ML20247E019
| ML20247E019 | |
| Person / Time | |
|---|---|
| Site: | Arkansas Nuclear  |
| Issue date: | 03/23/1989 |
| From: | Office of Nuclear Reactor Regulation |
| To: | |
| Shared Package | |
| ML20247E017 | List: |
| References | |
| NUDOCS 8903310271 | |
| Download: ML20247E019 (10) | |
Text
___
ENCLOSURE 1.
urg'o.
UNITED STATES g
8 NUCLEAR REGULATORY COMMISSION o
- T Q
R WASHINGTON, D. C. 20155 SAFETY EVALUATION BY THE.0FFICE OF NUCLEAR.REACTORIREGULATION RELATED.TO. INSTRUMENTATION.FOR DETECTION.0F. INADEQUATE-CORE COOLING' ARKANSAS. NUCLEAR ONE.. UNITS.I AND 2 DOCKET NOS. 50-313.AND.50-368 1
1.0 INTRODUCTION
-On August 10, 1988, a trip was made to the Arkansas Nuclear One, Units I and 2 Nuclear Power Plants of the Arkansas Power and Light Company (AP&L) to evaluate the installation, performance, and maintenance history of Inadequate Core Cooling Instrumentation (ICCI) installt.d in response to the NRC
. requirements of NUREG-0737, Item II.F.2, and Order for Modification of License.
issued on December 10 1982. ThevisitteamconsistedoftheNRCStaff(ICCI technical monitor and project manager) and a consultant from Oak Ridge National Laboratory.
The purpose of the visit was to complete the evaluation of the licensee's conformance to and implementation of the requirements of the NUREG. The participants collected and evaluated information derived from experience of
. installation, calibration and operation of the ANO-1 and 2 ICCI systems installed at the plants. The ANO visit was one of several visits planned to obtain a cross-section of the several different types of systems that were installed by l
different utilities. The information gained will-be used to assist in the j
assessment of the effectiveness and impact of the implementation of the ICCI requirements for all plants.
2.0 DESCRIPTION
OF ICC. INSTRUMENTATION. SYSTEM l
As required by NRC, the ICCI systems at ANO Units 1 and 2 consist of Subcooling Margin Monitoring (SMM), Core Exit Thermocouple (CET), Reactor Vessel Level Monitoring Systems (RVLMS), and for Unit 1 a Hot-Leg Level Measurement System (HLMS,. Unit 1 is a Babcock and Wilcox type reactor and Unit 2 is a Combustion Engineering design. The system installations were approved by NRC and became operational in 1986-87. The level systems have been operational through one or more drain-downs for each unit.
Technical Specification changes for reactor coolant inventory tracking system have not yet been propored by AP&L but are required by NUREG-0737.
2.1 Subcooling Margin Monitoring Unit'I had previously installed a Class IE SMM system which was approved by the NRC in an evaluation dated July 28, 1983. The earlier system used 8903310271 890323
{DR ADOCK 0500 3
hot-leg temperatures and did not include the use of CETs for margin calculations. AP&L has elected to install an additional system as part of the Inadequate Core Cooling Monitoring and Display System (ICCMDS) which includes margin and superhest calculations based on average core exit temperature using qualified CETs. Three-digit digital displays for each channel are provided on the Mimic Display Panel. All calculations are performed in the ICCMDS computers and alarms are provided for sensor or system failures and when subcooling margin temperature falls below the setpoint. Margin can also be displayed on the Fluorescent Display Units (FDU) on the instrument cabinets.
The Unit 2 Subcooling Margin Monitoring system was reviewed and approved by NRC in August 1983 and has not been upgraded further.
2.2 Core Exit Thermocouple System In Unit I there are 32 non-qualified CETs which are monitored by the plant computer and SPDS computer. There are 16 qualified CETs, including 4 per core quadrant, which provide signals in redundant channels to the ICCMDS. The two redundant ICCMDS cabinets provide signal processing and cold-junctiori compensation for the CETs. Eight additional qualified CETs (4 per channel) will be added during refueling outage IR8. The primary CET display is on the SPDS via isolated outputs from the ICCMDS. Qualified backup displays are provided on the Mimic Display Panel by one four-digit LED indicator per channel.
The LEDs display the average of all valid CET inputs. Additional plasma displays are available on the ICCMDS cabinets for individual CET readouts.
