plot also shows wide band random stresses between 0 and 50 H:.

Figure 3.7.2 shows the power spectral density plot of Sfr2 at 530*F, four-pump operation.

3.7.2 Shroud Tube Strain Caces 3.7.2.1 Location of Shroud Tube Strain Gaces.

Four strain gages were located around the circumference of the shroud tube.

Each gage was at the same elevaticn and spaced every 90 degrees. R ese gages were designated ST01, ST02, ST03, and STC4.

Figure 1.4.1 shcws the location of each gage, Volume I describes the installation. All four gages perfomed satisfactorily throughout the hot functional test.

3.7.2.2 Results fron Shroud *iube Strain Caces.

Figure 3.7.3 shows the measured stresses from the shroud tube strain gages at four-pump and no pump operation at 530*F. The signal-to-noise ratio varies from 1.2 to 2.5.

The crest factor is approximately 2.6 on all four gages. The maximum measured stress amplitude is 190 psi.

Figure 3.7.4 shows the spectral plot of STO4 at 530 F and four-pump operation. After eliminating the background noise, the only discernible peak is 37.5 H:. The balance of the plot shows random response between O to 50 H:.

3.7.3 Discussion of Results f.om Surveillance Scecimen Holder Tube Sensors The maximum measured stress amplitude on the surveillance specimen holder tube is 235 psi and 190 psi on the shroud tube.

For both stnictures, this is 3-55 O

e r-1 L

" Q AMPLITUDE 4 PUMPS, 5300F QRMS 0 PUMPS, 5300F E = 26.3 psi j

PREDICTED STRESS 10,000 psi 200 7-1 ACCEPTANCE CRITERIA 13.500 psi

[

i i

=

E a

d H

M 100

, _. 1 0

2

~

PUMPS 4

0 4

U 4

0

'i SENSOR SHT2 SHT3 SHT4

._ i i

FIGURE 3.7.1 SURVEILLANCE SPE*:IMEN HOLDER TUBE STRESSES, RMS AND AMPL.TUCE L

r.

1 r

.=

4 1

s l

an d

l 3-56

]

._1 6

I p

o

~

10-4 9-8-

7-6-

5-TAPE 214 4-SET 172 3

2 ptimv d

3-0 526.4 F 2155 psi i

2-4 PUMPS 12.0 mv RMS f4 7

10-3 _

eq 8-k 6-5-

n 4-3-

8

.J l

L g

g 7

2-1 It I

10-2 I

I l

t 0

10 20 30 40 50 l

FREQUENCY,H g

FIGURE 3.7.2 POWER SPECTRAL DENSITY, SHT2 m

em.

Y 3 57 I

t

6 e

Ge

• t

• I e

h O'

b.

i 300

,m.

4 PUMP 5, 5300F 0 AMPLITUDE

, O PUMPS, 5300F QRMS E = 26.3 psi

=

PREDICTED STRESS 10,000 psi 200 ACCEPTANCE CRITERIA 13,500 psi

[

g w

I E

t'

.t jon l-P

?

/

l I

?

?

p l

l l

(

m i

i

/

0

?

C

^

^

PUMPS 4

0 4

0 4

0 4

0 SENSOR STO1 STO2 STO3 STO4 FIGURE 3.7.3 SHROUD TUBE STRESSES, RMS AND AMPLITUDE B

i ir F'

.a.

e.

e a

3-58

• [

.. s 1

1 v

10-5 3

9-j 8-7-

d 1

6-g s_

4-j, p

l)

.j 3-

'J

~1 2-VS

~'

Ihi i

' Y)

=

4

~

N\\~ 10 _

9-

[A'

~

TAPE 214 6-1.,

SE'r 172 S-2mv/pc 526.4 F 4-2155 psi 4 PUMPS 4.2 m y RMS 3-2-

7

.1 10-3 I

I I

I 0

10 20 30 40 50 FREQUENCY,H g

r FIGURE 3.7.4 POWER SPECTRAL DENSITY, STO4 4

m>

3-59

Yl

..n.-~.

