ML19260G770
| ML19260G770 | |
| Person / Time | |
|---|---|
| Site: | Sequoyah  |
| Issue date: | 03/16/1981 |
| From: | Miller R, Sharp D, Tuley C WESTINGHOUSE ELECTRIC COMPANY, DIV OF CBS CORP. |
| To: | |
| Shared Package | |
| ML19260G769 | List: |
| References | |
| NUDOCS 8103190183 | |
| Download: ML19260G770 (23) | |
Text
.............................
WESTINGHOUSE REACTOR PROTECTION SYSTEM / ENGINEERED SAFETY FEATURES ACTUATION SYSTEM SETPOINT METHODOLOGY C. R. Tuley D. R. Sharp R. B. Miller WESTINGH0USE ELECTRIC Nuclear Energy Systems P. O. Box 355 Pittsbur gn, Pennsylvania 1E230 8103190 tis 3 '
nta s enorivusc e nvrna t inKt LLMd3 J TABLE OF CONTENTS Section Title Page
1.0 INTRODUCTION
1-1 2.0 COMBINATION OF ERROR COMPONENTS 2-1 2.1 Methocology 2-1 2.2 Sensor Allowances 2-2 2.3 Rack Allowances 2-4 3.0 RESPONSE TO NRC QUESTIONS 3-1 3.1 Approach 3-1 3.2 De finitions for Protection Sys tem 3-1 Setpoint Tolerances 3.3 NRC Questions 3-5 i
LIST OF TABLES TaDie Title Pace 3-1 Tavg Channel Accuracy 3-7 3-2 Overtemperature 6T Channel Accuracy 3-8 1-3 Overpower 6T Channel Accuracy 3-10 3-4 Reactor Protectior. System /Engineerea 3-12 Safety Features Actuation System Channel Error Allowances Notes for Table 3-4 3-13 h
9 4
g'.
e ii
WESTINGHOUSE PROPRIETARY CLASS 3
1.0 INTRODUCTION
In March of 1977, the NRC requested several utilities with Westinghouse Nuclear Steam Supply Systems to reply to a series of questions concern-ing the methodology for determining instrument setpoints.
This document contains the Westinghouse response to those questions with a correspond-ing defense of the technique used in determining the overall allowance for each setpoint.
The information desired pertains to the various instrument channel com-ponents' analysis assumptions, i.e., a channel breakdown and values, for the Reactor Protection System (RPS) and the Engine = red Safety Features Actuation System (ESFAS).
Some of the information requested is already available in public documents, e.g., Chapters 15 and 16 of the Safety Analysis Report.
The rest of the information has not been released and is drawn from equipment specifications or analysis assumptions.
This information is considered proprietary and is noted as such.
The basic underlying assumption used by Westinghouse is that several of the e ror components and their parameter assumptions act independently, e.g.,[
]ta,c, This allows the use of a statistical summation of the various breakdown components instead of a strictly arithmetic summation.
A direct benefit of the use of this technique is increased margin in the total allow-ance.
For those parameter assumptions known to be interactive, the technique uses arithmetic surmiation, e.g, [
] a,c The explanation of the overall approach is provided in Section 2.
Section 3 presents the information requested along with three examples of individual channels, Tavy, Overterstratore AT, and Overpower AT.
Also located in this section are oescriptions, et definitien, of the 1-1
WESTINGHOUSE PROPRIETARY CLASS 3 various parameters used.
This insures a clear uncerstanding of the breakdown presented, in nearly all cases a significant margin exists between the statistical sumation and tne total allowance.
1-2
2.0 COMBINATION OF ERROR COMPONENTS 2.1 Methodolooy The methodology used to combine the error components for a channel is basically the appropriate statistical comoination o f those groups o f components which are statistically independent, i.e., not interac tive.
Those errors which are not independent are added arithmetically into groups.
The groups themselves are independent ef fects wniCn Can then be systematically combined.
The metnodology used for this combination is not new.
Basically it is
]t a,c.e whicn nas been the[
utilized in other Westinghouse -eports.
Th is technique, or other s ta-tistical approaches of a similar nature, have been used in WCAP-9130(1) and WCAP-8567(2)It should be noted that WCAP-8567 nas been approved by the NRC Staff thus noting the acceptability o f statis-tical techniques for the application requested.
It should also be re-cognized that the Instrument Society o f America approves o f the use o f statistical techniques in determing safety-related instrumentation set-po in ts (3)
Thus is can be seen tnat the use of statistical approaches in analysis techniques is becoming more and more widespread.
