ML20154J676
| ML20154J676 | |
| Person / Time | |
|---|---|
| Site: | University of Michigan |
| Issue date: | 10/09/1998 |
| From: | Jun Lee MICHIGAN, UNIV. OF, ANN ARBOR, MI |
| To: | Michaels T NRC (Affiliation Not Assigned) |
| References | |
| 20, NUDOCS 9810150305 | |
| Download: ML20154J676 (8) | |
Text
{{#Wiki_filter:. a
- j. .
M:CHIGAN MEMORIAL-PHOENIX PROJECT gg PHOENIX MEMORIAL LABORATORY FORD NUCLEAR REACTOR ANN ARBOR, MICHIGAN 48109-2100 (he of the Director October 9,1998 Theodore S. Michaels, Senior Project Manager Non-Power Reactors and Decommissioning Project Directorate Division of Reactor Program Management Office of Nuclear Reactor Management U.S. Nuclear Regulatory Commission Washington, D.C. 20555-0001 Docket 50-2, License R-28 ' Re: Reportable Occurrence No. 20, Inadequate Implementation of Calorimetry Procedures
Dear Mr. Michaels:
On September 25,1998, Reactor Operators made an error at the Ford Nuclear Reactor (FNR) in implementing calorimetry procedures to raise the power level from 1.0 MW to the license power of 2.0 MW. This is written to follow up on a preliminary report on the incident submitted to you on September 28. Together with a dscription and review of the event, we present an analysis of calorimetry data which indicates that the FNR did not operate above the license power in spite of the procedural error, ,
- 1. Description of the Event On Friday, September 25,1998 at 18:39 the FNR was started up in preparation for 2.0 MW operation following a three-week shutdown. At 19:29 reactor power was increased from low power to a nominal 1.0 MW (50% on the 2 MW range of the Linear Level channel) and the reactor placed in automatic control for a calorimetric determination of core thermal power through OP-106, Power Level Determination.
The calorimetry was completed and analyzed to determine the true 1.0 MW Linear Level control setpoint at approximately 20:51. At 20:58 the setpoint for automatic control was reduced from 50% to 49% to bring the true reactor power to 1.0 MW as determined by the calorimeter. The swing-shift Lead Reactor Operator contacted the Assistant Manager for Operations, who was the On-Call Supervisor that night, and obtained permission to increase reactor power to 2.0 MW. Power was increased and at 21:12 the reactor was put into automatic control at a Linear Level channel setpoint of 99%. The calorimetry procedure required a Linear Level automatic control setpoint of 98%. I At approximately 23:30 the Interim Reactor Manager visited the Control Room to check on progress prior to leaving the facility. The Lead Reactor Operator brought to the Manager's 1 attention that he believed the power level to be low. He felt this was the case because the primary coolai inperature nse AT across the core was significantly less than what would be expected at 2.0 MW. The AT was 12.8 0F at 23:00. The 2 MW core AT at the current flow condition of 1000 gpm is expected to be approximately 13.8 oF. In addition, a physical adjustment to the position of Power Level channel ion chambers had been 3 required, to bring the indicated power level up to the 2 MW operating band of 2.0 to 2.1 indicated power. A physical adjustment to the ion chamber positions had not been [ p j1 9810150305 981009 / PDR ADOCK 05000002i 8 PDR Phone:(734) 764-6213 Fax: (734) 763-7863 www umichwiu/~mmpp/ J
2 anticipated. There were no changes to the core configuration or maintenance performed on
, the Power Level channels that would have affected their indicated power to the extent of )
requiring'a physical adjustment to their position subsequent to the last reactor operation at 2.0 MW on September 4. The Interim Reactor Manager reviewed the calorimetry data and did not see any obvious errors. He told the Lead Operator to maintain the current power level correcting for shim rod shadow and that the discrepancy would be investigated on Monday, September 28. The Interim Reactor Manager was not informed at the time nor was he aware of the unnecessary 1% addition to the Linear Level setpoint the i ead Operator made when power , was increased from 1 to 2 MW at 21:12. Saturday morning, September 26, the Assistant Manager for Operations reviewed the calorimetry data with the intent of settir$g the target operating band for the core AT. It was at this point that he noticed the incorrect Linear Level setpoint (99% instead of 98%) that had been initially established for 2 MW at 21:12 on September 25. At 10:47 the Assistant Manager for Operations lowered the Linear Level setpoint by 1% to compensate for the prior error and contacted the Interim Reactor Manager to discuss the event and determine what additional actions should be taken.
