ML20212R226

Forwards Addl Info Re Suppl 1 to NULAP-5 Manual,Per 870312 Request.Author Understands That Info Needed to Ensure Review of NULAP-5 Code Prior to 1987 Refueling Outage

ML20212R226
Person / Time
Site:	Haddam Neck File:Connecticut Yankee Atomic Power Co icon.png
Issue date:	04/16/1987
From:	Mroczka E, Romberg W CONNECTICUT YANKEE ATOMIC POWER CO.
To:	NRC OFFICE OF ADMINISTRATION & RESOURCES MANAGEMENT (ARM)
References
A06421, A6421, NUDOCS 8704240355
Download: ML20212R226 (19)

Text

e-CONNECTICUT YANKEE ATOMIC POWER COMPANY B E R L I N, CONNECTICUT P.o. box 270

• HARTFORD. CONNECTICUT 06141-0270 TELEPHONE mese April 16,1987 1 Docket No. 50-213 A06421 Re: 10CFR50.46 U.S. Nuclear Regulatory Commission Attn: Document Control Desk Washington, D.C. 20555

Reference:

(1) F. M. Akstulewicz letter to E. 3. Mroczka, NULAP-5 Questions, dated March 12,1987.

Gentlemen:

Haddam Neck Plant Response to Request for AdditionalInformation on Supplement I to NULAP-5 In Reference (1), nine requests for additional information were made concerning Supplement I to the NULAP-5 manual. Attachment I contains the response to these questions.

It is Connecticut Yankee Atomic Power Company's uryde that satisfactory and timely response to these additional questionstl) will enablerstanding the Staff to complete their review of the NULAP-5 code prior to the 1987 refueling outage.

Should you have any further concerns, please contact us.

Very truly yours, CONNECTICUT YANKEE ATOMIC POWER COMPANY

& P1 fo < t ke E. 3. Mroczka Senior Vice President 8704240355 870416 PDR ADOCK 05000213 L P PDR By: W. D. Romberg Vice President (1) F. M. Akstulewicz letter to E. 3. Mroczka, dated February 12, 1987, forv arded the first round of questions on NULAP-5. This was responded to in an L 3. Mroczka letter to F. M. Akstulewicz, dated March 19,1987. D h

\

t i

Docket No. 50-213 A06421 4-4-

i.

Attachment 1 4

i l Response to Requests for AdditionalInformation on j Supplement I to the NULAP-3 Manual

,i

?

4 1

1

,l i

1 1

i t

1 4

i F

1 1  !

l i

1 i

1 i

f i

i l.

't i

!~

4 i

1 April,1987 0

i

\

.- - - _ , . . . - .. . .- . - - - , . - . . - . - - - - . , - . , . . .-_ _--- -..--.-.~.-._--..

,O r._

1. Table 2.1 of . Supplement 1 presented the ANS decay beat curve to be used f or licensing analyses. This table showed that f or 0.0 s after scram the normalized power was 0.072. Clarify whether this is correct because it would seem that at 0.0 s after scram the normalized power should still be 1.0.

Table 2.1 of Supplement 1 presents only the ANS decay heat curve. This table includes both fission product decay and actinide decay heat power and has been normalized to initial core power conditions. As descri bed in Section 3.1.4 of the NULAPS manual, reactor kinetics is used to calculate the total power production in the core. In general, when the kinetics model is used, total core power is the sm of fission power, fission product decay power, and actinide decay power. For small break LOCA licensing calculations, however, the total core power during the transient is the se of fission power, calculated fra the reactor kinetics model, and decay heat power, calculated fra Table 2.1. As required by Appendix K to 10CFR50, a + 20% uncertainty is applied to Table 2.1 f or licensing calculations. For licensing calculations, the fission product decay power and actinide decay power kinetics calculations are replaced with Table 2.1 values.

m - _ _ _ _

2. Section 2.2 of Supplement 1 descri bed changes made to the interphase condensation model to tetter account for condensation due to ECC injection into vertical cmponents. The following questions are related to these changes:

a. To account for interphase condensation fem ECC injection into vertical cmponents, the option of specifying the Akers, Deans, and Crosser ( ADC) correlation to calculate the condensation heat transfer coefficient in vertical volmes was added to NULAPS. The ADC correlation was originally added to NULAP5 to account for wall condensation in the steam generator tubes as described in the NULAPS manual, Section 3.2.2. Clarify how a wall condensation heat transfer correlation can be used to model interphage condensation.

