ML20090D618

Ro:On 720228,unit Tripped Offline Due to Malfunction in Continuous Power Supply.Malfunction Caused Feedwater Upset & Reactor Low Water Level Scram.Caused by Transfer of Motor Generator Set Drive from Ac to Dc

ML20090D618
Person / Time
Site:	Nine Mile Point
Issue date:	04/18/1972
From:	Burt P NIAGARA MOHAWK POWER CORP.
To:	Skovholt D US ATOMIC ENERGY COMMISSION (AEC)
Shared Package
ML20090D621	List: ML20090D618
References
NUDOCS 8303020175
Download: ML20090D618 (5)

Text

{{#Wiki_filter:c , NBAGARA MOH AWK power CORPORATION NIAGARA MOHAWK Nine b!ile Point Nuclear Station Unit #1 Post Office Box 32 Lycoming, New York 13093 April 18, 1972

                                                                                                   ~ ' ' -

lir. Donald J. Skovholt Assistant Director for Reactor Operations Division of Reactor Licensing / W- - - ~ ' United States Atomic Energy Commission /h) .. Ej  ; Washington, D. C. 20545 fM u - i

                                                                                           .-        ba          ;

Q.f' ' '..f )6

Dear lir. Skovholt:

                                                 '\fV . PthC'y',,;l i~.4;   /s/

Re: Provisional Operating License DPR-17 Docket No. 50-220 {-{ On February 28, 1972 at 13:32, the Nine t!ile Point Nuc1 car Station, Unit #1 tripped off line as a result of a malfunction in a continuous power supply.. Initial Operating Condition i Steady state operation ll ilNth - 1760 Reactor pressure - 1011 psi

                            !.!We - 609 (gross)             Steam flow         -  6.25 X 106 lbs. per hr.

Introduction . [ Prior to the trip, the Station was operating at a steady stat'e output. A l reactor protection system continuous power supply motor generator set mal-i functioned causing loss of c1cetrical power to half of the reactor protec-i tion system and part of the feedwater control system. The malfunction caused a feedwater upset and a reactor low water level scram. Following , the scram, the water level increased and spilled into the main steam lines.  ; i Sequence of Etents At 13:31:31 low output voltage alarm 162 }!/G set 13:31:43 All sensors in channel 11 reactor protection system tripped. (Fail safe on loss of voltage.) 8303020175 720418 PDR ADOCK 05000220 S pyg

g . p-w s Mr. Donald J. Skovholt April 18, 1972 Sequence of Events (cont'd) 13:31 loss of feedwater control and feedwater valve lockup due to loss of the M/G set. 13:31 Clean Up systen isolated 13:31:45 Reactor scram due to low 1cyc1 trips (approx) 13:32 Main Steam isolation valves closed, reactor pressure less than 850 psi in run mode. 13:34:43 M/G set auto transferred back to AC drive. 13:40 Unsuccessful attempts to restore Clean Up system 13:41:15 M/G set transferred to DC drive again, loss of continuous power supply voltage. 13:43:23 Reactor 1cvel +3 ft. above nininun normal water IcVel. 13:44:32 #111 electromagnetic relief valve open (Note - closing time may not be nonitored after five seconds.) 13:44:32 #112 electromagnetic relief valve open. 13:44:34 M/G set auto transferred back to AC drive. 13:44:36 #112 electromagnetic relief valve closed. 13:46:29 #121 relief valve open 13:46:57 #111 relief valve open 13:46:57 #112 relief valve open 13:47:02 #121 relief valve closed 13:47:02 #112 relief valve closed 13:47-14:14 Repeated relief valve operation 13:50 Clean Up System restored to service 14:14 Reactor level and pressure under nanual control. Reactor level below emergency condenser no::les. 14:14-14:24 Operating relief valve to reduce pressure 14:24 Emergency condenser placed in service manually. L

r- u

                                    /%-                              p.