- 2. Safety Significance of the Event At the time of discovery it was believed that the reactor might have operated for a period of approximately 14 hours, between 21:12 Friday, September 25, and 10:47 Saturday, September 26, at a steady-state power level of 2.02 MW thermal, in violation of the license power imel of 2.0 MW thermal. The 1% power increase is believed to be well within experimcaal uncertainties associated with our method of dett. aining thermal power level.
In addition, the safety significance of operating at 2.02 MW thermal is minimal. The FNR Safety Analysis Report shows a steady-state power level of 4.68 MW at the minimum primary coolant flow of 900 gpm and 18 feet of pool head is required before boiling may occur in the hot channel. A second calorimetry experiment was performed on October 1,1998 to establish the true reactor power level. A conservative analysis of the seccnd calorimetry . lata indicates the reactor operated at an actual power not exceeding 1.97 MW during the 14-hour period before the setpoint error was corrected. The analysis of the October I calorimetry is enclosed. l The FNR management feels obligated, however, to consider the possibility that .a different t combination of reactor system parameters, subject to a similar procedural error, could have resulted in a larger power adjustment, e.g., a few per cent, possibly resulting in an overpower condition. Thus, we consider the incorrect implementation of the calorimetry , procedure, OP-106, and delayed discovery of the error as a serious mistake on the part of l the operations staff and management. l l 3. Immediate Corrective Actions Upon discovery on Saturday. September 26, of the error made in the Linear Level setpoint, reactor power was reduced by 1% to compensate for the error. With this 1% power reduction, the actual reactor power level during the subsequent reactor operation until the end of the operation cycle on Friday, October 2, is estimated to have been less than 97.5% oflicense power or 1.95 MW.
. 3 m
- 4. Root Causes of the Event In their ' written descriptions of the event and in oral discussions with the FNR management, the Lead Operator and Reactor Operator, responsible for the procedural error, ,
did not demonstrate sufficient understanding of the significance of the calorimetry in establishing thermal power level of the reactor. There were reasons to suggest that reactor system parameta deviated somewhat from normal ranges, as indicated by the need to adjust the Power Level channel positions. This, however, did not justify any ad hoc adjustment in the Linear Level setpoint, although the Lead Operator indicated that he had no intention ofincreasing the power level beyond the erroneous 1% adjustment. There is a'so indication that the two operators on swing shift did not have sufficient discussica on the system parameters observed and steps required for proper implementation of the calorimetry procedure. Furthermore, they did not bring to the attention of the On-Call Supervisor any concerns they had regarding the results of the calorimetry or the 1% erioneous adjustment made to the Linear Level setpoint. This indicates lack of proper communication and questioning attitudes between the operators themselves and between the operators and supervisory personnel. Finally, the operators did not follow the steps delineated in OP-106 correctly, indicating their lack of attention to detail.
- 5. Subsequent Corrective Actions The Lead Operator responsible for the incident has been relieved of Lead Operator responsibilities for an indefinite period of time. Additional training of the Lead Operator and Reactor Operator, who were on swing shift at the time of the incident, has been initiated this week and will continue over the next two weeks. This includes stringent written examinations on OP-106 and calorimetry performance tests under the direct supervision of the Assistant Manager for Operations.