In Unit 2 the CET system has 42 CETs with 21 CETs in each of two independent channels. The primary display is the SpDS. The qualified backup displays are Plasma Display Units located in seismically qualified equipment cabinets in the control room. The system complies with the requirements of the NUREG, except that replacement of cable and connectors inside containment with qualified types was not complete as of the site visit.
The replacement was completed during the refueling outage in May of 1988.
2.3 Reactor Vessel Level Monitoring System In Unit I the RVLMS consists of two redundant Radcal level Instruments (RLIs) which contain axially distributed differential thermocouple (DTCs) to detect reactor coolant inventory above the core. The RLI probes extend from the top of the dome to the fuel alignment plate.
Each of the instruments contains nine sensors and an electrical heater which is arranged to heat one junction of each DTC. The resulting temperature difference measurement is related to the surface heat transfer and differentiates the presence or absence of liquid water at the sensor. An annular gas-filled chamber is placed over the region of the heated junction to insn hte the junction from rapid changes of surface heat
7 l'
j.
~..
transfer caused by splashing. The RLIs were developed by the Technology.
for Energy Corporation. The development, testing and qualification program has been reviewed in' conjunction with ICCI reviews of the ANO-2
'and Palisades plants. The ANO-1 probes use only the " slow" sensors'as described-in those reports.
The operating principles of the RLis are, in general, the same as those of the heated junction thermocouple-(HJTC).. The response of the RLIs on uncovery or recovery is qualitatively the same. Because of the differences in some details of their construction, the quantitative-performance is somewhat different.- Specifically, the response times of the RLIs is greater due to the larger heat capacity of the sensors.
The RLIs traverse the entire region to be monitored and are inserted through the ' vessel head.in place of the center control rod assembly.
RLIs are surrounded by manometer tubes and an overall tube within the existing control rod assembly guide tube. The manometer tubes are used to hydraulically isolate the coolant adjacent to the RLIs, thereby enabling the DTCs to sense collapsed liquid level. Hydraulic isolation for dome versus plenum monitoring is accomplished by insertion of hydraulic isolaters in the manometer tubes. The four DTCs in the dome region can be read when the reactor coolant pumps are running. The indication of the DTCs in the plenum region is invalid with the pumps running due to-turbulence.
'The resolution of level indication is limited by the spacing between sensors which is approximately 1.8 ft. The total effective probe length from the head _to the alignment plate is 14 ft. Additional transient error is introduced during level changes due to the sensor response times, data acquisition delays and manometer drain and fill delays. For a maximum small break situation (0.1 ft2) the transient error was estimated to be-7.4 ft for the top sensors decreasing to 2.2 ft for-the bottom sensors.
This represents the actual change in level during the interval between uncovery of the sensor and indication of uncovery on the display.
For smaller break sizes and level change rates, the transient error will be correspondingly smaller.
l The level information from the RVLMS is processed and displayed by the Inadequate Core Cooling Monitoring and Display System (ICCMDS). The ICCMDS consists of two Class IE functionally and physically independent processing and display cabinets, and one dual-channel remote Mimic Display Panel. The display panel is designed to accommodate two independent instrumentation display channels. Each of the cabinets contains all necessary hardware and software required for proper system operation including microprocessor, data communications, multi-bus, input / output and
-real time data acquisition, fluorescent display unit (FDU), and other support electronics and hardware.
Isolated serial data communications channels are provided to the non-Class IE Safety Parameter Display System
t.
l (SPDS). The primary RVLMS level display is considered to be the graphic display on the SPDS. Qualified backup displays are provided on the Mimic Display Panel and the FDUs on the cabinets. The mimic displa level consists of nine pairs of Light Emitting Diodes (LEDs) y of vessel for each channel, indicating the " wet" and " dry" status of each sensor.
In Unit 2 the RVLMS is very similar to that of Unit I and is composed of redundant Radcal Level Instruments (RLI) which contain axially distributed Radcal Gamma Thermometer (RGT) sensors, associsted cabling and signal processing hardware and software. Each RLI probe contains two sensors in the upper head dome region and four sensors in the upper plenum region.
The probes extend into the core and have additional sensors in the core region which are not considered to be part of RVLMS. The guide tubes for the probes function as manometers so that the level sensed by the probes is equivalent to the collapsed liquid level of a two-phase mixture. The sensors detect liquid uncovery at discrete levels within the probes. The I
basic performance of the RLI system is very similar to that of the Combustion Engineering Heated Junction Thermocouple system, except that response times are significantly longer.
