ML20207B429
| ML20207B429 | |
| Person / Time | |
|---|---|
| Site: | Calvert Cliffs  |
| Issue date: | 07/14/1986 |
| From: | Tiernan J BALTIMORE GAS & ELECTRIC CO. |
| To: | Thadani A Office of Nuclear Reactor Regulation |
| References | |
| RTR-NUREG-0737, RTR-NUREG-737, TASK-2.A.2, TASK-TM NUDOCS 8607180004 | |
| Download: ML20207B429 (11) | |
Text
. - -
e B ALTIMORE GAS AND ELECTRIC CHARLES CENTER P. O. BOX 1475 BALTIMORE, MARYLAND 21203 JosEPN A.TIERNAN
%cE PRESIDENT NUCLEAR ENERGY July 14,1986 U. S. Nuclear Regulatory Commission office of Nuclear Reactor Regulation Washington, D. C. 20555 ATTENTION:
Mr. Ashok C. Thadani, Director PWR Project P' ectorate #8 Division of PWR Licensing-B
SUBJECT:
Calvert Cliffs Nuclear Power Plant Unit Nos.1 & 2; Docket Nos. 50-317 & 50-318 Emergency Preparedness, Dose Assessment Capability
REFERENCES:
(a)
Letter from Mr. R. E. Denton (BG&E), to Mr. J. R. Miller (NRC),
dated April 15,1985 Gentlemen:
Reference (a) forwarded a preliminary report (first three months of 1985) on the reliability of the dose assessment computer (MIDAS) used for initial dose estimates. As requested by your staff, a report addressing MIDAS reliability for the full 1985 calendar year is provided in Enclosure (1).
As the enclosed report indicates, the availability of the MIDAS initital dose estimate has increased from 77% for the first three.nonths of 1985 to 85% for the full year. This improvement is due in large measure to an increased surveillance frequency which began in May 1985, and to certain hardware changes as discussed in the report. The results of this report are conservative in that no credit is taken for manual input of information upon an unsatisfactory initial automatic response of the MIDAS system. The availability of the MIDAS initial dose estimate for the first five months of 1986 has been 91%.
It is important to note that the availability of the meteorological parameters required by Supplement I to NUREG-0737 is independent of the availability of the MIDAS initial dose estimate. The required meteorological parameters can be obtained directly from an onsite Data Recording Terminal, bypassing the offsite MIDAS system. Therefore, the availability of the MIDAS initial dose estimate should not impact the acceptability of BG&E actions regarding TMI Action Item III.A.2.2. Since May 16, 1985, the date when new meteorological instruments were incorporated into the Technical Specifications, joint availability of wind speed, wind direction, and atmospheric stability data in the control room has been greater than 93%. This performance satisfactorily meets the O05 8607180004 860714 PDR ADOCK 05000317 i
\\\\
F PDR t
F~
Mr. Ashok C. Thadani July 14,1986 Page 2 guidance, at least 90%, provided in American National Standards Institute /American Nuclear Society - 2.5 - 1984.
Should you have further questions regarding this reply, we will be pleased to discuss them with you.
Very truly yours, 3AT/MTF/GCR/ dim Enclosure cc:
D. A. Brune, Esquire J. E. Silberg, Esquire l
D. H. Jaffe, NRC T. Foley, NRC
ENCLOSURE 1 REPORT ON CALVERT CLIFFS' DOSE ASSESSMENT COMPUTER RELIABILITY AS USED IN INITIAL DOSE ESTIMATE This report summarizes the findings of the Emergency Planning Unit twelve month MIDAS Dose Assessment Computer Reliability Study.
HARDWARE
- 1. Digital Graphics Incorporated's mainframe computer, a VAX 11/750 (VAX '11/750).
- 2. Digital Graphics Incorporated's backup computer, a Sperry Univac V77 (DGI V77).
- 3. Calvert Cliffs' CRT terminals, four Techtronix 4014 terminals.
- 4. Calvert Cliffs' meteorological sensors, a set of MRI 60 meter windspeed, wind direction and temperature differential sensors (MET).
l
- 5. Calvert Cliffs' meteorological and radiological Data Recording Terminal computer, a Sperry Univac V77 (V77 DRT).
- 6. Chesapeake and Pot'omac's phone lines and Calvert Cliffs' modems (C&P lines / modems).
l SOFTWARE l
- 1. Pickard, Lowe, and Garrick's MIDAS dose assessment system (MIDAS).
- 2. Data Recording Terminal meteorological data collection system.
PROCEDURE The study period was from January 1,1985 to December 31,1985. Data collected consisted of an initial Dose Assessment Computer (DAC) printout, available in the Control Room or Technical Support Center Annex. The printout consists of an EPZ map showing plume isopleths for whole body gamma exposure rate, windspeed, wind direction, stability class, release rate, and emergency action level classification based on whole body exposure rate.
- Page 1 -
e Acceptance criteria for operability was that the DAC would produce the printout automatically after being given the five commands listed in the Emergency Response Plan Implementation' Procedures (ERPIP 3.0, page7).