A staff meeting was held on October 5 to discuss and emphasize tne importance of (a) teamwork and communication, (b) thoroughness and attention to detail, and (c) understanding of the physical basis for operational procedures. A summary of the staff meeting was distributed as a memorandum to the entire staff on October 6. The memorandum has supplemented the existing codes of perfomiance for the FNR staffin the ! retraining of the entire licensed staff this week. Sincerely, 4Q, John C. Lee Interim Director Enclosure l i l 4 4
ANALYSIS OF THE OCTOBER 1,1998 FNR CALORIMETER The Priinary Coolant System at the Ford Nuclear Reactor. . 1 Figure I shows the primary coolant system for the FNR. The primary coolant water from . the pool is sucked through the reactor core. The water enters a holdup tank to allow for N- l 16 decay. Temperature sensor Tl is located at the inlet plenum of the holdup tank. Water is then directed to either of two primary coolant pumps. Pump 2 is currently in use. Pump 2 3ushes the water to the heat exchanger. Temperature sensor T2 is located in the piping xtween pump 2 and the heat exchanger. Water flows from the heat exchanger through an orifice plate and returns to the pool approximately 27 feet below the pool surface. Two ' temperature sensors are located in the pool for bulk pool temperature measurement. T4 is 2 feet below the surface. T5 is at core inlet height 20 feet below the pool surface just to the I north of the core. The Calorimeter Absolute power determination is performed by treating the entire primary coolant system e a large calorimeter. The reactor power is kept constant at some nominal level with the primary coolant pump running'and the secondary coolant system secured. The rate of temperature increase in the primary coolant system is directly proportional to the reactor power level under these conditions. The proportionality constant is equal to the inverse of the total mass of the primary coolant water multiplied by the specific heat capacity of water. Heat losses during the calorimeter originating from conduction through the pool walls, l convection and radiation from the primary process piping and components, and evaporation . and radiation losses from the pool surface are expected to be less than 3.6% 1 and are , assumed to be offset by heat gain from the 20 hp primary coolant pump. Analysis of the Calorimeter Performed October 1,1998 to Verify True Reactor Power. A second calorimeter experiment was initiated after the event that resulted from the calorimeter performed on September 25 prior to 2 MW operation to verify the reactor power level. The reactor power was kept constant at 2 MW as indicated by the September 25 calorimeter i dring the period from September 26 (when the 1% downward correction in power was i made) up to and including the second calorimeter experiment on October 1. Constant power was verified by maintaining the reactor core AT at the 12.8 0F. Results of the second calorimeter are presented in Table 1. h was suspected that true mwer was less than 2 MW based in large part on the observed AT=12.8 0F. The expected 2 MW AT under existing FNR flow conditions is 13.8 0F. However, the power level of 1.76 MW initially determined by the second calorimeter was also below what was expected. This led to further analysis. First,it became apparent that the calorimeter as currently performed puts too much emphasis on the coolant heat-up rate measured in the process piping as compared to the rate measured in the pool. Only 7% of the 48,000 gallons of primary coolant water i resident
_ . . _ . - . _._.m__._.__._ _ _ _ __ _ _ . _ . _ _ ... ._ _ . _ . _ l l . l in the holdup tank, heat exchanger, and piping. The two reactor pool temperature sensors,
. T4 and T5, should be primarily relied upon for this reason.
Second, the pool water had a thermal layer established and had not stabilized from the ovemight cool down at about 0.5 oF/hr that had occurred in preparation for the calorimeter. The temperature measured by T5 would have continued to decrease at approximately this rate until the thermal layer was eliminated g'. -n the absence of the calorimeter experiment. The maximum estimated true power after the event on September 25 and prior to the corrective action taken on September 26 was determined to be 1.97 MW thermal. This estimate was calculated by conservatively taking the highest pool heat-up rate measured during the second calorimeter of 15.84 0F/hr measured by T5, and compensating for 0.5 0F/hr cool down giving an overall heat up rate of 16.34 0F/hr. This gives a measured power of 1.88 MW. Assuming the maximum heat loss equivalent to 3.6% of power increases the estimate to 1.95 MW. Finally, adding the 1% procedural error that was made on September 25 results in a maximum estimated true power 1.97 MW. Table 2 shows the calorimeter experiment result for September 25 when the procedural , error was made and is included for information purposes.