The-systeni is qualified for the small LOCA environment in containment and has redundant channels and power sources as required. The uncertainty of level indication at the discrete sensor locations under steady state conditions is less than one foot. The resolution of indication is limited by the number of sensors (six) distributed over the height of 20 ft from the top of the head to the top of the core. The maximum distance between adjacent sensors is 4 ft. The total response time for indication of uncovery at a sensor was originally specified to be less than one minute.
The actual response is slower, up to 2.5 minutes. The slow response gives rise to dynamic error during a transient such as the changing level associated with a small-break LOCA. That is, if the coolant level is falling rapidly, by the time uncovery is indicated at one sensor the actual level may be at a lower level. This response may limit the usefulness of the devices for the maximum size small-breaks considered.
2.4 Unit 1 - Hot-Leg Level Monitoring System Hot-Leg water level monitoring is based on one wide range and four narrow range differential pressure measurements per hot-leg. The wide range covers the full-length of the hot-leg from the top of the candy cane to the bottom of the hot-leg on the decay heat drain line (588 in.). The four narrow range transmitters per hot-leg are used to increase the accuracy of the level measurement and to monitor water level in four areas of interest to the plant operator during an ICC event. The narrow range transmitters also provide additional redundancy. The top narrow range transmitter NRI has a span of 106 in, and is located in the upper candy cane portion of the hot-leg and provides early detection of voiding. The second transmitter, NR2, is located from near the midpoint of the hot-leg l
to the bottom of the range of NR1 and has a span of 200 in. This range j
I l
m:
x
_t.t.L c.
L
-5.
1...
I:
}
includes the area of EFW nozzles and provides information on recovery of level in the reflux boiling mode. The third range, NR3, extends 141 in, below NR2 and also provides information on recovery of level in the reflux boiling mode..The bottom narrow range transmitter, NR4, extends'141 in..
from the bottom tap of NR3 to the decay heat drain line tap. This range monitors level in the region around the reactor coolant pump cold-leg sp111over and also indicates reactor vessel level from the tap to the-bottom of the hot-leg nozzle.
Each dp transmitter uses an RTD on its reference leg for reference leg.
temperature compensation _and valid CET average values for process temperature compensation. The redundant microprocessors in the ICCMDS cabinets perform the data processing and compensation calculations. The estimated uncertainty for the wide range hot-leg indication is approximately +20 to 24%. Uncertainties for.the narrow range indications r4nge from 16% to 40%.
Primary display of hot-leg level information is on the SPDS. Redundant
. qualified backup displays are provided on the Mimic Display Panel and on the FDUs located on the ICCMDS cabinets. The Mimic Panel Displays for hot-leg level are bar-graph indicators. All of the transmitters, connectors, cables, signal processing equipment, and backup displays are Class 1E, qualified for the small-LOCA environments.
c 3.0. INSTALLATION AND MAINTENANCE-EXPERIENCE
.The completed ICCI systems have been operational in both units for at least one fuel cycle. Some difficulties were encountered during installation of the level probes in Unit 2 which required modification of the guide tube design and delay of the installation for one fuel cycle.
Since installation, neither RYLMS system has had any major problems nor have they required extraordinary maintenance.
In both units individual sensor failures have occurred, but these failures were not sufficient to consider the probes inoperable. No significant problems were encountered with the SMM or CET systems for either unit.
The upcoming refueling outage of Unit I will be the first opportunity to perform on-line calibration of the HLMS since installation. During installation, it was discovered that venting of the Narrow Range differential pressure systems somewhat tricky because of the multiple narrow range systems at different elevations.
It was indicated that using system pressure worked better for the venting process than using an external water source.
It is expected that some of the instrument locations along the hot-leg may be near radiation " hot-spots" and that special shielding may be necessary during maintenance and calibration to satisfy the ALARA criteria. Calibration and venting is expected to require approximately two man-weeks.
10
'e?
L Since qualification and ' upgrade of the CET and SMM systems, very few' problems have been experienced.- Qualification of cables and connectors in containment-were completed at the last shutdown for Unit 2 (2R6).
4.0; PERFORMANCE. TECHNICAL SPECIFICATION
S. PROCEDURE
S AND TRAINING 4.1L Operational Problems The CET and SMM portions of the system are used routinely and perform well.
