ML13330B511
| ML13330B511 | |
| Person / Time | |
|---|---|
| Site: | San Onofre  |
| Issue date: | 03/23/1989 |
| From: | Nandy F SOUTHERN CALIFORNIA EDISON CO. |
| To: | NRC OFFICE OF INFORMATION RESOURCES MANAGEMENT (IRM) |
| References | |
| NUDOCS 8903310173 | |
| Download: ML13330B511 (12) | |
Text
Southern California Edison Company P. 0. BOX 800 2244 WALNUT GROVE AVENUE ROSEMEAD. CALIFORNIA 91770 F
R. NANDY TELEPHONE MANAGER OF NUCLEAR LICENSING March 23, 1989 (818) 302-1896 U. S. Nuclear Regulatory Commission Attention: Document Control Desk Washington, D.C. 20555 Gentlemen:
Subject:
Docket No. 50-206 Responses to Thermal Shield Questions San Onofre Nuclear Generating Station Unit 1 By letter dated March 21, 1989, SCE provided information on the SONGS 1 thermal shield which was requested in NRC letter dated March 3, 1989, and during discussions held during a meeting at NRC headquarters on March 14, 1989. Subsequent to that meeting, telephone discussions between SCE and the NRC were completed which resulted in four additional questions concerning the thermal shield. Responses to these questions are provided as Enclosure 1.
If you have any questions regarding this subject, please let me know.
Very truly yours, cc: 3. B. Martin, Regional Administrator, NRC Region V F. R. Huey, NRC Senior Resident Inspector, San Onofre Units 1, 2 and 3 C. M. Trammell, Nuclear Regulatory Commission 8903310173 890 PDR ADOCK O0 020U PNtJ P
ENCLOSURE 1 Responses To NRC Questions On Thermal Shield
Ouestion 1. Ref. for 6-50 of 2/17/1989 letter The paragraph "Seismic" should explain in more detail the loads and stresses stated therein. The explanation should be accompanied with sketches and additional calculations as needed to support the statements and conclusions contained in this paragraph. Also justify the allowable stress of 3.6 Sm limit for the faulted condition.
Response: A sketch of the flexure and its connection with the thermal shield and core barrel is provided in Figure 1. The thermal shield flexure was modeled as a horizontal beam located between the core barrel and the top of the thermal shield. Originally there were six flexures located around the circumference of the thermal shield with each flexure representing a beam having a length of 5 inches in the radial direction, a width of 8 inches in the tangential direction and a thickness of 0.375 inches vertically. This orientation provides for good load carrying capabilities in the radial and tangential directions, but still remains relatively flexible in the vertical direction allowing for relative thermal growth between the thermal shield and the core barrel.
An overall view of the core barrel, thermal shield, and the locations of the flexures at the top of the thermal shield are shown in Figure 3. A detailed view of the flexure and its attachment to the barrel and the shield is shown in Figure 1.
The properties of a thermal shield flexure, when modeled as a beam are:
The cross sectional area of each flexure is (width times thickness = 8*0.375) = 3.0 square inches (See Figure 2).
The moment of inertia about the vertical axis (thickness times width cubed divided by 12 = 0.375*83/12) = 16 in4.
The loads for the flexure evaluation are referenced from page 7-26 of WCAP-12148. Based on use of the previously indicated beam properties, the indicated radial seismic load of 40,940 pounds produces a direct stress of 13,650 psi.
Similarly, the referenced tangential seismic load of 34,400 pounds produces an average shear stress of 11,470 psi.
In addition, the tangential load also produces a bending stress of 21,500 psi, if the flexure is treated as a clamped guided beam. The clamped guided beam assumption is justified since by observation, the core barrel and the thermal shield are much larger than the beam connecting them. This leads to the observation that the moment reaction at either end of the flexure beam would be significant (greater than zero), and hence a clamped guided beam model.
The stresses combine to give an average membrane stress intensity of 26,700 psi and a membrane plus bending stress intensity of 35,150 psi at the extreme fiber. The stress intensity is determined from the square root of the sum of the direct stress squared plus four times the shear stress squared (Sd2+4*t2).5.
The faulted allowable for membrane stress intensity is 2.4 Sm which is 39,360 psi and for membrane plus bending stress intensity is 3.6 Sm which is 59,040 psi.
By observation the above stresses clearly meet these limits.
The basis for the stress intensity limits is explained in the following paragraphs.
The justification of the 3.6 Sm stress intensity limit is based on the rules of the ASME Boiler and Pressure Vessel code,Section III, subsection NG. These limits were developed for the design of the critical core support structures.
Although the thermal shield and the flexure are not core support structures, these limits were conservatively applied to the evaluation of the flexure since no specific code criteria exist for these structures. Subsection NG of the ASME code references Appendix F which states (Paragraph F-1331.1 (b)) that for component elastic analysis, the stress limit for the level D situation, (the faulted conditions) for primary membrane plus primary bending is 150% of the limits for general primary membrane stress intensity. Further, Paragraph F-1331.1 (a) states that the general primary membrane stress intensity shall not exceed the lessor of 2.4 Sm or.7 Su. Applying the 150% value to 2.4 Sm yields 3.6 Sm, the value specified above. Applying 150% to the.7 Su yields a value greater than ultimate. Now for this evaluation, Sm = 16,400 psi @ 600*F and the code value of Su is 70-75 ksi.