~-

approximately 2.4 percent of the predicted stress of 10,000 psi and 1.7 e

percent of the acceptance criteria of 13,500 psi.

b The only discemable frequency peak on the suncillance holder tube strain gages is at 17.5 !!: and on the shroud tube at 37.5 H:. All gages show random response between 0 and 50 H:. The frequency peaks may be

[

caused by the themal shield, since it is vibrating at both 17.5 II: and

~

37.5 H:.

.l.

3.8 TAPE RECCRLER NOISE 3.8.1 Types of Recorded Noise Noise seen in the playback was caused by one of three sources:

F' 1.

Sensors and conditioning equipment.

~'

2.

External noise sources; i.e., pumps, cooling fans, or other electrical equipment.

I

~l 3.

Electrical noise in the repmduce and record electronics.

j F

In looking at the recorded sensor signals with no pumps operating, the

_[

magnitude of both one and three together were checked.

In general, the results of the second noise source cannot be analyzed.

'}

n a l

3.8.2 Analysis of Tace Recorder Noise

• 7 e

F.1, In order to determine the noise caused by the tape recorders only, the inputs to the recorders were shorted and a tape was recorded. The playback i

gives the noise of the record and playback recorders.

In general, the only significant noise was below 10 H:.

Figure 3.8.1 shows the frequency content from recorder 2, channel 1.

This is similar to most other channels.

T.. 3 ras value was 2.7 my and the major frequencies were 3, 4, and 8 H:. The

~~ }

frequency content of the noise is occasionally a prob 1cm when analyzing the

{'"

various sensors, especially when double integration is involved. However,

M the total noise level is within specifications of the tape recorders.

-4 3-60 m

i n..

f 1

i i

ATTACIDfENT C SURVEILLANCE SPECIMEN HOLDER TUBE VIBRATION DATA

'I

'l i

j i

i I

I L

i

)

t l

l l

7 Jb HST SECTION >500 Hz hV 2 ~"

- 2ND SECTION, >1000 Hz m

m s_I D

s 3RD SECTnON >1000 Hz e

3 o

N 4TH SECTION >1000 Hz 5

s:::,

n s

A l

f ;1) 1_

STH SECTION >1000 Hz r.-'"

i

+

i dv

,% 2 1

4r i

r

/

TANGENTIAL 125 Hz -

e l

F.* RADIAL 125 Hz

,,g

~

o o

o C

- r

, !==:14 -

sY h

\\

M M-1 j

RADIAL 115 Hz

\\

c 1

1 y

/

"l e

o, L

- E

- g s

._g 9

+

z 4

IN-AIR NATURAL FREQUENCIES

.. FNURE 1

s.

TABLE I Plant Operating Component Condition Freauency-H:

E!S Stress-psi SSHT 4 RCP, 75

<500 530*F 420 Shroud Tube Same (See Note 2)

Very Low (Note 1) i NOTES:

(1)

PSD's for 0 and 4 RCP are essentially the same indicating I

stresses are extremely low i

(2)

PSD's indicate broadband response to the random pressure field I

I INSTRUMLNTATION SENSOR IDENTIFICATION LOCATION Ato ORIENTATION Ref Dwg 34502F 4

3r. cried

~

ST02 ST01 Y=,

Top of RV Y,

e

/

C"-

c

)

STRAIN GAGE AXIS PARALL ^m TO T//////////////////D TUBE AXIS

~

Y=J ST04 STO3 g7 Y

h Y-Y SHROUD TUBE STRAIN GAGES STOl ST02 ST03 STO4 Fi so ne_

2__

e 1

O

- =...