Also clarify how the upper limit of 3500 Btu /hr-ft gF f or the heat transfer coefficient calculated by the ADC correlation was chosen.

The NULAP5 assessment eff orts included simulation of LOFT Tests L3-5 and L3-6. In these two tests the high pressure safety injection was directed into the upper part of the reactor vessel downcmer. The existing interphase condensation model in NULAP5 (the Northwestern University correlation) is not applicable to vertical cmponents. The original RELAP5 interphase condensation model overpredicts the interfacial energy exchange in horizontal safety injection cmponents and its applicability to flows in vertical cmponents could not be ascertained. It was for these reasons that an additional condensation model was added to NULAPS to simulate transients with atypical safety injecti on . The Akers, Deans and Crosser's correlation, even though it is mainly applicable to wall condensation heat transfer, was found to predict heat transfer coefficients that were in the range of the expected interphase condensation coefficients for the given flow rate and phasic temperature difference during safety injection. The correlation is a strong f unction of the flow Reynolds nmber which is important in predicting the condensation rate during injection. Owing to a general lack of data for these conditions, the ADC correlation was implemented f or simulation of this condensation process in the two LOFT tests since it was deemed the more appropriate. It should be noted that the licensing calculations for Haddam Neck do not apply this option in the safety injection model, instead, the Northwestern University correlation is always used.

The upper limit of 3500 BTU /hr-ft.2- F is a built-in restriction on the calculated heat transfer coefficient and is based on the approximate upper limit characteristic of the maximm range of heat transfer coefficients in the Northwestern University experiments. The NULAPS simulations of the LOFT tests L3-5 and L3-6 have shown that the code did not calculate values near this limit in the injection volme for the transients when the HPSI was activated.

b. For voltnes downstream of the ECC injection point the range of static quality where the liquid was allowed to be subcooled was changed from 0.0 - 0.5, as used in the RELAP5/M001 base code, to 0.0 - 0.999 Discuss the basis f or this change, i.e., discuss why the quality range 0.0 - 0999 was chosen over 0.0 - 0.4 or smething else. Also, discuss how far downstream of the injection point this change allows the liquid to remain subcooled.

The selection of the above static quality range was based on the f ollowing reasons:

1) A review of the Northwestern University experimental data shows that the liquid temperatures at the exit of the test channel remained in the subcooled range for all the tests, although the subcooling continually decreased fra the entrance to the channel towards the exit. A retention of the original RELAP5 static quality range f or subcooled liquid (0.0 - 0.5) in the NULAP5 model would lead to totally unrealistic calculations of the interphase condensation rate in the voltmes representing the safety injection.

A selection of a wider static quality ca igc f or liquhi subcooling is based on the ecdai reproducing the experimental data. This also insures that the liquid fra the cold legs entering the reactor vessel downcmer would still be subcooled and that the calculated condensation rate would not be overpredicted in licensing t.al c el s i ons .

2) The upper limit of static quality range equal to 0.999 was chosen to make sure that the code would calculate a subcooled liquid temperature in the injection sections of the cold legs over the range of safety injection flow rates at the Haddam Neck Plant. An examination of early RELAPS assessment results showed that the quality in those volumes rarely approaches the 0.999 limit with the safety injection actuated.

The interphase model in NULAPS licensing calculations is used only in the safety injection volines. The Haddam Neck reactor vessel the downcmer flow is verticalmodel anddoes the not use this option Northwestern since,in University this region,is correlation not applicable f or vertical flows. The subcooled liquid flowing into the downcmer is well mixed and may becme saturated if the downcmer liquid inventory is saturated and the heat addition fra the vessel wall is still high during a SBLOCA transient. The results of the Haddam Neck break spectrta analyses indicate that the lower portion of the downcmer and Ine lower plents contain saturated liquid f or the limiting SBLOCAs.

~

c. Supplement 1 stated that the dependence of'the interphase mass transfer on the temperature difference between the phases was

. changed to an equivalent quality difference. -Clarify whether this is true f or condensation calculated by both the Northwestern University model as well as that calculated by the ADC correlation.