Therefore, use of the 3.6 Sm limit is justified by the application of the ASME code rules.
Question 2. It appears that the load combination, the seismic and other hydrodynamic loads were not investigated using their time history to ascertain that their peak values do not coincide. The SRSS method of combining the loads is acceptable only if the combination of the maxima of the loads is precluded.
Response: The Flow Induced Vibration (FIV) loads are generated due to flow turbulence and thus subject the reactor components to random vibration loadings. The vibration analysis which has been performed determines the root mean square (RMS) values of these loadings. Since the mean value of the vibratory loading is zero, the RMS value is equal to the standard deviation (sigma) of the distribution of the vibratory loads.
For conservatism, the analysis was performed using the 4 sigma vibratory loads.
Thus peak vibratory loads used in the analysis correspond to 4 times the RMS values.
The seismic loads also subject the reactor components to random vibration loadings. A response spectrum seismic analysis was performed to determine the peak (rather than RMS) seismic loads in each direction and the total seismic load was obtained as the square root sum of the squares (SRSS) of these peak loads in accordance with standard practice for combining seismic loads.
Component stresses resulting from the peak vibratory loads were combined via SRSS with stresses resulting from the peak seismic loads and compared with established stress limits, as shown in WCAP 12148, in order to demonstrate acceptability. The justification for combining the vibratory and seismic responses via SRSS rather than direct addition follows:
o The seismic responses and the FIV responses are generated by two random vibration forcing functions.
Each of these has a zero mean value so that the standard deviation of the loads corresponds to the RMS load.
The analysis was performed using the 4 sigma FIV loads and the peak seismic loads. A breakdown of the sigma values and the percentage of occurrences for a normal distribution follows:
Expected % of Time Interval Sigma Occurrences Between Occurrences 0+ to 1 68.26%
.24 sec 1+ to 1.5 18.38%
.9 sec 1.5+ to 2.0 8.80%
1.9 sec 2.0+ to 2.5 3.32%
5.02 sec 2.5+ to 3.0
.98%
17 sec 3.0+ to 3.5
.22%
75.8 sec = 1.26 min.
3.5+ to 4.0
.04%
417 sec = 6.94 min.
Also, given in the above table is the time interval between each occurrence of loads which are within a specific range of sigma values. This time interval is based on a frequency of 6 Hz which conservatively bounds the frequencies of the dominant modes of vibration for the worst credible degraded condition.
The time interval is calculated from the following equation:
Time Interval =
1 (6 Hz) (Frequency of occurrence) o Using the preceding table; for a normally distributed random variable, values in excess of 2 sigma only occur 4.56% of the time and values in excess of 3 sigma only occur 0.26% of the time. It is therefore very conservative to even consider 4 sigma FIV loads.
o A review of the design Acceleration Time History* for the SONGS 1 modified Hausner earthquake shows that it has a duration of only 20 seconds.
Further, the length of time for which the acceleration approaches.67 g's is too small to infer from the curve (1-2 "spikes" of essentially no duration throughout the 20 second time history).
This implies that there is a very small probability that peak acceleration levels would be reached during the postulated seismic event.
o Since the peak seismic loads would occur less frequently than even the 3.5 sigma seismic loads, it is conservative to assume that the probability of experiencing peak seismic loads is.04% or.0004.
o Based on the preceeding results, the probability that peak seismic loads could occur simultaneously with FIV loads in excess of 2 sigma is
.0456 X.0004 = 1.82 x 10 **(-5).
The analysis was performed by combining 4 sigma vibratory loads with peak seismic loads and the probability that these loads occur simultaneously is less than.0004*.0004=l.6xlO**(-7) since FIV loads in excess of 3.5 sigma would occur less than.4% of the time. Therefore, the probability of 4 sigma vibratory loads occurring simultaneously with peak seismic loads is negligibly small so combination of the loads by SRSS is acceptable.
Letter K. P. Baskin to D. M. Crutchfield, "SEP Topic 111-6, Seismic Design Consideration," July 9, 1982.
0 o Since the probability that 2 sigma vibratory loads would occur simultaneously with peak seismic loads was determined to be 1.82 x 10**(-5), an evaluation was performed to determine if direct combination of these loads would be expected to alter the conclusions of this study.
That evaluation follows:
(Ref. WCAP-12148)
Vertical DBE Seismic Load = 53,640 lb.
worst credible degraded case (Pg. 7-26)
Vertical RMS FIV Load = 19,700 lb.
Case 000, Block 240 (Table 6.2-2, Pg. 6-25)
Direct Combination For 2 Sigma FIV Loads Load = 53,640 + (2)(19,700) = 93,040 lb.
