United States Nuclear Regulatory Commission - Protecting People and the Environment

Information Notice No. 86-101: Loss of Decay Heat Removal due to Loss of Fluid Levels in Reactor Coolant System

                                                            SSINS No.:  6835
                                                            IN 86-101 

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

                              December 12, 1986

Information Notice No. 86-101:  LOSS OF DECAY HEAT REMOVAL DUE TO LOSS OF
                                   FLUID LEVELS IN REACTOR COOLANT SYSTEM 
Addressees: 

All holders of an operating license or a construction permit for 
pressurized-water reactor (PWR) facilities. 

Purpose: 

This notice is intended to advise licensees of continuing problems during 
PWR outages with procedures and instrumentation for control of water level 
in reactor vessels when reactor coolant systems (RCSs) are partially drained 
for maintenance. These problems have resulted in temporary loss of decay 
heat removal. 

It is expected that recipients will review this information for 
applicability to their reactor facilities and consider actions, if 
appropriate, to preclude occurrence of similar problems. Suggestions 
contained in this notice do not constitute NRC requirements; therefore, no 
specific action or written response is required. 

Description of Circumstances: 

A typical PWR has a decay heat removal system with two redundant trains. 
Generally, both trains take suction from the same RCS hot leg, and the 
connecting piping is attached to either the bottom or a lower quadrant of 
the hot leg. During certain maintenance activities, the water level in the 
reactor vessel must be lowered below the tops of the nozzles which connect 
the hot legs to the reactor vessel. Lowering the level too far can cause 
vortexing in the hot leg at the suction nozzle for the decay heat removal 
system, air entrainment in the water flowing to the operating decay heat 
removal pump, and air binding of the pump. If the other pump is started, it 
too is likely to become air bound. Consequently, all decay heat removal is 
lost until the water level in the reactor vessel and hence in the hot leg 
piping is raised and the decay heat removal pumps are vented and restarted. 

During outages in the last year and half, decay heat removal pumps at 
several PWRs lost suction because of vortexing. Four of these events are 
described in Attachment 1 to this information notice. Deficiencies which 
contributed to the events include: (1) lack of operator knowledge about the 
correlation between water level and pump speed at the onset of vortexing and 
air entrainment 

8612090402 
.

                                                       IN 86-101 
                                                       December 12, 1986 
                                                       Page 2 of 2 

(2)  operating procedures that did not adequately consider vortexing and air
     entrainment 

(3)  reactor vessel water level instrumentation which was erratic or inac-
     curate, did not have adequate range, was not checked adequately before 
     use, or was not monitored as frequently as necessary during use 

During one of these events, local boiling of reactor coolant and some 
release of radioactive contamination to containment did occur. 

Discussion: 

In the aggregate, licensees involved in the events described in Attachment 1 
have taken certain actions. These actions include additional operator 
training, improvement of instrumentation for monitoring water level in the 
reactor when the level has been lowered for maintenance, addressing in 
operating procedures the relationship between water level and flow rate for 
the onset of vortexing and air entrainment, and requiring in operating 
procedures that the performance of water level instrumentation be checked 
before water level is lowered. 

The nuclear industry has been previously made aware of this problem. IE In-
formation Notice 81-09 described an event that occurred at Beaver Valley 
Unit in March 1981. Further, the Nuclear Safety Analysis Center operated by 
the Electric Power Research Institute published NSAC-52 in January 1983. 
This report provides information on 12 PWR events which occurred from 1977 
through 1981 and which resulted in the loss of capability to remove decay 
heat because of reduction of water inventory in the RCS. Case Study Report 
AEOD/C503 issued in December 1985 by NRC's Office of Analysis and Evaluation 
of Operating Data presents similar information from 1976 through 1984. That 
case study indicates that there were 32 events during that period including 
6 in 1984. Although these reports are available to the industry, significant 
events continue to occur. 

This notice requires no specific action or written response. If you have any
questions regarding this matter, please contact the Regional Administrator 
of the appropriate regional office or this office. 


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

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

Attachments: 
1.   Loss of RHR events at PWRs 
2.   List of Recently Issued IE Information Notices 
.

                                                       Attachment 1 
                                                       IN 86-101 
                                                       December 12, 1986 
                                                       Page 1 of 4 

                         LOSS OF RHR EVENTS AT PWRs

San Onofre 2 

On March 19, 1986, San Onofre 2, a Combustion Engineering designed reactor, 
was in cold shutdown and preparations were made to partially drain the RCS 
and perform maintenance in a steam generator channel head. Before initial 
draining the reactor vessel to the midpoint of the RCS hot and cold legs, 
wide- and narrow-range RCS level instruments were put in service by 
installing their temporary connections and calibrating them. Because their 
readings oscillated when a portable RCS eductor for control of airborne 
radioactive contamination was operated tygon tubing was installed 
temporarily to provide a sight gauge for monitoring a er level. Thus, three 
devices were available for monitoring water level in the system. 