Hr. Donald J. Skovholt April 18, 1972 Analysis of the Trip All of the sensors in channel 11 reactor protection system tripped causing a 1/2 scram due to the fail safe design of the system on loss of power. The loss of power to parts of the feedwater system caused a feedwater up-set, control valve lockup and subsequent reactor scram due to low water level. The reactor protection system M/G sets are equipped with monitoring devices on the AC motor drive power which causes a transfer to DC motor drive from the station battery when the AC power source is outside specified limits. The input under-frequency monitor caused the transfer just prior to the scram. The transfer trips the AC supply to the motor drive and closes the contactor for the DC motor drive. Transfer to DC drive did not occur causing the M/G set to coast down. This resulted in loss of generator out-put voltage to the reactor protection system channel 11, parts of the feed-water control system and the clean up system.- The M/G set should have returned to normal AC drive after a two minute delay, restoring the generator output voltage as part of an automatic oper-ation. This occurs because the AC power to drive the M/G set is monitored after a transfer to DC drive and allows a transfer to normal if the AC power has been within specified limits for two minutes. Analysis of the data shows that the M/G set did transfer to AC drive after two minutes, but transferred a second time to DC drive resulting in coast down and loss of power for two minutes. 1he feedwater system was operating with the turbine shaft pump and two electric motor driven pumps in service at the time of the trip. The con-trol valves on all pumps locked up on loss of the M/G set output. The out-put of the MG set also supplies some of the control modules in the feed-water control system. The nature of the power failure has made it impossible to determine at what position the feedwater valves locked up. Loss of the ability to control feedwater flow after the scram caused increase reactor water level and the overflow of water into the main steam lines. Several conditions prevail after the scram which made the control of water level dif fi cult . The nain stean isolation valves closed. The clean up system isolated which is the only system through which water can be removed from the reactor. The control rod drive system delive*s 65 gpm to the reactor continuously, from a source outside the reactor. Analysis of data indicates that water overflowed into the main steam lines at about 13 minutes after the scram. The long time suggests that feedwater flow was at some low flow value. The clean up system provides the only means for renoving water from the reactor following a scram. This system isolated at the time of the trip because sensors are powered from the malfunctioned M/G set. Several attempts were made to restore this system after the scram. It was success-fully returned to service about 20 minutc; after the scram. L. ,

r

          -(                     v                               Q) 4 Mr. Donald J. Skovholt                                    April 18, 1972 Analysis of the Trip (cont'd)

The relief valves operated at approximately 13 minutes after the scram, and then repeated several operations for the next 30 minutes at which time water level was brought under control. After water level was under control the emergency condenser was placed in service to assist in holding pressure. Cause of the Malfunction

             'the transfer of the M/G set drive from AC to DC drive motor was caused by an under-frequency monitor. This device is adjusted for a 59.75 IC transfer with a 70 millisecond tine delay. The short time delay makes the device sensitive to transient c1cetrical noise resulting in unnecessary transfers.

This action by itself was not the cause of the malfunction. The malfunction was caused by a ilown fuse in the control circuit for the DC drive. A 6 amp fuse was installed in place of a 10 amp fuse. The fuse stamping was very difficult to read and was the reason for the installation of the wrong size fuse. No electrical probicas could be found with the control circuits. The fuse was replaced and transfer tests were- performed to test the control circuits. Problems with the feedwater control and clean up systems which resulted in high water level were caused by the M/G set malfunction, which in turn, caused these systems to become inoperable or uncontrollable. Corrective Action Tbc time delay of the under-frequency relay has been changed from 70 nilli-seconds to 110 nil 11 seconds. This adjustment makes the device insensitive to electrical noise thus preventing unnecessary transfers. 1he fusesSurveil-in all the M/G sets were checked to insure proper sizes are installed. lance of M/G set control panels will indicate a blown fuse and loss of the transfer control circuit. The'feedwater lockup circuits will be rewired so that all the circuits are not powered from the same M/G set. This will provide manual flok control on at 1 cast one motor pump on loss of power. Conclusion The malfunction was caused by a blown fuse in the transfer control circuit of a continuous N/G set power supply. The wrong size fuse was installed in the control circuit. This caused the M/G set to coast down and the loss of generator output voltage resulted in a feedwater upset and a reactor scram. The wrong sine fuse was initially installed because the fuse stamping was misicading, t

(.

        .-       . s
                                   ,                               r~
           ,                                                       s j                             '

Mr. Donald J. Skovholt April 18, 1972 Conclusivas (cont'd) Increased surveillance of tht M/G sets will afford early detection of control circuit problens and accrease the chance of a raalfunction of a con-tinuous power supply. Very truly yours, Q/ JJ,

                                                                     ,/r
                                                      / &s)w)),0:Jik CalMf0 c @--!

P.-A. Burt General Superintendent, Nuclear Generation J}}