The relationship between the error components and the total statistical error allowance for a channel is, ta,c
(
3 (Eq. 2.1)
(1) Little, C.C., Kopelic, S.
D.,
and Chelemer, H., " Consideration o f Uncertainties in the Speci fication of Core Hot Chani'el Factor Limi ts. " WCAP-9180 (Proprietary), WCAP-9181 (Non-proprietary),
Septemoer, 1977.
(2) Chel emer, H., Bowman, L. H., and Sharp, D. R., " Improved Theemal Design Procedure," WCAP-8567 (Proprietary), WCAP-8763 (Non-pro-prietary), July,1975.
(3) Instrument Society of America, proposed standard SP67.04, "f e tpo in ts for Sa fety-Related Instrumentation Used in Nuclear Power Pl nts."
2-1
wnere:
ta,c As can be seen in Equation 2.1, [
]ta,c allowances are interactive and thus not independent.
The [
]ta,c is not neccessarily considered interactive with all other parameters, but as an added degree of conservatism is added arithmetically to the statistical sum.
2.2 Sensor Allowances Four parameters are considered to be sensor allowances, [
]ta,c (see Table 3-4).
Of these four parameters, two are considered to be statistically independent, [
]ta,c, and two are considered interactive [
] a,c
{
] a,c are considered to be independent due to the manner in which the instrumentation is checked, i.e., the instrumentation is [
]ta,c An example of this would be as follows;
[
2-2 B
tg a,c
{
]ta,c are considered to oe interactive for the same reason that
[
]ta,c are considered indepencent, i.e.,
due to the manner in which the instrumentation is checked.
[
]ta,c Based on this reasoning, [
] tac have been added to form an. independent group which is then factorad into Equation 2.1.
An example of the impact of this treatment is; for Pressurizer Water Level-High (sensor parameters only):
- a,b,c using Equation 2.1 as written gives a total of; ta,c
- 1.66 percent 2-3
Assuming no interactive effects for any of the parameters gives tne following results:
ta.c (Eq. 2.2)
= 1.32 percent Thus it can be seen that the approach represented by Equation 2.1 wnich accounts for interactive parameters results in a more conservative sum-mation of the allowances.
2.3 Rack Allowances Four parameters, as noted by Table 3 4, are considered to be rack allow-ances,[
]ta,C Three of these parameters are considered to be interactive (for much the same reason outlined for sensors in 2.2), (
]ta,c
[
]Ta,c Based on this logic, these three f actors have been added to form an incependent group.
This group is then factored into Equation 2.1.
The impact of this approach (formation of an independent group based on interactive components) is rignificant.
For the same channel using the same approach outlined in Equations 2.1 and 2.2 the following results are reacned:
24
ta,b.c using Equation 2.1 the result is; ta,c
= 1.82 percent Assuming no interactive e f fects for any o f the parameters yields the following less conservative result;
_ta,c (Eq. 2.3)
= 1.25 percent Thus the impact o f the use of Equation 2.1. is even greater in the area o f rack e f fects than for the sensor.
There fore, accounting for inter-active ef fects in the statistical treatment o f these allowances insures a conservative result.
Finally, the [
]ta,c parameters are considered to bJ independent o f both sensor and rack parameters.
[
] a,c Thus, these parameters have been statistically factored into Equation 2.1.
2-5
WESTINGHOUSE PROPRIETARY CLASS 3 3.0 RESPONSE TO NRC OUESTIONS 3.1 Approach As noted in Section One, Westinghouse utilizes a statistical summation of the various components of the channel breakdown.
This approach is valid where no dependency is present, An arithmetic surrmation is required where an interaction between two parameters exists, Section Two The equation provides a more detailed explanation of this approach.
used to determine the margin, and thus the acceptability of the param-eter values used, is:
ta c (Eq. 3.1) ~
~
where:
[
] a,c, and all other parameters are as defined for Equation 2.1.
Tables 3-1 through 3-3 provide examples of individual cnannel breakdowns and margin calculations for Tavg, Overtemperature AT, and Overpower AT.
It should be noted that only those channels dich Westinghouse takes For credit for in the analysis are provided with detailed breakdowns.
those channels not assumed to be primary trips, there are no Safety Analysis Limits, thus no Total Allowance or Margin can be determined.