INS 11(UMI'!11 Al 10!l :,1 tt,UR Illi~.NTIFICATICt1

,e L O C A T 1 CI1 At:!: OHil.NI A110tl Ref Ong 34533F

~~X, Top of RV f

,_- f ~~

b

=

_/

)

X l

SHT4 SHTl STRAIN GAGE AXIS

\\

PARALLEL TO TUBE AXIS

\\

JSHT3 SHT2 X-X

' SURVEILLANCE SPECIMEN HOLCER TUBE

~

STRAIN GAGES SHTl s

SHT2 SHT3 SHT4 h

g

i D

?

i i

1 t

1

)

.I i

I I

j ATTACHMENT D 4

i OCONEE 1 REACTOR VESSEL SURVEILLANCE SPECIMEN HOLDER TUBE REPORT i

)

I i

i I

o i

f f

1 A

i l

i i

i i

--,-----v---r-m--...---y

,,,mw....,,,.

..-.-,m.,,

m,.--.rm

.-.--r-.we--r-

OCONEE 1 SURVEILLANCE SPECIMEN HOLDER TUBE REPORT t

i INTRODUCTION This report provides additional information relative to the reactor vessel surveillance specimen holder tubes (SSHT) in Oconee 1.

As a result of this additional information, further corrective action beyond that reported in Reference 1 is deemed necessary.

NEW INFORMATION Duke Power Company was appraised by the Babcock & Wilcox (B&W) Company on April 5, 1976 that inspections conducted at another B&W reactor, following removal of the reactor internals, revealed evidence of wear to the SSHT journal bearing located at the bottom of the shroud tube.

CORRECTIVE ACTION Reference 1 describes the evaluation and corrective action taken on Oconee 1 as a result of the inspections performed on the SSHT's.

The inspections and evaluation did not, however, consider the new information described above.

B&W advises that if excessive wear is present on the journal bearings, this has the potential for allowing excessive vibratory motion of the SSHT. Methods are available to compensate for this wear; however, they were not used on the Oconee 1 SSHT's.

In order to minimize the possibility of unacceptable vibratory motion and resultant SSHT degradation occurring during cycle 3, the SSHT's and spring loaded retaining devices will be removed for the balance of cycle 3 operation.

SAFETY EVALUATION Reactor Vessel The request for an exemption to the requirements of 10 CFR 50, Appendix H in Reference 1 provides the justification for operation with the reactor vessel surveillance specimens removed. The additional corrective action described above does not change the evaluation or conclusions in Reference 1.

Surveillance Specimen Holder Tube The evaluation in Reference 1 did not consider the new information described i

above. Should excessive wear exist on the journal bearings, the conclusion i

presented that the SSHT's are structurally adequate for extended operation of Oconee 1 could be invalidated. Therefore, operation of Oconee 1 for the brief period of time until implementation of the corrective action described above has been evaluated. Due to the brevity of the period of operation, a failure is extremely unlikely to occur; however, failures in the areas of wear have been considered.

Complete severance at the wear locations within the shroud tube would have no immediate effect since these portions are contained by the shroud tube.

Severance at the 4th spacer location could allow this lower portion of the holder tube to oscillate on the hinged mounting brackets (pintles). This motion could be expected to wear the anti-rotation portion of the mounting bracket at the dowel pin.

This wear could allow larger oscillations until eventually the upper portion of the holder tube and spring loaded retaining device could be free and could drop into the annulus between the thermal shield and the reactot vessel wall. Depending on the motion, and condition of the uppar portion of the tube, it may be in one or more sections, severed at the wear locations. These sections, depending on their length, could either wedge in the annulus between the thermal shield and the reactor vessel wall, or for shorter pieces, may wedge in the lower reactor vessel head.

Damage from these loose parts could occur to the reactor vessel clad, incore instrument guide tubes and the lower reactor internals. This damage would not represent an imminent threat to public health and safety, but could require extensive evaluation or repair to assure these structures remain serviceable for the life of the plant. The loose parts monitoring system at conee has proven able to detect parts much smaller than those from the failure t

of a holder tube and would allow an orderly shutdown should a failure occur.

CONCLUSION t

It is concluded that operation of Oconee 1 for the period until implementation of the corrective action described above is acceptable. This operation and the corrective action described above will not be inimical to the health and safety of the public.

REFERENCE 1.

Letter, Mr. William O. Parker, Jr., Duke Power Company, to Mr. Benard C. Rusche, NRC, RE: Oconee Unit 1, March 16, 1976.

-