Since RVLMS.and HLMS are post-accident monitoring systems and give no useful information in normal operation, their acceptance is less'well defined.
Operators have observed the tracking of the ICCI'1evel indications with other level measurements (e.g., tygon manometers) during draining of the primary systems and have found reasonable agreement. At the time of this visit, Unit l
2 was drained to the mid-line of the bottom of the hot-leg and the level was
~
indicated correctly on the mimic panel. The operators have noticed'the very slow response time of the RVLMS indications on both units and this may cause them to be somewhat distrustful.
Further improvement is needed.
After further review of the tests performed by Technology for Energy Corporation (TEC) and other Radcal data, AP&L has determined that the signal doubling time for a " slow" sensor is about 92 seconds. The 18 second time given in the Final Design Description was based on a TEC report which is purely theoretical. This change, in conjunction with recalculated worst-case I/O hardware delay times, means that the uncovered indication delay time for the TEC 601 display is changed from about 26 seconds to about 129 seconds for
'a " slow" sensor. The worst-care delay time for the SPDS display uncovery indication is about 160 seconds for the " slow" sensor.
In Unit I all sensors are of the " slow" type.
In Unit 2, the top sensor'in the dome is a " fast" sensor. The worst-case delay for the response of this sensor is 48 seconds at the TEC 601 display and 78 seconds at the SPDS.
It is stated in the Implementation Letter Report for Unit 2 that, "This sensor is important for early detection of ICC and its response time is significantly less than the 2.5 minutes (Ref. 1)." This response time may still be short enough compared to.the 20-30 minutes required to uncover the core in a maximum small-break event (0.1 ftr),
4.2 Technical Specifications Technical Specifications covering ICCI, except reactor coolant inventory tracking system, have been submitted for NRC review. AP&L has proposed that no Technical Specifications are required for RYLMS and HLMS, citing the Commission Interim Policy on Technical Specifications Improvement (Ref. 2).
This is not acceptable for two reasons:
(1) It is staff policy that licensees should not be allowed to remove existing requirements from their Technical Specifications based solely on the Policy Statement criteria pricr to completing the program of converting to a new STS.
t
.4
'o
' (2) It was. concluded, during the Technical Specification split review phase of the Technical Specification Improvement Program, that Post-Accident Monitoring Instrumentation that satisfies the definition of Type A variables in Regulatory Guide 1.97, " Instrumentation for Water-Cooled Nuclear Power i
Plants to Assess Plant and Environs Conditions During and Following an Accident," meets Commission Criterion 3 and should be. retained in the TechnicalSpecifications(Ref.2). Type A variables provide the primary information required to permit the control room operator to take specific manually controlled actions for which no automatic control is provided and that are required for the safety systems to accomplish their safety functions for design basis accident events. Although Type A variables are to be identified on a plant specific basis, the December 10, 1982 Order to B&W Owners requires that the ICC instrumentation be integrated into emergency procedures and submitted for NRC review and approval. Therefore, we expect that all B&W Owners who have integrated the ICC instrumentation into emergency procedures in a manner acceptable to the staff will identify the variables monitored as Type A variables. On that basis, we expect that ICC instrumentation will be retained in the Post-Accident Monitoring table of the new STS.
Thus for Unit 1 and Unit 2 the RVLMS should be included in the appropriate Instrumentation and Surveillance Requirements TS sections. For Unit 1, the HLMS should be included in the appropriate Instrumentation and Surveillance Requirements sections.
4.3 Procedures The Emergency Operating Procedures (EOP) have been revised to include ICCI (Ref.3). The procedures are symptom-oriented and do not refer the operator to specific instruments. No specific entry points using vessel level indications are cited.
Information required for procedural steps is usually available from several sources and cross-confirmation is expected. The procedures for both units are generally based on the B&W ATOG guidance, which is still under the staff review for unresolved issues and the Reactor Vessel Level Monitoring System is one of them. As stated in Section 4.2 of this evaluation, the RVLMS shall be integrated into the E0Ps.
4.4 Training Operators have received specific training concerning the indications and use of ICC instrumentation. The primary displays for ICCI are the SPDS monitors which include other plant information. The operators are also instructed in the use of the mimic displays. The mimic displays have been installed at the plant simulators, but have not yet been made software operational.
4.5 Human Factors and Operators'. Perceptions The mimic displays ar.d the SPDS screens seem to have been adequately planned and are readily interpreted by the operators. These have been reviewed as part of the CRDR process. A visit to the control rooms
l
+
l.
