United States Nuclear Regulatory Commission - Protecting People and the Environment

Information Notice No. 85-95: Leak of Reactor Water to Reactor Building caused by Scram Solenoid Valve Problem

                                                       SSINS No.:  6835
                                                            IN 85-95 

                                UNITED STATES
                           WASHINGTON, D.C. 20555 

                             December 23, 1985 

                                   CAUSED BY SCRAM SOLENOID VALVE PROBLEM 

All boiling water reactor (BWR) facilities holding an operating license (OL)
or a construction permit (CP). 


This information notice is to alert addressees of the potential for leakage 
of reactor water to the reactor building caused by problems associated with 
the scram pilot air solenoid valves. Recipients are expected to review the 
information for applicability to their facilities and consider actions, if 
appropriate, to preclude similar problems occurring at their facilities. 
However, suggestions contained in this information notice do not constitute 
NRC requirements; therefore, no specific action or written response is 

Description of Circumstances: 

On September 19, 1985, Dresden Unit 3 experienced leakage of reactor water 
into the reactor building because of what appears to be a generic problem 
associated with the scram pilot air solenoid valves used in most BWRs. The 
event began when the reactor scrammed from 83% power on high average power 
range monitor (APRM) level caused by a pressure spike resulting from an 
inadvertent closure of the turbine control valves. During the scram 
recovery, a problem with the reactor mode switch prevented resetting channel 
"B" of the reactor protection system (RPS). During this half-reset 
condition, steam entered the reactor building and radiation levels were 
elevated in the first three floors of the reactor building. (A half-reset 
condition is equivalent to a half scram condition immediately following a 
full scram.) 

Although it was not immediately recognized, the leakage was caused by the 
scram outlet valves being open when they should have closed as soon as RPS 
channel "A" was reset. For about 23 minutes, the leaking reactor water was 
flashing to steam and causing contamination of the first three levels of the
reactor building. The leakage path was from the scram outlet valves, to the 
scram discharge volume (SDV) vent and drain valves, and then to the reactor 
building equipment drain tank (RBEDT) and the reactor building atmosphere. 

When RPS channel "A" was reset, air header pressure (to the scram valves) 
increased to about only 38 psig instead of the normal 83 psig. The air 
pressure was sufficient to open the SDV vent and drain valves but was 
insufficient to close the scram outlet valves. 



                                                         IN 85-95 
                                                         December 23, 1985 
                                                         Page 2 of 3 


The current understanding of this event suggests that it could occur at most
other BWRs if the reactor is first scrammed, bringing the air header 
pressure to zero, and then the "A" channel of the RPS is reset and the "B" 
channel is left tripped. In this condition, some of the scram pilot air 
solenoid valves tend to vent air preventing the air header pressure from 
rising to the proper value. With the resulting low pressure, the scram 
outlet valves will likely remain open providing a path for reactor water 
leakage to the reactor building. 

Although the mode switch was part of the initiating sequence for the event 
at Dresden Unit 3, other possible causes of half scrams, such as misapplied 
maintenance or surveillances, could start this type of event provided that a 
full scram occurred first. The problem with the mode switch is believed to 
be that it was inadvertently left between positions. 

Testing in place and on the workbench shows that the leakage of the scram 
pilot air solenoid valves only occurs when the "A" channel valve (number 
305-117) is reset and the "B" channel valve (number 305-118) is tripped. 
(See the attached drawing of the scram pilot air solenoid valves.) The 
problem does not occur in the reverse situation when the "B" channel is 
reset and the "A" channel is tripped. The difference in the low pressure 
operating characteristics of the 117 and 118 valves is believed to be 
associated with back pressure on the exhaust diaphragm "E2" induced by 
diaphragm "E1" which causes diaphragm "E2" to leak under the described 

No mechanism has been identified associated with scram pilot air solenoid 
valve leakage that would adversely affect the ability of any plant to scram.
The problem has only been found to apply to reactors that have two separate 
ASCO scram pilot air solenoid valves as shown in the attached drawing. The 
problem has not been associated with the single combined function ASCO scram
pilot air solenoid valves, commonly referred to as "T" ASCO valves. 

Operators may be misled as they were during this event, when there were 
false indications that the reactor water cleanup system relief valve was 
causing the leakage to the RBEDT. If trained, operators should be able to 
recognize the real condition. The full core display contains a blue light 
for each control rod indicating that the scram outlet valve is open, and the 
SDV vent and drain valves also have position indication in the control room.

In response to this event, the licensee has adopted corrective actions that 
deal with the mode switch and with the SDV vent and drain valves. In regard 
to the mode switch, the corrective actions include visual inspection for any
impairment. Another mode switch related corrective action includes revision 
of the scram procedure so that the reactor operator will be directed to move
the mode switch to the "shutdown" position after any scram occurs. This will
,prevent any future mis-positions of the mode switch. If the mode switch is 
replaced in the future with a more reliable type, this instruction may be 
removed. (The reactor operator had been placing the mode switch in "refuel" 
position in order to obtain a white light single rod permissive that 


                                                         IN 85-95 
                                                         December 23, 1985 
                                                         Page 3 of 3 

all rods were in following a scram. Because some rods usually bounce out to 
position 02 following most scrams, the white light permissive has not been 
that useful.) 

In regard to the SDV vent and drain valves, corrective actions include two 
changes to the scram procedure. One change is that the reactor operator will
be directed to close the SDV vent and drain valves using the individual 
control switches in the control room before resetting any scram. This will 
prevent any possible steam releases in the future if the scram air header 
pressure were to become degraded. 

The other change is that a caution statement will be added to the procedure.
If the SDV vent and drain valves will not close during any half scram condi-
tion, following a full reactor scram reset, the reactor operator will be 
instructed to manually scram the reset RPS channel. (A full scram will cause
the SDV vent and drain valves to close.) 

No specific action or written response is required by this information 
notice. If you have 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 Contact:  Eric Weiss, IE 
                    (301) 492-9005 

1.   Figure of Scram Pilot Valve Configuration 
2.   List of Recently Issued IE Information Notices 
Page Last Reviewed/Updated Friday, May 22, 2015