Information Notice No. 86-106, Supplement 1: Feedwater Line Break

                                                 SSINS No.: 6835 
                                                 IN 86-106, Supplement 1

                                 UNITED STATES
                         NUCLEAR REGULATORY COMMISSION
                     OFFICE OF INSPECTION AND ENFORCEMENT
                             WASHINGTON, DC 20555

                               February 13, 1987

Information Notice No. 86-106, SUPPLEMENT 1: FEEDWATER LINE BREAK

Addressees:

All nuclear power reactor facilities holding an operating license or a 
construction permit.

Purpose:

This supplement to Information Notice (IN) 86-106 is intended to provide 
addressees with additional information about a potentially generic problem 
which resulted in thinning of secondary system piping at both units of an 
operating nuclear power station, catastrophic failure of a main feedwater 
suction pipe, and injuries and fatalities of workers in the vicinity of
the  pipe. Recipients are expected to review this information for
applicability to  their facilities and consider actions, if appropriate,
to preclude the  occurrence of similar problems at their facilities.
However, suggestions  contained in this information notice do not
constitute NRC requirements;  therefore, no specific action or written
response is required.

Discussion:

IN 86-106 was issued on December 16, 1986 in response to a feedwater line 
break at Surry Power Station Unit 2. That notice provided information 
regarding the feedwater line break. NRC regional offices are collecting 
information about the condition of piping in other plants.

In addition to the failure of the suction line to feedwater pump "A," the 
licensee (Virginia Electric and Power Company,) found that the check valve
in  the discharge line for the pump had failed. One of two hinge pins had 
apparently been missing for some time, and the disc/seat assembly was 
dislodged from the valve body. Dislocation of the disc/seat assembly had 
resulted from failure of two clamp assemblies which hold the disc/seat 
assembly in the valve body. Failure of the clamping assemblies appears to
be  the result of erosion/ corrosion. Loss of one hinge pin probably would
not  prevent the valve from performing its function. Dislocation of the
disc/seat  assembly did not contribute to failure of the suction line but
may have  contributed to the volume of water released.


.

                                                 IN 86-106, Supplement 1 
                                                 February 13, 1987  
                                                 Page 2 of 5

On January 14, 1987, the licensee submitted a detailed account of the 
circumstances surrounding the Surry 2 feedwater piping failure of December
9,  1986./*/ On February 10, 1987 the licensee shared results of their 
determinations thru meetings with industry representatives. The licensee's 
conclusion is that thinning of the feedwater pipe wall was caused by 
erosion/corrosion. On January 15, 1987, the NRC staff met with technical 
experts from several engineering disciplines--piping design, metallurgy, 
nondestructive testing, water chemistry, corrosion, and fluid mechanics--
to  participate in a technical panel discussion on the parameters believed
to have  had an important role in the pipe break at Surry 2 and the means
to predict  and reduce the effects of erosion/corrosion in piping systems.

The technical panel concluded that the Surry pipe break failure mechanism
was  erosion/corrosion. Erosion/corrosion is flow-assisted corrosion.
Corrosive  action is initiated by erosion of protective metal oxide. The
role of those  parameters which could have potentially contributed to
erosion/corrosion in  the feedwater piping system are summarized below:

o Piping Design

The configuration of the piping where the break occurred at Surry 2 is 
believed to have played a major role in establishing conditions which
promoted  erosion/corrosion. The pipe failure occurred on the outer radius
of a  90-degree, (long radius), 18-inch diameter elbow which was connected 
downstream of a flow-splitting tee in the 24-inch diameter condensate
header.  The break initiated in the carbon steel elbow material (A234
Grade WPB) -- not  in either weld. The average bulk flow velocity of the
water was calculated to  be 17 ft/sec in the 18-inch branch line. The
elbow-tee configuration has been  identified as an undesirable design
arrangement which is believed to have  caused direct flow impingement on
the inside of the elbow and to have  established secondary flow paths in
the elbow causing even higher, turbulent  flow velocities.

The use of a 45-degree lateral fitting rather than a tee would have
reduced  the direct flow impingement effects and reduced the local flow
turbulence in  the elbow. The piping design stresses in the elbow were low
and believed not  to have contributed to pipe wall degradation.

o Fluid Dynamics

Oxide dissolution is believed to play a key role in the mechanism of
erosion-corrosion wear and is highly interactive with flow velocity. An
increase in  flow velocity generally tends to increase the erosion/
corrosion rate in  carbon steel piping although the effect is more
pronounced in two-phase flow  conditions. Experimental tests have further

__________

/*/ Virginia Electric and Power Company, "Surry Unit 2 Reactor Trip and 
Feedwater Pipe Failure Report," Revision 0, January 14, 1987. __________


.

                                                 IN 86-106, Supplement 1 
                                                 February 13, 1987  
                                                 Page 3 of 5

established that local flow velocities in an elbow can be two-to-three
times  higher than bulk flow velocities.

Pipe wall measurements performed at both Surry Units have shown that
piping  erosion/corrosion effects are most severe at locations where local
flow  velocities are high (e.g., downstream of restricting orifices, flow-
control  valves, reducers, and in elbows and tees).

Based on the system operating temperature and pressure at the time of the
pipe  failure (374xF/367 psi, 50xF subcooled), the conditions for
cavitation to  exist in the elbow was not likely. However, cavitation-
erosion cannot  completely be excluded as a contributory factor under
different operating  modes such as single-pump operation.