This procedure is designed for Control Room use in initial assessment of a radiological release. It may be used by a Control Room Operator, the Shift Supervisor, the Interim Radiological Assessment Director or the Radiological Assessment Director.
All meteorological data and radiological data are acquired automatically by MIDAS from the V77 DRT. No manual inputs were used for the purpose of the study. No backup meteorological data, radiological data or calculations were used in the study.
Surveillance of the system was on a twenty-four hour basis, seven days per week for the time of study. Data samples were taken on a once per shift basis for three shifts per day.
The system was considered to be unavailable from the time a test failed until the time a successful test was performed. Conversely, the system was considered to be available from the time a successful test was performed until a failure was experienced or notification of downtime was given.
For this study, operating time was considered to be the time one or both reactors were above cold shutdown status. The time both reactors were in cold shutdown (507 hours0.00587 days <br />0.141 hours <br />8.382936e-4 weeks <br />1.929135e-4 months <br />) was deducted from the total annual time (8760 hours0.101 days <br />2.433 hours <br />0.0145 weeks <br />0.00333 months <br />) leaving the total annual operating time (8253 hours0.0955 days <br />2.293 hours <br />0.0136 weeks <br />0.00314 months <br />). No credit was taken for any successful use of MIDAS during cold shutdown of both reactors.
The total number of hours unavailable was ratioed to the total hours the system was tested during the study.
Since a MIDAS software program is maintained on a backup computer (DGI V77) for support during routine maintenance, time for routine maintenance was not deducted from l
l availability unless a test of the backup system failed.
All data collected are saved in EPU files. Hard copies of the DAC printouts are maintained in a logbook.
l l
l
- Page 2 -
l l
RESULTS is a summary of cumulative and per month availability. The DAC was available eighty-five percent of the time in 1985. is an hourly breakdown of the availability per month and per annum.
Out of 8,253 hours0.00293 days <br />0.0703 hours <br />4.183201e-4 weeks <br />9.62665e-5 months <br />, the system was available for 7,054 hours6.25e-4 days <br />0.015 hours <br />8.928571e-5 weeks <br />2.0547e-5 months <br />. is a monthly summary of downtime and cause of unavailibilty.
January had the highest downtime, September had zero downtime. is a system diagram provided for information.
DISCUSSION The single component contributing most to the system unavailibility is the V77 DRT. Thirty-two percent of the system unavailibility (388 hours0.00449 days <br />0.108 hours <br />6.415344e-4 weeks <br />1.47634e-4 months <br />) was directly associated with the hardware and software of this computer. Most of this time occurred in January when intensive maintenance was required to correct the problem. Difficulty in obtaining replacement parts extended the maintenance time.
The cause of twenty-one percent of the unavailibility (246 hours0.00285 days <br />0.0683 hours <br />4.06746e-4 weeks <br />9.3603e-5 months <br />) was not easily determined. Most of the undetermined time lost occurred throughout the year. Typical of this kind of lost time resulted when the VAX'll/750 was unable to obtain current MET data from the V77 DRT. The cause of this poor current MET data communication was hard to trace but judging from the high recovery of daily MET data (greater than ninety-five percent; ref: Monthly Meteorological Report, EPU File 10.3.2) the V77 DRT to VAX ll/7501nterface appears to have been at fault. This interface could have been weakened by line and modem failures as well as calculation stalls. When these undetermined causes arose, subsequent checks of the system would eventually prove successful with the problem being resolved without maintenance effort. Another type of problem in this category resulted from heavy use of the VAX 11/750 by other users causing long delays in
- Page 3 -
a aquisition of MET data by MIDAS. In most cases if the system delayed for more than five minutes, the user logged off the DAC and reported a test failure, when the problem was actually orily a delay.
Line and modem problems accounted for twenty percent (242 hours0.0028 days <br />0.0672 hours <br />4.001323e-4 weeks <br />9.2081e-5 months <br />) of thelost time. Replacement of modems has improved communication between the VAX 11/750 and the V77 DRT. Two percent of the 'line and modem pro 5lems was due to' lightning damage.
Maintenance on the VAX 11/750 accounted for nineteen percent (233 hours0.0027 days <br />0.0647 hours <br />3.852513e-4 weeks <br />8.86565e-5 months <br />) of the lost time. Eighty-eight percent (205 hours0.00237 days <br />0.0569 hours <br />3.38955e-4 weeks <br />7.80025e-5 months <br />) of this time was lost as a result of lightning damage. Four hours existed where the backup system was unavailable during preventive maintenance downtime on the VAX 11/750.
Seven percent (90 hours0.00104 days <br />0.025 hours <br />1.488095e-4 weeks <br />3.4245e-5 months <br />) of lost time was due to MIDAS software problems.
PLANNED ACTIONS Evsluate methods for improving the V77 DRT to yield increased reliability. Higher levels of msintenance should be able to bring the overall reliability up.
Continue routine surveillance by EPU (preventive maintenance procedure number ET-D-18) on a daily basis. This surveillance will be continued in order to provide data for improving system reliability.
Evaluate the installation of a dedicated ~line for interface with the mainframe computer. This will reduce chil up time and line weaknesses that result in periodic line noise and calculation interference.
