ML20141M235
| ML20141M235 | |
| Person / Time | |
|---|---|
| Site: | Palisades  |
| Issue date: | 03/31/1992 |
| From: | Office of Nuclear Reactor Regulation |
| To: | |
| Shared Package | |
| ML20141M234 | List: |
| References | |
| NUDOCS 9204030215 | |
| Download: ML20141M235 (4) | |
Text
- -_ ._ - ._ . . - - - _ - - . - - - - - _ - - . _ - . . - .-
a stav
+
'3" \o UNITED STATES
! .. E NUCLEAR REGULATORY COMMISSION
- d @tg v[/ ,
WASHINoToN. o C. 20f46
$AFETY EVALVAIJAN BY THE OFFICE OF NUCLEAR REACTOR REGULATION RE0 VEST FOR RELIEF FROM ASME CODE REEAIR--CONTAI.NJENT AIR COOLERS CONSUMERS PQRER COMPANY EAL1}ADES PLANT DOCKET NO. 50-255 1.0 INTRODUCTIOR By letters dated September 9, 1991, January 16, and March 18, 1992, Consumers Power Company (CPCo), the licensee, requested relief from Section XI of the ASME Code. ~he request is for the purpose of making non-code repairs to three (of four) containment cir coolers at the Palisades Nuclear Plant. The air coolers are air handling and cooling units located entirely in the containment building. Plant service water from the critical service water header is circulated through the units' cooling coils. The units are designed to control containrent temperature during normal operation and condense steam during certain a:tident scenarios to help limit peak containment pressure.
The coolers are chssified as Safety Class 3 per Regulatory Guide 1.26.
This request is a follow-on to a previous relief granted by the NRC concerning cooling coil leakage during hydrostatic tests. In that relief, the licensee was required to attain zero leakage during hydrostatic tests of the cooler coils by the end of the 1992 refueling outage. Tha request for additional relief arises from inherent design problems that preclude employing code repair methods.
The air cooler cooling coils are fabricated from 5/8-inch OD (1/2-inch type "K") copper tubing. The individual coil tubes are brazed to a 1-5/8 ',nch OD copper manifold (1-1/2-inch type "M" copper tubing). The manifold to tube connections are configured such that they are difficult to access for repair.
Operating experience has demonstrated that these joints are susceptible to developing pinhole leaks.
These leaks are reported to be the result of several causes. Brazing defects remaining from the time of manufacture were cited in several instances. Some water erosion of the copper and braze material has been observed and some leaks attributed to this mechanism. Finally, some leaks ar2 attributed to hydraulic surge induced by start-up of the service water pumps during emergency diesel generator tests.
Attempts at rebrazing the leaking connections have resulted in adjacent tube-to-m mifold connections becoming heat-affected and creating additional leaks on previously non-leaking connections. As a result, the licensee has deter-mined that due to design, geometry, and materials, the ability to complete a code repair on the three remaining original coolers is impractical. These factors led to the replacement of one cooler (of the four total) during 1990.
Due to the inability to perform a c-de repair, the licensee has proposed a temporary non-code repair to assure zero leakage. The repair is in the form of an engineered clamp with an epoxy sealant. A clamp would-be applied to a 920403021S 920331 5 DR ADOCK 0500
. l leaking joint upon discovery of leakage, regardless of the operational mode of
-the unit. Clamps would be used on an as-needed basis to stop leaks, up to the maximum permitted by seismic support considerations (approximately half the joints per raalfold).
A second joint type in the assembly, a brazed miter forming an ell in the 1-5/8 inch OD copper manifold, has been r.n sccastoral leak site. Although these miter joint leaks have proven to be Code repairabla, a separate clamp has been designed for this geometry and is available if needed.
The proposed clamps would be employed to meet the mandated zero leakage after hydrotest until the three subject coolers could be replaced with an upgraded design. The licensee has proposed the following schedule for replacement of the air cooler coils (as amended in the letter of Jan. 16,1992):
verification of new cooler design during 1993 refueling outage installation of two coolers during 1994 refueling outage
- installation of the one remaining cooler during the 1996 refueling outage ,
This schedule is reported to be controlled by long lead times for engineering and procurement and other conflicting outage wort 2.0 pjSCUSSION The NRC strff reviewed CPCo experiences and bases for requesting relief from the ASME Code,Section XI requirements fe* repairs to the air cooler coils.
The staf f has determined that because of the design of the tube to manifold connections, the materials of construction, and the limited access, the coil joints in question cannot be reliably repaired in accordance with the code.
The staff has determined that the burden involved in requiring a code repair could result in adjacent connections becoming heat affected, thereby creating potent 1:1 new leak locations.
It is determined that coil leaks do not cause or result in additional failures to other systems. Furthermore, the licensee has performed analyses of post loss-of-coolant accident (LOCA) conditions assuming impaired or completely non-operational containment air coolers (CACs). These analyses indicate adequate cooling margins oxist with only the containment sprays in operation.