The interphase mass transfer as a function of the temperature-equivalent quality difference applies to both the Northwestern University model and to the Akers, Deans, Crosser's model.

This is to allow for calculational consistency in the optional NULAPS interphase mass transfer models.-

f l

l t______- _ , _ _ . - __ . ._ ~ _ _ . . . _ _

3 Section 2.4 of Supplement 1 stated that a change to the original RELAPS condensation heat transfer logic was made to allow the code to remain in a condensation heat transfer mode. Previously the code would call the condensation correlation and the natural convection correlation and chose the maximum of the two coefficients. A review of the assessment runs showed the code tended to pick the natural convection coefficient most of the time. Clarify how the new logic works in NULAP5. Does the code f orce the calculation to remain in a condensation mode by only.

evaluating the condensation correlation?

The RELAP5 original heat transfer logic calls the condensation subroutine whenever the wall surface temperature is at or below the vol ume saturati on temperature. Having calculated the condensation heat transfer coefficient, the RELAPS logic would next call the natural circulation /pogi boiling correlation if the volumetric mass flux fell hel ow 200 kg/m -sec. The maximum of the two heat transfer fluxes would then be chosen by the coda.

The initial sensitivity studies with NULAPS concerning the Haddam Neck small break LOCA model have shown that the code heat transfer selection .

logic favored the natural circulation during the " pressure plateau" period of a SBLOCA transient. This period is characterized by condensation inside the steam generatgr tubes. The mass flux inside the tubes may f all bel ow 200 kg/sec-m during the plateau period.

The natural circulation logic in RELAP5 has a tendency to promote better heat transfer to the secondary which results in more safety injection into the primary side. The logic does not readily account for the presence of non-condensable gases inside the steam generator t u bes .

It was for these reasons that NULAPS heat transfer logic was modified to calculate a more conservative heat transfer rate to the secondary during the condensation period of a SBLOCA licensing calculation. The code still selects the condensation logic if the wall surface temperature is at or below the vol ume saturation temperature. However ,

the code bypasses the call to natural circulation / pool bolling correlations, following the condensation heat transfer calculations.

This minimizes the heat transfer from the primary to the secondary. In addition, as reported in Supplement 1, the logic for the condensation correlation using the Akers, Deans and Crosser's option f urther minimizes the heat transfer rate by choosing the minimtm of the two heat transfer coefficients calculated by this correlation. The heat transfer can be further degraded by specifying in the input a heat transfer degradation multiplier to account f or the presence of a non-condensable gas inside the tubes.

..m u - - -. - .- . . . ,q q , . . . ~ , -

In sumary, the new NULAP5 heat transfer logic will calculate condensation if conditions inside the steam generator tubes are appropriate f or condensation. For licensing calculations, the logic will minimize the f orward heat transfer inside the steam generators and theref ore will also minimize the amount of safety injection into the primary side since this modeling technique maximizes RCS pressure during a small break LOCA.

. 4 Changes to the critical flow models to account for stratified flow in the break volune were described in Section 2.5 of Supplement 1. When the liquid level is within or below the break, the junction void fraction is used to define a junction quality or to apportion the break flow between the Henry-Fauske and Moody models. Clarify how the junction void fraction is calculated. -Is the junction void fraction still calculated as discussed in Section 3.2.1.5 of the NULAP5 manual?

Also this section stated that when the liquid level is below the break the Moody model is used alone. Clarify whether liquid entrainment is still accounted for in this situation as discussed in Section 3.2.1.5 of the NULAPS manual .

a) The junction void fraction in a stratified critical flow is calculated as descri bed in Section 3.2.1.5 of the NULAPS manual .

The section describes how the junction void calculation is modified if the break junction elevation is specified at a location other than the pipe centerline.

4 Section 2.5 of Supplement 1 f urther describes the stratified i critical flow in NULAP5 for the case when the phases in the blowdown volune are in thermal nonequili briun. The early version of the NULAPS critical flow model (reported in the NULAP5 manual) did not account f or stratified non-equili briun critical flow, rather the model was based on an assunption that if the phases were stratified, they were also in thermal equili brium.

In NULAP5, the junction void fraction is calculated in a manner similar to that for the RELAPS void fraction calculation for stratified flow conditions. -However, in NULAP5 the break junction elevation is not restricted to the pipe centerline. For licensing analyses, to maximize liquid lost fran the system, the break is placed at the bottan of the pipe. The void fraction calculation is calculated in the following manner.