- l. Bullock, J.B., Absolute Power Measurements of the Ford Nuclear Reactor, July 1965, p.4 l
1 l i 1
l
. . l n .-.qure 1 ?.- mary Cooiant System g
,9f * .!
I.. . . ..i..; __ . . _ _ - .
,- _l.i )
)
I.:: , S .* il ' WCDP -@
- 1 I
. l .'
-{,
** Core .,
* . ,e , '
'. ' . 1
....a..
'l -l
*i
( .5==== > h ,,, ,.,. . - Il ! ~ besleriev n
, mas
- a V I 4esman f &
A h(N
.' l I mm s~
f l l [. da memes cz- x- .
<. - ~s ,1 9.a 2m t
- s h
e e ,' O' W
%)/
- %/ . .
.CL"*
. g == ' ,
h e=a nae' / ,
* .no Ree J T* M ~ -umews Of4ae G . ("%
%./
.' # Ibl#
Caas emuse cisunwee W e llads - 5% . ... C'""' ,
$ N.0 . P'****
- U N
't. #
I in 6
-# b WUN MS MO N Leersent fil G. M Us earer m
N'*y fy{ . k yse
}
(N
......o I - , .e
[- OP-l% Power Level Deternunation Revision 17: 041698 0 i Table 1 Cycle No: O Calculated oy: # M '? Date: />II "/ Checked by: 2 2Wd% y Initial Data InitialLinear Level Setpoint 9I % Nonunal Power Level: - MW Poollevel: OM in y Corrected pool heatup rate: /27O
- F/hr/MW W Calculation of Heatup Rate:
BID TIMECoefficient(*F/ min.) Heat Up Rate (*F/hr) n 15 > x 60 min.= nA T3 M Ox x 60 min.= /4. O T4 , Mu x 60 min.= ><,o T5 . eM V x 60 min.= 15. 9Y Total 50 N Average (Total /4) /f, N ~ - (. Actual Thermal Power = Average HUP / Corrected HUR
, , _ . en . -
=( /IM F/hr)/(D N1 F/hr/MW)
= l 79 MW Corrected Lm:arLevel Setpoint
= (Initial LL setpoint) x (Nominal Power / Actual Thermal Power)
=( % %)x( k MW)/(l 767 MW)
= l1),& %
N.I. Chambers Adjusted (if needed): w 'WWed No Ir'u4. hQ On-Call Supervisor Contacted at 1 MW:_. N/A /d g.g g dg Reviewed By: be(db Q ( Review Date: lo[2-/ frr Page 4
OP-106 Power Level Determination Revision 17: 041698 [ Table 2. Cycle No: "1 Calculated by: 5dAGP Date: Rh6' Checked by: NC Initial Data Initial Linear Level Setpoint: TS S % Nominal Power Level: / MW Poollevel: "7 . in
<<724 l
Corrected pool heatup rate: F/hr/Mw Calculation of Heatup Rate: BlQ TIME Coefficient (*F/ min.) Heat Up Rate (*F/hr) T2 . l6/ x 60 min.= 9.06 T3 .If0 x 60 min.= 9,00 T4 ,1W x 60 min.= L 6'/ T5 ,N9 x 60 min.= MV Total 35M Average (Total /4) T 9/ l I Actual Thermal Power = Average HUR/ Corrected HUR l
=(N
- F/hr)/(47M F/hr/MW)
= _/. 0 Q MW Corrected Linear Level Serpoint
= (Initial LL setnoint) x (Nominal Power / Actual Thermal Power)
= ( MO 4 x ( /,00 MW)/( / o'^ MW)
= 09:0 %
N.I. Chambers Adjusted (if needed): // o On-Call Supervisor Contacted at i MW: c }. Reviewed By: U wa N
\
f Review Date: 9 /E8,/9T Page6
, -- wa y}}