- Page 4 -
ATTACHMENT 1 DOSE ASSESSMENT COMPUTER AVAILABILITY 3ANUARY 1985 TO DECEMBER 1985 11 donth A'.erage 61 %
76 %
7'%
80 %
83 %
85 %
82 %
84 %
86 %
85%
86 %
85 %
3AN FES MAR APR MAY 3UNE JULY AUG SEPT OCT NOV DEC 100 %
I 100 %
95%
94 %
94 %
95%
93 %
84 %
80%
78%
78%
70%
64 %
61 %
60%
50%
40%
30%
20%
10%
0%
Percent Availability = (8253) total hours - (1199) hours available = (85)% for Year to Date (8253) total hours 1
1985 MIDAS OPERABILITY DATA Monthly Operability Hours Monthly Percent Operability Accumulative Yr. Hrs, PercInt Operability Ycar ~
Theoretical Actual Actual
. Theoretical Theoretical
- Accumlative Actual Ops.
Theoretical Ops. Hrs.
Down Time Ops. Hrs.
Ops. Hrs.
Accumulative Down Time Hrs. Accumulated' Hrs. Accum61ated O ps. H rs.
January 744 - 293 = 451 4514 744 = 61%
744 - 293 = 451 451 + 744 = 61%
February 672 - 50 = 622 622 + 672 = 93%
1416 - 343 = 1073 1073 + 1416 = 76%
March 744 - 165 = 579 579 + 744 = 78%
2160 - 508 = 1652 1652 + 2160 = 77%
April (Both Units in Simultaneous Cold Shutdown 94 hrs.
720 - 94 = 626 Theoretical Hours) 626 - 41 = 585 585 + 626 = 94%
2786 - 549 = 2237 2237 + 2786 = 30%
May (Both Units in Simultaneous Cold Shutdown 197 hrs.
744 - 197 = 547 Theoretical Hours) 547 - 25 = 522 522 + 547 = 95%
3333 - 574 = 2759 2759 + 3333 = 83%
June 0 - 45 = 675 675 + 720 = 94%
4053 - 619 = 3434 3434 + 4053 = 85%
July (Both Units in Simultaneous h
Cold Shutdown 168 hrs.
- 744 - 168 = 576 Theoretical Hours) g 576 - 205 = 371 3714 576 = 64%
4629 - 824 = 3805 3805 + 4629 = 82%
August (Both Units in Simultaneous Cold Shutdown 48 hrs.
744 - 48 = 696 Theoretical Hours) 696 - 32 = 664 664 + 696 = 95%
5325 - 856 = 4469 4469 + 5325 = 84%
i September 720 - 0 = 720 720 + 720 = 100%
6045 - 856 = 5189 5189 + 6045 = 86%
I October 744 - 164 = 580 580 + 744 = 78%
6789 - 1020 = 5769 5769 + 6789 = 85%
November 720 - 56 = 664 664 + 720 = 92%
7509 - 1076 = 6433 6433 + 7509 = 86%
December 744 - 123 = 621 621 + 744 = 84%
8253 - 1199 = 7054 7054 + 8253 = 85%
Page 1 of 2 ATTACHMENT 3 MONTH HOURS DOWN CAUSE OF UNAVAILABILITY JANUARY 49 V77 DRT hardware maintenance 244 V77 DRT teletype maintenance 293 total FEBRUARY 50 undetermined causes 56 total MARCH 54 V77 DRT sof tware 33 C&P lines / modem 4
MIDAS sof tware file error 74 undetermined causes IB total APRIL 19 MIDAS sof tware 7
V77 DRT software 15 undetermined causes Tf total MAY 15 V77 DRT modem 10 MIDAS software 25 total JUNE 18 C&P lines / modem lightning damage 2_7 MET tower modem lightning damage 7
45 total JULY 139 VAX 11/750 lightning damage 66 VAX 11/750 maintenance 205 total AUGUST 32 undetermined causer 32 total
Page 2 of 2 ATTACHMENT 3 MONTH HOURS DOWN CAUSE OF UNAVAILABILITY SEPTEMBER 0
OCTOBER 149 C&P line parity error 15 undetermined causes ITV total NOVEMBER 34 V77 DRT hardware 22 undetermined causes 36 total DECEMBER 19 MIDAS software 4
VAX 11/750 scheduled maintenance with no backup DGI V77 available 38 MIDAS system change 24 VAX 11/750 hardware 38 undetermined causes 123 total
ATTAC11 MENT 4 MIDAS SYSID(
BMX-UP ET SDEORS rY ANAIOG MIDAS STRIP OIARIS g
ROCKVIllE VAX 11/750 MAllmE 7*
DISPULY L
a u x.
ROCN PRIMARY ET SENSORS fI ANAIOG DATA RECORDIIC STRIP OIAKIS THIMINAL ANAIOG PRIVATE LI!E DISPIAY V77 gg HISIORICAL ET g
g.7 gg TOER
'IRAIIER CENIBt IOCAL LIlE ET DATA 5
N N
l
/ l
/
l
(
4014 4014 4014 4014 CRT CRT CKr CKr EIFfIRIC CONIBOL ye 1sST me xxu
-