To determine the liquid level in relation to the area occupied by the break junction, the area of the liquid phase is first determined, or:

A where AF = at,FvoThe liquid void fraction (1)

A y g) = cross sectional area of the pipe Next, the areas of the pipe directly below the break ( A 2) and directly above the break ( A )'are calculated. Conparing A to A and Ay , one can determine if the level is either below or 5bove khe break . If the level is above the break, the junction vapor void fraction is set to zero (i.e. no vapor pull-through is allowed).

If the level is determined to be below the break, the j unction liquid void fraction is modified to allow for liquid entrainment or:

4

- -, --w-,g,-, r- ---% y - s - --.-

_- r y - -

"F" F,J IYGIV GLI (

, where a stratifi5d flow (RELAPS calculation)J = void fraction of liquid in the junction assum Vg = vapor phase velocity in the blowdown volume V = the limiting vapor velocity for stratified flow (Equation 164 ibtheRELAP5Usersmanual)

It should be noted that Equation 2 above is Equation 166 in the RELAP Manual .

If the level lies within the break area, the junction vapor void fraction is obtained by interpolation of the two void fractions at the boundary of the break, that is:

1 U1 - Ag fV g,g d + Ag u.- a F,K O,K N G,J * (}

G,J G

_FAAV_. FAAV where:

A A, - Ac 4

g = Ij - A'2 A3 = area below the top of the break Ap = area of the liquid phase, calculated in Equation 1 a bove A2 = area below the bottom of the break V

G,K

= Vapor velocity in the break vol une u p,g = liquid void fraction in the junction FAAV = Equation 161 in the RELAP5 manual, or the Taitel - Dukler criterion f or stratified flow The interpolation on void fraction follows the same procedure as used in RELAPS. The only difference is that the j unction elevation is now a -user option which means that the areas Ai , A an are calculated to reflect the location of the break jonct$on.d A3

, b. The statement in Section 2.5 of Supplement I which states that if the break junction is exposed to all steam, the Moody model is used alone, may have been misleading. For the Haddam Neck break j spectrun licensing analyses the postulated breaks in the hot and cold legs are always located on the bottom of the pipe (i.e. the bottom of the break coincides with the botton of the pipe in the cross-sectional view of the pipe). If there is no liquid left in the break volune, the break is, of course, exposed to steam only.

_9 In this case, the Moody.model is used alone. In Lother cases, if the break junction is 1ocated anywhere else and the level is -below the break,-liquid.entrainment is ~ accounted for as described in part

. (a) a bove .

j 1

i

II. Assessment Calculations Presented in Supplement 1

1. In the LOFT L3-5 cmparison the difference between the calculated and measured hot leg temperatures after 1300 s (see Figure 3.1-5) was attributed to the low secondary pressure calculated by NULAPS.

However, because of the good cmparison of the temperatures in the cold leg and at the core outlet after 1300 s (where the measured temperatures are cmparable to the calculated hot leg temperature, see Figures 3.1-6 and 3.1-7) it would seem that another explanation f or the difference in the hot leg temperatures is needed. One possible explanation is the hot leg thermocouple in the experiment was influenced by a hot wall and thus was not actually measuring the fluid temperature. Review the hot leg temperature comparison again and clarify the reason f or the difference between the measured and calculated temperatures. Also, sme of the instabilities in the calculated hot and cold leg temperatures were attri buted to the steam generator steam control valve. Was the steam valve in the model controlled in the same ways as the steam valve in the experiment? Also, what type of control on the steam valve would cause the cold leg temperature to initially decrease and at the same time cause the hot leg temperature to initially increase?

a) In preparing the response to this question, both the NULAP5 output of the LOFT L3-5 analysis and the Experimental Data Report (EDR) for LOFT Experiment L3-5/L3-5A (NUREG/CR-1695, November 1980) were reviewed. As a result, the f ollowing observations can be made.

The NULAPS temperatures plotted in Figures 3.1-5, 3.1-6, and 3.1-7 are consistent with the code's printed output. The trends of the temperatures at the core outlet (Fig. 3.1-7) and in the intact loop hot leg (Fig. 3.1-5) are consistent, i .e. both temperatures are approximately equal and decrease at the same rate.