.~
O
~,.
revealed that the primary and secondary displays are adequately located-for operator access and visibility. This observer had difficulty reading the smaller legends on the mimic panel from the operating console position. Operator experience should overcome this minor difficulty. The operators we interviewed were acutely aware of the slow response of the RVLMS systems and suggested that their utility was reduced for draining and filling operations as a result.
5.0
SUMMARY
AND CONCLUSION The staff concludes that the ICC instrumentation systems are acceptable for implementation, although the performance of the RVLMS systems is marginal because of the large level uncertainties and the slow response time. Some individual level sensor failures have occurred, but not enough to be an immediate concern. Individual sensors are not re expensive ($125,000) and require long lead time (pairable and new probes are 30 weeks) for delivery.
Technical Specifications for the level instrumentation must assure that significant further degradation of the RVLMS performance (because of too many failed sensors) is not permitted while the reactor continues to operate.
The Technical Specifications for ICCI systems as stated in the NUREG-0737 Item II.F.2 are required by the Order issued on December 10, 1982. Guidelines for the Technical Specifications ~for ICCI systems are given in GL 83-37.
6.0 REFEREtlCES 1.
NRC Memorandum from D. M. Crutchfield to R. Lee, dated October 8,1986,
" Evaluation of Arkansas Power and Light Company's Inadequate Core Cooling Instrumentation for ANO-2."
2.
NRC Memorandum from J. H. Sniezek to Division Directors, dated March 30, 1988, " Implementation of the Commission Interim Policy Statement on Technical Specifications Improvement."
3.
ANO Inadequate Core Cooling Monitor and Display, 1105.08, Revision 0, November 11, 1986.
Dated: March 23, 1989 Principle Contributor:
T. L. Huang
ENCLOSURE 2 4
TECHNICAL SPECIFICATION. GUIDANCE FOR ANO-1 AND.ANO-2 REACTOR COOLANT INVENTORY TRACKING SYSTEM (RCITS)
Per " Order for Modification of Licensee" issued on December 10, 1982, all of the B&W reactors are required to have hot-leg d/p measurement extending from the top of the hot leg candy cane to the low point in the hot-leg in addition to a Reactor Vessel Level Monitoring System (RVLMS). According to NUREG-0737 Item II.F.2, Technical Specifications for the instrumentation will be required. Guidance on Technical Specifications for NUREG-0737 items was issued on November 1, 1983 via Generic Letter No. 83-37.
The RCITS for ANO-1 consists of a Reactor Vessel Level Monitoring System (RVLIMS) and a Hot Leg Level Monitoring System (HLLMS). The RVLMS consists of tworedundantRadcalLevelInstruments(RLI)whichcontainnineaxially distributed differential thermocouple (DTCs) (4 in the dome region and.4 in the upper plenum region). The HLLMS consists of two redundant channels of one wide range (WR) and four narrow range (NR) differential pressure measurements.
The RCITS for ANO-2 consists fo a RVLMS and is composed of two redundant RLIS
'which contain 7 axially distributed differential thermocouple (2 in the dome regionand4intheupperplenumregion).
The Technical Specifications (TS) for the ANO units are plant specific since the RCITS designs are different from the generic systems which were approwd by the staff. The staff position on the Technical Specifications for both ANO-1 and ANO-2 RVLMS is the same as the position taken for CE RVLMS. That is the RVLMS shall be included in the Accident Monitoring Instrumentation Table 3.3-10, and in Surveillance Requirements, Table 4.3-7 for Unit 2 and Tables 3.5.1-1 and 4.1-1 for Unit 1.
The guidance for this part of Technical Specifications with the action statements is described in GL 83-37. However, the channel operability of the RLI is defined as folhys. For ANO-1, a channel is operable if a minimum of three sensors in the upper plenum region and two sensors in the dome region are operable. For ANO-2, a channel is operable if a minimum of two sensors in the upper plenum region snd one in the dome region are operable.
In addition, the HLLMS for ANO-1 shall be
?-
L considered as a part of the RCITS. Therefore, the HLLMS shall also be included in the Accident Monitoring Instrumentation and in Surveillance Requirements in accordance with the GL 83-37 guidance. The channel operability of the HLLMS shall be defined as follows: A channel is operable in one WR and any two NR d/p transmitters at the same hot leg are operable.
. - _.