ML13330B519

Concludes That Existing Administative Controls Restricting Incore Axial Offset & Operation W/Reduced Tavg Provide Sufficient Margin to Compensate for Identified Vol Differences

ML13330B519
Person / Time
Site:	San Onofre
Issue date:	05/17/1991
From:	Rosenblum R Southern California Edison Co
To:	NRC/IRM
References
NUDOCS 9105280080
Download: ML13330B519 (29)

Text

1 Southern California Edison Company 23 PARKER STREET IRVINE, CALIFORNIA 92718 R.M.ROSENBLUM May 17, 1991 TELEPHONE MANAGER OF (714) 454-4505 NUCLEAR REGULATORY AFFAIRS U. S. Nuclear Regulatory Commission Attention: Document Control Desk Washington, D.C. 20555 Gentlemen:

Subject:

Docket Nos. 50-206 Evaluation of Volume Differences in Accident Analyses San Onofre Nuclear Generating Station, Unit 1

References:

1. Letter, R. M. Rosenblum (SCE) to NRC Document Control Desk, "Underestimation of Refill Volume Assumed in Large Break LOCA Analysis," March 29, 1991.

2. Letter, R. M. Rosenblum (SCE) to NRC Document Control Desk, "Volume Differences Identified in Accident Analyses,"

May 10, 1991.

Reference 1 notified the NRC of 1) a discrepancy in the value used by Westinghouse for the refill volume in the SONGS 1 large break loss of coolant accident (LBLOCA) analysis-and 2) administrative controls established to compensate for the underestimated refill volume. Additionally, reference 2 described the preliminary results of our follow-on effort to confirm the volume information used in accident analyses performed by Westinghouse. This letter presents the final results of our evaluation of the identified volume differences and confirms the plant can be safely operated within its design basis. Westinghouse has participated in our evaluation and has documented their associated engineering analyses in the enclosure.

BACKGROUND On March 29, 1991, we informed the NRC that the refill volume used in the LBLOCA analysis was underestimated by Westinghouse by approximately 182 ft3 (total RCS volume is approximately 7200 ft3). At that time, we also established administrative controls on incore axial offset to compensate for the discrepancy. Those restrictions provided 237.51F in additional margin on the fuel peak clad temperature (PCT), more than enough to offset the 188.5oF net PCT penalty that resulted from the refill volume underestimation.

910528 0080 910517 PDR ADOCK 05000206 P

PDiR

Document Control Desk May 17, 1991 Subsequent to March 29, 1991, we extended our confirmatory efforts to assure the accuracy of the volume information used in other accident analyses performed by Westinghouse. The preliminary results of our expanded investigation were described in reference 2 and included additional volume differences that affect LBLOCA/PCT, LBLOCA containment mass and energy, and non-LOCA analyses. Although some of the volume differences may represent acceptable variations consistent with the specific modeling methodologies used for the various analyses, we are conservatively treating all of the identified volume differences as non-conservative errors. Our evaluation confirms sufficient margin exists in each of the affected accident analyses to adequately compensate for the volume differences. Our assessment for each of the affected accident analyses is discussed below.

LBLOCA/PCT ANALYSIS The 550 ft3 RCS volume difference in the LBLOCA/PCT analysis that was identified in reference 2 is comprised of 1) the 182 ft3 refill volume underestimation identified in our March 29, 1991 letter, 2) 150 ft3 in the core baffle region (the volume between the reactor vessel baffle and core barrel was not modeled), and 3) small volume differences in other areas of the RCS due to differences in calculational approach which total 168 ft3.

We have conservatively assumed these volume differences represent non-conservative errors. On that basis, Westinghouse completed.an evaluation of the effect of these volume differences on the LBLOCA/PCT analysis by considering the three phases of a LOCA, i.e., blowdown, refill, and reflood.

As discussed in the enclosure, Westinghouse performed a sensitivity analysis to determine the effect of the volume differences on the blowdown and reflood phases of the LBLOCA. Based on the sensitivity study and comparison with other plants and test results, Westinghouse concluded that no penalty is expected for the blowdown and reflood phases due to the volume differences between the IAC calculation and current methods.