A review of the similar temperatures in the EDR for LOFT L3-5 has shown that all temperatures f or f uel assemblies 1, 2, 4, 5 and 6 above the core outlet (see Figures 3M-46, -52, -57, -61, and -64 in the EDR) are consistent with one another and are comparable to the NULAP5 temperature in Figure 3.1-7. In all cases, the minimum temperature reported in the EDR or calculated in NULAPS between 0 and 2400 seconds is 495 K and 500 F, respectively.

However, the consistency in the EDR temperatures is not maintained as one checks the hot leg temperatures in the intact loop and f urther downstream near the inlet plenn of the intact steam generat or. Figure 3M-37 in the EDR (coolant temperature in the intact loop hot leg, sensor TE-P139-32-1) shows approximately 15 0K increase over the core outlet temperature at 2400 seconds. The same temperature difference is indicated in Figure 3A-30 of the EDR which shows the te'mperature recorded near the steam generator inlet

7 plenm. If the data in Figure 3M-37 of the EDR (also shown in Figure 3.1-5 of the Supplement) was influenced by too close proximity of the sensor to the pipe hot wall, then the steam generator inlet plenum temperature seems to show the same effect.

It is possible that the post-experiment data consistency checks unknowingly preserved the EDR inconsistency between the core outlet and the hot leg temperatures. The EDR does not contain any information that would explain this inconsistency. The NULAP5 calculated temperatures are consistent, however, b) The description of the calculated results in Supplement 1 indicated that the instabilities in Figures 3.1-5 and 3.1-6 are due to the steam generator steam control valve. A more detailed plotting of the calculated temperatures and the steam flow re+e on an expanded time scale f or the first 100 seconds of the transient suggests a more consistent reason f or these temperature swings. In the experiment, the reactor trip occurred 4.2 seconds bef ore the break.

The hot leg temperature in the intact 1 cop initially decreased due to the mismatch in f orward heat transfer in the steam generator (steam control valve has not closed yet). After the valve closed and bottled up the secondary, the hot leg temperature started to increase. The increase peaked at 13 seconds after the initiation of the break and then the temperature gradually started to decgease. In NULAPS, the peak temperature was calculateo to be 591 K at 13 seconds. In the EDg the recorded temperature fell to 535 K and then increased to 566 K at 40 seconds (see Figure 3S-47 in the EDR), somewhat below the initial steady state hot leg temperature of 577 K. In general, both the calculated and the recorded data show the same trend, although the peaks are somewhat delayed in time.

A similar trend is maintained in the plots of the cold leg temperatures (Figure 3S-48 in the EDR and Figure 3.1-6 in Supplement 1 - NULAP5 data). The NULAPS data shows the cold leg temperature achieving a minimtm of 550 K at 26 seconds after the break, while the recorded data shows the lowest temperature equal to 556 K at 10 seconds after the break opening.

The uncertainty in the steam operator steam control valve flow could not be determined fran the data presented in the EDR, whleh contains only one plot of valve position as a f unction of time, starting at 2000 seconds after the break initiation. The model of the valve in the NULAPS simulation of the transient which was included in the input deck of LOFT L3-5, was developed by EG8G and also used in NULAP5 benchmarking. ,

l

1

2. In the discussion on the LOFT L3-6 assessment it was stated on page 40 that the discharge coefficient during the saturated portion of the blowdown was 0.55. On page 55 the discharge coefficient was said to be 0.50. Clarify what discharge coefficient was used for the saturated blowdown in the L3-6 assessment calculation.

The discharge coefficient for the saturated critical flow model of the LOFT L3-6 analysis was set equal to 0.55 in the NULAPS input deck.

Page 44 (as opposed to page 55 identified in the above question) of the Supplement contains a misprint.

3. In the Sem: scale Test S-07-10D assessment calculation, problems that were calculated to 6cc~ur during ECC injection were alleviated by directing the accumulator flow into three cold leg volmes instead of one volme and injecting the HPSI and the LPIS into the downceer instead of the cold leg as was done in the experiment.

Clarify whether the conditions that resulted in having to model the ECC injection in this way could ever be encountered in a licensing calculati on . If they could be encountered in a licensing calculation, would an approach similar to that used in the S-07-10D assessment be used to overcome the ECC injection problems? If alternate ECC injection locations are used in licensing calculations discuss whether core heat transfer could be overestimated.