On March 26, the water level in the reactor vessel was below the vessel 
flange, the RCS was vented to the containment atmosphere via incore detecter
nozzles in the vessel head, a low-pressure safety injection (LPSI) pump was 
running to provide decay heat removal via the shutdown cooling system 
(SDCS), and a temporary dam was installed in the cold leg nozzle of the 
steam generator to facilitate maintenance which was to be performed on it. 
To permit repair of,the nozzle dam which had been leaking, the water level 
in the reactor vessel was being lowered to 17.5 inches above the bottom of 
the 42-inch diameter hot legs. 

One of the hot legs supplies water to the inlet side of the SDCS. The nozzle
for the connecting pipe to the SDCS is located on the bottom of that hot 
leg. While the water level was being lowered, a vortex formed on the suction 
side of LPSI Pump 16. The vortex entrained air causing the pump to become 
air bound, loss of SDCS flow, and thus loss of decay heat removal. To avoid 
damage to the pump, it was secured. The redundant pump, LPSI Pump 15, was 
started, and it too became air bound and was secured. To again establish 
flow through the SDCS, the system was vented, and the water level in the 
reactor vessel was raised. Seventy minutes after the first indication of 
vortexing, decay heat removal was #gain established when LPSI Pump 16 was 
returned to service. During the time that decay heat removal was lost, the 
hot leg temperature increased from 114 F to 210 F, and local boiling 
occurred in the reactor core. Steam and 2 curies of radionuclides were 
released to containment. 

The wide- and narrow-range level instruments are connected to taps on the 
RCS hot leg drain line and on the pressurizer. Instrument zero for the 
narrow-range instrument is at the level of the bottom of the hot leg, and 
its range is from zero to +42 inches, i.e., the top of the hot leg. 
Instrument zero for the wide-range instrument is at the reactor vessel 
flange, and its range is from -120 inches (or 19.5 inches below the bottom 
of the inside surface of the hot leg) to +300 inches. The operators distrust 
these two instruments because their readings oscillate when the RCS eductor 
is operating and because low points in flexible tubing at the upper pressure 
tap collect condensate. 

.

                                                       Attachment 1 
                                                       IN 86-101 
                                                       December 12, 1986 
                                                       Page 2 of 4 

The RCS eductor is a portable device which is temporarily installed by 
maintenance personnel when the RCS is opened for repair work. The eductor 
takes suction on the air space above the reactor coolant surface and 
discharges to the containment purge system. Its function is to minimize the 
exposure of maintenance personnel to airborne radioactive contamination. 

While installing and filling the tygon tubing, an air bubble was 
inadvertently trapped in the tubing causing it to read high by 10.5 inches. 
Further, the reference scale for the tubing was displaced by 2.5 inches in 
the upward direction causing a total error of 13 inches on the high side. 
The operators were relying on this device while reactor water level was 
being lowered. The licensee intended to lower the level to 17.5 inches above 
the SDCS nozzle; however, the level was actually being lowered to 4.5 inches 
above the SDCS nozzle. After the level reached 9.5 inches, vortexing 
started. Although, the operator did not have confidence in the narrow range 
instrument, its reading was approximately correct at that time. 

The operator did not have at hand a formal correlation of the potential for 
vortexing as a function of water level and SDCS flow rate. Lack of knowledge
about the performance of the system at low water levels and unreliable 
instrumentation for monitoring water level were the principal causes of this 
event. 

Zion 2 

On December 10, 1985, Zion 2, a Westinghouse designed reactor, was in cold 
shutdown with the water level in the reactor vessel below the flange, the 
RCS vented to atmosphere, a residual heat removal (RHR) pump running to 
provide decay heat removal, and a charging pump running to provide makeup to 
the RCS. The water level in the reactor vessel had been lowered to 
facilitate repair of an RHR valve. A recorder in the control room.was 
connected to the refueling water level transmitter and was being used to 
monitor the water level in the reactor vessel. 

Between December 10 and 14, enough additional water was inadvertently 
removed or lost from the RCS to lower the water level in the vessel far 
enough to cause vortexing and air binding of RHR Pump B. Pump B was 
immediately secured. The redundant RHR pump was started, and it too became 
air bound and was secured. Because of anomalous performance of the 
refueling, water level instrumentation, an operator entered containment to 
read the tygon standpipe that had been installed temporarily to monitor 
water level in the reactor vessel. The licensee concluded that suction to 
the RHR pumps had been lost and started to raise the water level in the 
reactor vessel. After level had increased 10 inches, an RHR pump was 
restarted, but had to be secured because it still had inadequate suction 
pressure. To provide pressure quickly and to increase level further, RHR 
suction was transferred from the RCS to the refueling water storage tank. 
The water level in the reactor vessel was raised an additional 2-1/2 feet. 
Approximately 75 minutes after loss of decay heat removal, RHR Pump B was 
vented and successfully returned to service. RHR Pump A was vented, 
demonstrated to be operable, and deenergized. The reactor had been shut down 
for approximately 100 days, and the increase in RCS temperature was 
15F. 

.