3.2 Definitions for Protection System Setooint Tolerances To insure a clear unders+anding of the channel breakdown used by West-inghoue.e in this report, the following definitions are noted:
3-1
WESTINGHOUSE PROPRIETARY CLASS 3 1.
Trio Accuracy The tolerance band containing the highest expected value of the difference between (a) the desired trip point value of a process variable and (b) the actual value at which a comparator trips (and thus actuates some desired result).
This is the tolerance band, in percent of span, within which the complete channel must perform its intended trip function.
It includes comparator setting accuracy, channel accuracy (including the sensor) for each input, and environ-mental effects on the rack-mounted electronics.
It comprises all instrumentation errors; however, it does not include process measurement accuracy.
2.
Process Measurement Accuracy Includes plant variable measurement errors up to but not including the sensor.
Examples are the effect of fluid stratification on tenperature measurements and the effect of changing fluid density on level measurements.
3.
Actuation Accuracy Synonymous with trip accuracy, but used where tha word " trip" does not apply.
4 Indication Accuracy The tolerance band containing the highest expected value of the difference between (a) the value of a process variable read on an indicator or recorder and (b) the actual value of that process variable.
An indication must fall within this tolerance band.
It includes channel accuracy, accuracy of readout devices, and rack environmental effects, but not process measurement accuracy such as fluid stratification.
It also assumes a controlled environment for the readout device.
3-2
4 5.
Channel Accuracy The accuracy o f an analog channel wnich includes the accuracy of the primary element and/or transmitter and modules in the chain where calibration o f modules intermediate in a chain is allowed to compen-sate for errors in other modules o f the chain.
Rack environmental ef fects are not included here to avoid duplication due to dual inputs, however, normal environmental e f fects on field mounted hard-ware is included.
6.
Sensor Allowable Deviation The accuracy that can be expected in the field.
It includes drift, temperature e f fects, field calibration and for the case o f d/p trans-mitters, an allowance for the e ffect o f static pressure variations.
The tolerances are as follows:
Re ference (calibration) accuracy'- [
]tabc percent unless a.
other data indicates more inaccuracy.
This accuracy is the SAMA reference accuracy as de fined in SAMA standard PMC-20-1-1973(1) b.
Temperature e f fect - [
]tabc percent based on a nominal temperature coef ficient of [
] tab c percent /100 F and a 0
maximum assumed change o f 50 F.
c.
Pressure e ffect - usually calibrated out because pressure is constant.
If not constant, nominal [
] abc percent is used.
Present data indicates a static pressure e f fect o f approximately [
]tabc percant/1000 psi.
d.
Ori ft - change in input-output relationship over a period o f time at reference conditions (t:.g., [
]ta,c,
[
]ac o f s an.
3-3
7.
Rack Allowable Deviation The tolerances are as follows:
a.
Rack Calibration Accuracy The accuracy that can be expected during a calibration at refer-ence conditions.
This accuracy is the SAMA reference accuracy as defined in SAMA standard PMC-20-1-1973(1) This includes all modules in a rack and is a total of [
]tabc percent of span assuming the chain of modules is tuned to this accuracy.
For simple loops where a power supply (not used as a converter) is the only rack module, this accuracy may be ignored.
All rack modules individually must have a reference accuracy within
[
]tabc percent, b.
Rack Environmental Effects Includes effects of temperature, humidity, volt.ge, and frequency changes of which temperature is the most significant.
An accuracy of [
]tabc percent is used which considers a 0
0 nominal ambient tempercture of 70 F with extremes to 40 F and 0
120 F for short periods of time.
c.
Rack Drift (instrument channel drift) - change in input-output relationship over a period of time at reference conditions (e.g.,
[
]ta,c
+1 percent of span, d.
Comoarator Setting Accuracy Assuming an exact electronic input, (Note that the channel accuracy" takes care of deviations from this ideal), the (1) Scientific Apparatus Manufacturers Association, Standard PMC-20-1-1973, " Process Measurement and Control Terminology."
3-4
WESTINGHOUSE PROPRIETARY CLASS 3 tolerance on the precision with which a comparator trip value can be set, within such practical constraints as time and effort expended in making the setting.
The tolerarces are as follows:
(a) Fixed setpoint with a single input - [
]tabc percent accuracy.
This assumes that comparator nonlinearities are compensated by the setpoint.
(b) Dual input - an additional [
] abc percent mus+. be added for comparator nonlinearities between two inputs.
Total [
]tabc percent accuracy.