The simulation of the S-07-10D test with NULAP5, reported in Supplement 1, was perf ormed with the code changes necessary to circmvent the problems incurred when using the RELAPS interface condensation model (these changes are described in Section 2.2 of the Supplement) . During the calculation, occasional calculational problems with water packing in the injection volunes had occurred

! after the ECC system activation. With the physically small size of the Semiscale facility and the injection of a highly subcooled ECC water into steam-filled volunes, a very small time step size was required in the input deck to overcme the water packing problem.

Also, to allow NULAPS to execute with the small time steps, it was necessary to model the Semiscale test case with the safety injection flow distributed to more than one volume along the length of the cold leg and into the upper part of the downcomer.

Other code benchmarking calculations, such as the LOFT tests in which the physical sizes of the safety injection volumes more closely approximate a PWR plant, have not shown any calculational insta bilities. The Haddam Neck break spectrun analysis was characterized by very high injection rates from one HPSI system (see Figure 2.6.2.4-1 of the Haddam Neck Topical Report) into mostly voided vol unes. No instabilities were noted, even with a time step size of 0.01 seconds specified f or all the postulated break s . There was no need in the licensing calculations to use an approach similar to that used in the-S-07-10D assessment. The safety injection was modeled at the locations corresponding to'the physical location of the ECCS nozzles in the cold legs at Haddam Neck.

7 r

-1~4-4 In the discussion on the Semiscale Test S-07-10D calculation the pressure camparison was shown in Figure 3.3-5. Clarify why the calculated pressure decreased more rapidly than the measured 2

pressure from 100 to 200 s.

Inspection of the core collapsed liquid level plot in Figure 3.3-6

. of the NULAPS calculation shows a deeper core 'uncovery after 100

seconds than the recorded data. Less- energy is theref ore predicted to be transferred to the RCS and at the same time more decay heat I energy is retained in the f uel in the region exposed to steam.

l This is evident in a higher peak clad temperature predicted by i NULAPS and shown in Figure 3.3-8. Since less energy is being

transferred to the RCS, the system will depressurize at a faster j rate and with steam discharged through the break, the i depressurization will even be more effective in depressurizing the

primary side. This is evident in Figure 3.3 -5 after 100 seconds _

! when the primary pressure departs fram the pressure plateau, i

I

't 4

I I

i i

l

.~- .

3

/

[. t 4

5. Figures 3.3-7A and 3.3-7B that capare the calculated and measured ,

cold leg and . hot leg temperatures for Semiscale Test S-07-100 only' ,

, cover the period fra 0 to 100 s. Provide figures of these parameters for review with the same time scale (0 to 800 s) as the other figures provided f or Test S-07-10D.

e i The figures of the calculated and measured cold leg and hot leg temperature f or Semiscale Test S-07-10D are enclosed.

1 I

s i

i i

4 1

i 1

x k

W n

d 5

1 i

I i

i 4

I

7 FIGURE 3.3-7A SEMISCALE TEST S-07-10D COLD LEG TEMPERATURE IN BROKEN LOOP 650-600-F L e U -

.A '

I D \,

T 550- '"'s, '

E ,

M P .

E R ',

's R '.

T N, U \,

R 500- \-

E ',

s K , l.

'~. , l ,'s,

's,'"""""-

450- I,\ If

'A l

400 ,,,,,,,,,,,,,, ,,, ,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,, ,,,,,,,,,,,,,,,

0 100 200 300 400 500 600 700 800 ,

l TIME (SEC)

CURVE -------

NULAPS TEST DATA

FIGURE 3.3-7B SEMISCALE TEST S-07-10D HOT LEG TEMPERATURE IN BROKEN LOOP 650-600- ,

F L

U J 1 , .

0 '),

-el T 550- al E -

\,

tI M t P \,  !!

E ',

' ll R

A T

\

ll l

lI U

, I[ >

R 500-E -,

-  : 1

, l ;

, l K \, l l ,,

'~,, -;' ;l'~,

ll 's,

-~~

450- f 400,,,,,,,,,i.........,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,

l 0 100 200 300 400 500 600 700 800 T IME (SEC) l CURVE ------- N U L A P S TEST ORTA