                                                       Attachment 1 
                                                       IN 86-101 
                                                       December 12, 1986 
                                                       Page 3 of 4 

For Zion 2, the suction lines to the RHR pumps connect to a horizontal run 
of RCS hot leg piping. The nozzles for the suction lines are located on the 
underside of the hot leg piping and at a 45 angle from the bottom of 
the line. The internal diameter of the hot legs is 29 inches, and the 
internal diameter of the suction lines is 11 inches. When reactor water 
level falls approximately 5.5 inches below the centerline of the hot leg, 
uncovering of the RHR nozzle commences, and when the water level falls below 
approximately 13.5 inches below the centerline, the RHR nozzle is completely 
uncovered. During the event of December 14, 1985, vortexing started with 
water level at 6 inches above the centerline with RHR flow at 3000 gpm. 

A 10-inch line returns water from the RHR system to the RCS and is connected
to the top of one of the RCS cold legs. The water level sensing line for the
refueling water level transmitter is connected to a 4 inch line which is 
connected to the same cold leg. Both nozzles are in the same vertical plane. 
The 4-inch nozzle is located at 90 with respect to the 10-inch nozzle. 
When the cold leg is partially filled as it was during this event, water 
from the RHR return line impinges with appreciable force on the water 
surface close to the nozzle for the 4-inch line. Because of possible dynamic 
effects of this impingement, the operators believe that water level readings 
from the refueling water level transmitter are inaccurate and erratic when 
the water level in the reactor vessel is low. Furthermore, when the water 
level in the reactor vessel is anywhere below the nominal midpoint of the 
cold leg, the refueling water level instrument will indicate erroneously 
that the water level is at the midpoint. 

Notwithstanding these problems with the refueling water level 
instrumentation, the tygon standpipe was not being continuously monitored 
while the water level was low. Further, the operator did not know the 
correlation of RHR flow rate and the water level for the onset of vortexing 
at the suction of the RHR pumps. 

Sequoyah 1 

On October 9, 1985, Sequoyah 1, a Westinghouse designed reactor, was in cold
shutdown with the water level in the reactor vessel 4 inches below the 
centers of the hot leg nozzles, RHR Train B in service for removal of decay 
heat, and normal letdown and makeup out of service. The water level in the 
reactor vessel had been lowered to facilitate plugging and eddy current 
testing of tubes in a steam generator. During an evolution to put Train A in 
service, RHR Pump A was started and then Pump B was secured. Running both 
pumps simultaneously with low reactor vessel water level caused initiation 
of vortexing and air binding in Pump A. The pump was secured immediately, 
Pump B was restarted, and it operated normally. The alignment of Train A was 
verified and the pump was vented . Pump B was secured, and Pump A was 
restarted, became air bound, and was again secured. Pump B was restarted, 
but this time it became air bound and was secured immediately. After 
verifying that personnel were out of the steam generator, the water level in 
the vessel was raised to the centerline of the hot legs by adding water to 
the RCS from the RWST. Approximately 43 minutes after loss of decay heat 
removal, Pump A was vented and returned to service. Pump B was vented, 
demonstrated to be operable, and deenergized. 

.

                                                       Attachment 1 
                                                       IN 86-101 
                                                       December 12, 1986 
                                                       Page 4 of 4 

At Sequoyah 1, both RHR pumps take suction from the same hot leg (as they do
at San Onofre 2, Zion 2, and Catawba 1). The water level in the hot leg was 
such that initially it would support operation of one RHR pump, but not both
pumps. Starting the second pump without first securing the operating pump 
caused vortexing, air entrainment, and air binding of Pump A, which is 
apparently more sensitive to this problem than Pump B. The procedure for 
operating the RHR system with low water level in the reactor vessel did not 
adequately reflect the relationship between RHR flow rate and water level 
for the onset of vortexing in the suction line for the RHR pumps. 

Catawba 1 

On April 22, 1985, Catawba 1, a Westinghouse designed reactor, was in cold 
shutdown with RHR Train A inoperable because of maintenance, and RHR Train B 
in service to remove decay heat. Although one RHR train was inoperable, the 
licensee started to lower the water level in the reactor vessel to 
facilitate RCS pump seal maintenance. While draining was in progress, 
erratic performance of RHR Pump B indicated that vortexing, air entrainment, 
and air binding were occurring. The pump was secured; a charging pump was 
aligned to take suction from the RWST; and the water level in the reactor 
vessel was raised. Approximately 81 minutes after the first indication of 
vortexing, RHR Pump B was vented and returned to service. Temperature of the 
RCS peaked at 177 F. 

For Catawba 1, the operating procedure for lowering water level in the 
reactor vessel does limit RHR flow as a function of level, apparently to 
preclude the onset of vortexing. However, the licensee believes that water 
level information obtained from inaccurate instrumentation contributed to 
complete loss of RHR flow. Further, the licensee incurred an increased risk 
of loss of RHR flow by lowering water level with one train of RHR cooling 
out of service. With the reactor in cold shutdown and the vessel partially 
drained, a limiting condition for operation in the Technical Specifications, 
for Catawba 1 requires that one RHR train be operating and that the other be 
operable. Nevertheless, the operators concluded incorrectly that water level 
could be lowered if corrective action had been initiated to comply with the 
action statement for that limiting condition for operation. 

Page Last Reviewed/Updated Tuesday, November 12, 2013