The temperature of the feedwater at the Surry 2 carbon steel pipe elbow
that  failed is nominally 374xF. The temperature effects of erosion/
corrosion on  carbon steel are greatest in the 250-340xF range. Below
250xF and above 340xF,  erosion/corrosion wear rates decrease rapidly.

o Piping Material

Carbon steel can be vulnerable to erosion/corrosion when certain
unfavorable  conditions are present. However, by increasing the alloy
content (e.g.,  chromium, molybdenum, copper), its resistance to erosion/
corrosion can be  increased significantly. Field experience has shown that
the use of 2 1/4 Cr-1  Mo steel improves piping resistance to erosion/
corrosion by a factor of four.  Chemical analyses of the failed pipe elbow
from Surry 2 have disclosed  unusually low amounts of these elements,
particularly chromium (less than 0.02  percent).

Austenitic stainless steel has been proven to be highly resistant to 
erosion/corrosion under normally expected flow conditions.

o Water Chemistry

Water chemistry is believed to have been another important factor in
causing  the pipe wall thinning at Surry 2. The erosion/corrosion wear
rate of carbon  steel is greatest when the pH levels are between 7 and 9
or below pH 5.  Erosion/corrosion rates drop sharply at pH levels above
9.2. The Surry 2 pH  levels were reported to have been maintained between
pH 8.8 and 9.2, however,  local values could vary significantly./*/

__________

/*/ It also has been noted that during the initial years of operation, 
ineffective control of water chemistry and condenser in-leakage may have 
contributed to the degraded feedwater piping in Surry 2, especially at 
locations of high flow velocity. Subsequent to 1981, condensate polishing 
units have been used at Surry 2 to remove impurities from the condensate.
__________


.

                                                 IN 86-106, Supplement 1  
                                                 February 13, 1987  
                                                 Page 4 of 5

A preliminary finding by the Surry 2 licensee that extremely low oxygen 
content in its secondary side water contributed to the recent pipe break
has  been questioned. Although it is known that oxygen content of about
100 parts  per billion (ppb) is beneficial for neutral water (pH of 7.0)
because this  improves repassivation of carbon and low alloy steels, many
fossil plants and  foreign nuclear plants have operated at extremely low
oxygen content with no  evidence of piping erosion/corrosion in the
feedwater piping. At Surry 2, the  oxygen content has been maintained at 4
ppb to minimize steam generator tube  degradation.

Predictive measures to detect erosion/corrosion in piping systems also
have  been identified and these measures are summarized below.
Erosion/corrosion  failures in two-phase systems and erosion in single-
phase systems containing  suspended solids were noted to be much more
prevalent than erosion-corrosion  in condensate and feedwater systems.
Consequently, these single phase and  two-phase systems have received more
attention in operation and maintenance.  Many utilities have programs to
control and reduce damage in two-phase systems  such as steam extraction
lines and turbine wet steam piping and in single-phase systems containing
suspended solids such as service water piping. Except  for concerns with
suspended solids, erosion in service water systems would not  likely
result in a catastrophic failure because of the low operating  temperature
and pressure.

Currently, Section XI of the ASME Boiler and Pressure Vessel Code does not 
require any inservice inspections-specifically for measuring pipe wall 
thickness. Although not required to do so, many utilities had elected to 
perform wall thickness measurements routinely in piping systems where 
erosion/corrosion caused by wet steam or raw water impurities had been
found  to result in severe pipe wall degradation. The use of the zero-
degree  ultrasonic beam technique is generally utilized for measuring pipe
wall  thickness. Although Section XI does not require measurement of wall
thickness,  a volumetric examination for indication of flaws is required
in Code Class 1  and 2 piping systems, but examination is limited to the
weld area and heat-affected zone using 45 and 60 degree shear wave
techniques. There are no  inservice volumetric requirements for ASME Code
Class 3 piping, nor any  inservice inspection requirements for ANSI B31.1
piping systems.

The effectiveness of ultrasonic measurements in piping is enhanced by 
predictive methods that can locate potential areas of maximum wear. The 
Department of Mechanical Engineering of MIT (Cambridge, MA) has been 
developing a computer program to predict location and extent of wear in 
two-phase piping systems. The program is reported to also be applicable to 
single-phase piping systems. The NRC staff has not performed any
substantial  review of this computer program.

In other studies, as reported in EPRI Report NP-3944, "Erosion/Corrosion
in  Nuclear Plant Steam Piping: Causes and Inspection Program Guidelines,"
April  1985, the influence of the flow path configuration for various
fittings has  been established and expressed as a range of empirical
values from 0.04 (least  harmful flow occurring in straight pipe) to 1.0
(most harmful flow occurring  in flow-splitting tees) to be used for
calculating erosion/corrosion


.

                                                 IN 86-106, Supplement 1  
                                                 February 13, 1987 
                                                 Page 5 of 5

wear in two-phase (wet steam) piping systems. The Surry-2 elbow-tee 
configuration albeit not in a two-phase system would have been predicted
as an  arrangement potentially susceptible to severe erosion/corrosion
wear based on  the high empirical value of the elbow-tee configuration.

Additional information pertaining to erosion/corrosion in wet steam piping
can  be found in Information Notice No. 82-22, "Failure in Turbine
Exhaust  Lines," dated July 9, 1982. Other erosion/corrosion events
pertaining  specifically to the feedwater systems (including emergency and
auxiliary feed)  have occurred in feed pump minimum flow lines, J-tubes in
steam generator  feedwater rings, and emergency feedwater supply to a
helium circulator.

No specific action or written response is required by this information
notice.  If you have any questions about this matter, please contact the
Regional  Administrator of the appropriate NRC regional office or this
office.



Edward L. Jordan, Director
Division of Emergency Preparedness 
  and Engineering Response 
Office of Inspection and Enforcement

Technical Contacts: Roger W. Woodruff, IE
                    (301) 492-7205

                    Vincent Panciera, Region II
                    (404) 331-5540

                    David Terao, NRR
                    (301) 492-7037

Attachment: List of Recently Issued IE Information Notices