Thus, the licensee concluded that the CACs are not needed to mitigate LOCA conditions. A main steam line break (MSLB) accident analysis was also performed to explore containment conditions with assumed impairment of the CACs due to leakage. As shown in the Final Safety Analysis Report (FSAR), the containment pressure and temperature are held below the design limits with only three cf the four CACs operational and with one containment spray pump.
A sensitivity study was done to determine the impact of degraded air cooler performance in this scenario. The MSLB analysis yielded acceptable results .
with a 40'/. reduction in CAC heat removal capacity due to leakage. This analysis also ass 9mes only three of the four CACs are functional.
CPCo has completed an extensive analysis of the proposed clamp designs, subjected prototype clamps to field and laboratory tests and provided two l samples for NRC staff examination. The staff finds the licensee's effort
! toward the clamp design engineering to be appropriate. The two clamp samples l
4 submitted for staff evaluation (one each, for the tube to manifold j(int and the manifold miter joint) are amply sized and constructed to perform in the intended manner.
The staff has also examined the cooling coil joint designs and finds some concerns relative to joint strength. These concerns arise due to the fact that socket type brare joints were not used for the tube to manifold connections or in place of the miter joints on the manifolds and headers.
These joint designs depart from typical brare joint design practice (reference: Welding Handbook, Chapter 11, Vol. 2 Seventh Ed., American WeldingSociety,1976),andthecurrentASMESectIon!!!, Para.ND-3671.6 requirements.
In the designs in questian, the joint area wetted by the brare material is quita small cohpared to a socket type braze joint. This reduces the joint strength because the area of controlled joint clearance (gap) is reduced significantly. The strength of a brace (or solder) joint depends entirely upon the establishment and maintenance of a thin (0.003" to 0.010") braze material layer between the two parts joined. properly executed braze joints can equal or exceed the base material strength. Instead, the staff noted that a braze metal fillet was employed (for the tube to header joints) in an attempt to compensate for the small joint area. This type of design cannot ,
achieve more taan a fraction of the strength of a properly designed braze co nection, regardless of fillet size. Illustration of the inherent weakness was demonstrated by the licensee's tests performed during the qualification of a repair brazing procedure in 1986 (this repair method was not practical due to aforementioned difficulties). Rather than the 2: .(SI minimum tensile strength required, the strength actually achieved with butt joints was 8 to 18 KSI. Therefore, the staff finds that due to the joint design, the jcint strength is indeterminate. Thus, calculation of joint performance is impractical and assurance of leak free performance under all design conditions was of concern.
Due to the staff concerns regarding joint strength, several related questions were posed to tne licensee regarding the seismic capabilities of the coolers.
The most significant result of this enquiry shows the amplitude of the seismic demand is low. The coolers are on the lower level of the containment building where the peaks of the input response spectra are low with respect to the rest of the plant. Thus, the burden imposed upon the joints is low. Since the cooler configuration is redundant, with 16 tubes per man: fold, and three l
manifolds per cooler, a single joint lask or failure does not compromise overall functionality. As discussed previously, overall functionality has been shown to have minimal impact on accident mitigation due to redundancy of l coolers with other systems.
l l In view of the joint design uncertainties and other operational experiences such as hydraulic losses and flow induced erosion, the licensee is reminded of their responsibility under Code Paragraph IWA-7220. This paragraph states, in part: "If cause of failure appears to be a deficiency in the specification .
for the existing item, the specification for the item to be used for replacement shall reflect appropriate corrective provisions."
L. , . -_. -
1 1 l 3.0 (pNCLU$10NS The stat. finds that pursuant to 10 CFR 50.55a(g)(6)(1) the use of the proposed clamps is appropriate and in the interest of the public health and !
safety, with due regard to the burden upon the licensee that could result if the Code requirements were imposed upon the facility. Application of the submitted clamp designs may proceed, at the licensee's discretion, on any or all of the four CAC's, up to the three percent weight limit imposed by seismic calculations. This relief is granted for the period as described in Section 1.0, Introduction, which describes an air cooler replacement schedule completed by the end of the 1996 refueling outage. Additionally, the licensee is urged to consider application of the miter joint clamps to all miter joints as a precautionary measure, regardless of leakage. Use (or reuse) of the clamps is permitted for the remainder of the operating life of each individual cooler, with intended replacement schedule as proposed.
Additionally, pursuant to 10 CFR 50.55a(g)(6)(i), the licensee shall specify appropriate corrective provisions to ensure that replacement of CAC coils meet the requirements of ASME Code,Section XI, paragraph IWA-7220.
The staff recommends that if braze joints are used in the redesigned coils, they be socket type, designed and tested as described in ASME Section III paragraphs ND-3671.6, ND-4500, ND-4520; Fig. ND-4511-1; ANSI /ASME AG-1-1985, Para. CA-6130, CA-5240; and Addenda ANSI /ASME AG-la-1985, Paragraphs AA-6410, AA-6420, and AA-6430.
Principal Contributor: G. Hornseth, EHCB D3te: Parch 31,1592 i
I I
u
_ _