Note:
The following four definitions are currently used in the Str.ndardized Technical Specifications (STS).
8.
Nominal Safety System Setting The desired setpoint for the variable.
Initial calibration and subsequent recalibrations should be made at the nominal safety system setting (" Trip Setpoint" in STS).
9.
Limiting Safety System Setting A setting chosen to prevent exceeding a Safety Analysis Limit
(" Allowable Values" in STS).
(Violation of this setting represents an STS violation).
- 10. Allowance for Instrument Channel Drift The difference between (8) and (9) taken in the conservative direc-tion.
3-5
WESTINGHoubt FRuexitiARY LLads a
- 11. Safety Analysis Limit The setpoint value assumeo in safety analyses.
- 12. Total Allowable Setpoint Deviation Same definition as 9, but the difference between 8 and 12 encompasses
[
, ta c s
3.3 NRC Ouestions The information requested by the NRC for each channel is:
1.
What is the technical specification trip setpoint value ?
2.
What is the technical specification allowable value?
3.
What instrument drift is assumed to occur during the interval between technical specification surveillance tests?
4.
What are the components of the cumulative instrument bias (e.g.,
instrument calibration error, instrument drift, instrument error, etc.) ?
5.
What is the margin between the sum of the channel instrumentation error allowances and the total instrumentation error allowance assumed in the accident analysis ?
The Westinghouse response to these questions is:
The response to Question 1 will be found as Column 14 of Table a.
3-4 in this section.
3-6
WESTINGHOUSE PROPRIETARY CLASS J
b.
Column 13 of Table 3-4 provices the information requested in Question 2.
c.
The instrument drift assumed is the difference between the trip setpoint and the allowable value in the technical specifications, this can be found as Column 11 of Taole 3-4.
d.
The bulk of Table 3-4 provides the breakdown values required by Question 4.
The margin requested by Question 5 is noted in Column 17 of Table e.
34 It should be remembered that Westinghouse is providing responses only for those channels for which credit is taken in the accident analysis.
Again this is due to the fact that Question 5 cannot be answered if the channel is not a primary trip.
3-7
TABLE 3-1 Tavg Channel Accuracy Allowance
- Par ame ter tabc
- in percent o f span.
The margin, based on Equation 3.1, is calculated as follows:
tabc
-[
]tabc The Total Allowance is 4.0%, thus the margin is
[
]ta,b,c o f span.
3-8
atsiinonvuse e n v
.u
. ~...-~..
TABLE 3-2 Overtemperature ai Channel Accuracy Parameter Allowance
- tabc 1
- in percent of span, (1000F span.1!iO % power)
WEbl1N6Huust enve n t u nn i utxas a TABLE 3-2 (Continued)
The Margin, based on Equetion 3.1, is calculated as follows:
_ 'a,b,c t
=[
la.c.
The Total Allowance is 5.9%, thus the margin is (
Ja.c of t
Span.
3-10
WESTINGHOUSE PROPRIETARY CLASS 3 TABLE 3-3 Overpower AT Channel Accuracy Parameter Allowance
- tabc 0
- in percent of span, (100 F span - 150% power) 3-11
WESTINGHOUSE PROPRIETARY CLASS 3 TABLE 3-3 (Continued)
The margin based on Equation 3.1 is calculated as follows:
~
tabc
~
Jabc, The total Allowance is 4.1%, thus the margin is [
]tabc of t
[
spen.
3-12
WESTINGHOUSE PROPRIETARY CLASS 3 NOTES FOR TABLE 3 4 (1)
All values in percent span.
(2)
As noteJ in Table 15.1-3 of SAR.
(3)
As noted in Table 2.2-1 and 3.3 4 of Westinghouse STS.
(4)
Includedin[
t g a,c (5)
Not used in the Safety Analysis.
(6)
As noted in Figure 15.1-1 of SAR.
(7)
As noted in Notes 1 and 2 of Table 2.2-1 of Westinghouse STS.
(8)
[
] a,c (9)
Venturi.
(10) TVA supplied information.
(11)
Includedin[
]ta,c (12) As noted in Table 3.3-4 of Westinghouse STS.
(13)
Does not impact Safety Analysis results.
(14) Trip setpoint function of note (12) plus 16;,.f4.
(15)
Included in [
]ta,c (16)
Not found in Table 15.1-3 of SAR b'ut used in Safety Analysis.
(17)
[
t7ac (18)
[
ty a,c 3-14