SRSS Combination For 4 Sigma FIV Loads Load = SQRT (SQ 53,640 + SQ (4 x 19,700))
= 95,324 lb.
This comparison shows that combining the 4 sigma FIV loads by SRSS with the peak seismic loads leads to loads which conservatively bound the direct combination of 2 sigma FIV loads and peak seismic loads. Therefore, the higher probability of occurrence associated with coincidence of the 2 sigma FIV loads and the peak loads is not a concern since it is bounded by the case which has been evaluated.
o As a final consideration, it should be noted that the probability of occurrence of the DBE for SONGS 1 is on the order of 10 **(-5).
Question 3. It appears that the impactive loads resulting from rocking of the thermal shield were not accounted or. These loads should have been combined with the seismic loads obtained from the three-directional load combination to calculate the worst scenario regarding the support blocks.
Response
The rocking motion of the thermal shield, whether caused by flow-induced vibration or a seismic event or both, is limited in magnitude by the limiter keys (see Figures 3 and 4).
The amount of motion permitted by the limiter keys, in turn, depends on how much wear they have experienced during normal operation, due to flow-induced vibration. For the worst credible case, the radial motion of the thermal shield at the limiter keys is calculated to not exceed.480 inch. which means that the radial gap between a limiter key and keyway (see Figure 4) cannot become larger than.480 inch due to wear.
The tangential motion at a given key is limited by the keys which are located 90 degrees away.
In Section 7.5 of WCAP-12148, a model was presented for calculating impact loads at the limiter keys and support blocks which depends on the free vibration amplitude and the actual radial gap. The maximized load from these calculations was 37,562 lbs. for two blocks or 18,781 per block. To address the seismic issue, the same procedure was used to calculate seismic plus flow induced vibration impact loads. This was done by linearly adding the calculated free flow induced vibration amplitude of the thermal shield in the worst credible condition, to the maximum displacement amplitude calculated for the seismic event. The resultant,.218 inch, is the vertical free "lift-off" amplitude at the support blocks located on one side of the thermal shield during rocking that is appropriate for flow induced vibration plus seismic excitation. To provide additional conservatism, a free amplitude of 0.24 inch was used rather than.218 inch. Using this lift-off amplitude, a stress analysis is performed to assess the maximum impact load associated with dropping of the thermal shield from the tilted lift-off condition. The result was a maximized impact load of 75,028 lbs. for two blocks, or an impact load of 37,514 lbs. for one block. Conservatively assuming only one block would react the impact, a load of 53,050 lbs. would result. Both of these cases produced impact loads which are well below the limiting support block load of 78,800 (4 sigma 4 (19,700)) which was previously analyzed as reported in Section 6 of WCAP-12148, for the worst credible condition.
It can therefore be concluded that the combined seismic and flow-induced vibration impact loads at the support blocks are acceptable.
Question 4. In the response to the Overall Concern No. 4, it is stated that "if shield were to fall, its downward motion should be restrained by the lower radial keys."
The structural integrity of the lower radial keys has not been analyzed for the impactive loads resulting from the downward motion of the thermal shield. In that case, provide the basis for the statement that the shield will be stopped in its downward motion.
Response: It should be noted that for the postulated thermal shield drop onto the radial keys, the structural integrity of all impacted components has been evaluated. The methodology and the conclusions of the thermal shield drop analysis were summarized in Section 9.2 of the WCAP-12148 "Engineering evaluation of the SONGS I Thermal Shield Supports," February, 1989.
Results of the structural evaluations of the thermal shield drop on the impacted components, viz, a) radial keys, b) radial key weldments, c) support saddle and support saddle weldments, d) core barrel shell, and e) core barrel flange showed that stresses in these components are within the ASME Code allowables and, therefore, the structural adequacy of the impacted components is maintained for this postulated condition.
The stress summary of the thermal shield drop evaluations is given in the following table.
IMPACTED STRESS MIN. MARGIN OF SAFETY COMPONENT CATEGORY (ALLOWABLE/CALC. -1)
Radial Key Pm (SHEAR) 2.9 Pm + Pb Large (> 5.0)
Key Weldment Pm (Shear) 1.06 Support Saddle and Pm (Shear) 1.04 Saddle Weldment Pm + Pb 0.50 Core Barrel Pm Large ( 5.0)
Barrel Flange Pm + Pb Large (> 5.0) 04230 CHUTE SUPPORT WELD S CORE BARREL SURVEILLANCE SAMPLE CHUTES
-FLEX.UR, SUPPORT CHUTE
-THERMAL SHIELD WELDS THERMAL SHIELD Flexure 5pport and Sorvetitance Sample Chute Deteits FIGURE 1
IT p
-IL ItI
%D NV UP02 f~gI 2:!i
SPECIMEN TUBE SPECIMEN TUBE SPECIMEN BASKET EXPANSION JOINT FLEXURE FIXTURE (TYP 6 PLACES)
LIMITER KEY (TYP 4 PLACES)
THERMAL SHIELD SUPPORT BLOCK (TYP 6 PLACES)
FIGURE 3
-Ill irA ET+ L&
FIGUR 4)