The Westinghouse sensitivity study performed to assess the volume differences estimated a benefit of 122'F due to the extended blowdown. However, due to the complexity of the phenomena and the approximate nature of the sensitivity study, Edison has conservatively assessed a penalty of 211F. The 21oF penalty was calculated by assuming the worst case blowdown phase heatup rate (70*F/sec) and a 0.3 second increase in the blowdown duration. We have also assessed a penalty for the reflood phase of the transient. A penalty of 36.3,F was assessed by assuming a delay of reflood for 3 seconds (as determined by the sensitivity study) and an adiabatic heatup rate of 12.10F/second. Therefore, although Westinghouse has concluded no penalty is expected, Edison is conservatively assessing a penalty of 57.3oF.

As discussed in reference 2, our administrative controls on incore axial offset provide 237.50F of PCT margin. Additionally, reference 2 identified 460F of margin due to more realistic (but nevertheless conservative) assumptions on safety injection mini-flow. Therefore, the combined margin of

Document Control Desk May 17, 1991 283.51F associated with the incore axial offset restrictions and the safety injection mini-flow assumptions are sufficient to offset the 245.80F PCT penalty (188.5 + 57.3).

LOCA CONTAINMENT MASS AND ENERGY ANALYSIS Volume differences also appear in the Westinghouse LBLOCA mass and energy analyses. The total RCS volume used in the LBLOCA containment mass and energy design basis analyses was identified in reference 2 to be approximately 300 ft3 less than the volume utilized in the recent NOTRUMP analysis. This volume anomaly includes the previously discovered 182 ft3 in the lower plenum and other differences in various areas of the RCS. Assuming these differences represent non-conservative errors, they result in an underestimate of the RCS mass and energy released during the LBLOCA. This adversely affects the calculated peak containment pressure and temperature after a LBLOCA.

As discussed in the enclosure, Westinghouse has completed an evaluation of the impact of the 300 ft3 volume difference. The 300 ft3 volume difference translates into approximately 8 MBTU of increased energy released into the containment as a result of a LBLOCA. However, the SONGS 1, Cycle 11 design basis LBLOCA mass and energy analyses have, among other conservatisms, not credited plant operation at a reduced RCS average coolant temperature, Tavg.

Operating with a lower Tavg reduces the energy available to be released to the containment by approximately 11 MBTU. This margin is more than sufficient to offset the 8 MBTU penalty associated with the volume difference. We currently have administrative controls that maintain the lower Tavg limit and, consistent with our previous commitment, will submit by June 23, 1991, a license amendment to incorporate those administrative controls in the SONGS 1 Technical Specifications.

NON-LOCA ANALYSES We have also explored the possibility of volume differences affecting the results of non-LOCA analyses that use digital computer codes in which volume is a modeled parameter. We did not investigate the potential for volume differences in accident analyses performed with analog simulation techniques since those analyses are non-limiting, small volume differences are considered insignificant, and little documentation is available for these analyses performed in the late 1960's.

As discussed in the enclosure, Westinghouse has evaluated the effect of the identified volume differences for the following non-LOCA events: 1) rod withdrawal at power (RWAP), 2) dropped rod, 3) loss of flow, 4) locked rotor,

5) steamline break, 6) loss of normal feedwater (LONF), and 7) feedline break (FLB).

In addition, we have evaluated the impact of volume differences on the boron dilution analysis. The results of these evaluations are summarized below.

Document Control Desk May 17, 1991 LOFTRAN is the Westinghouse computer code used for non-LOCA events for which RCS volumes are modeled. Therefore, the LOFTRAN volumetric inputs were checked against equivalent volumes based on the Westinghouse plant component data base (that was used to calculate the RCS volumes for the recent NOTRUMP small break LOCA analysis). That effort uncovered a 366 ft3 difference between the total RCS volume used in the LOFTRAN analyses and that based on the component data base. Approximately 148 ft3 of the total difference is associated with 1) the dead volume in the reactor vessel head and 2) the total upper core plenum volume. The remaining volume difference is attributed to small differences in calculational approach in computing various volume inputs and has been confirmed not to be a discrepancy. The effect of the two volume differences on the limiting non-LOCA events is discussed below.

Steamline Break Analysis The effect of the identified volume differences on the LOFTRAN steamline break analysis was evaluated by re-performing the analysis with corrected volumes. The new analysis showed that the LOFTRAN volume differences resulted in 1) a 36.6 psi increase in RCS pressure, 2) a small increase in the peak heat flux from 0.403 to 0.407 (fraction of nominal), and 3) a change of -1.0F and

+0.5*F in the hottest and coldest core inlet temperatures, respectively. As discussed in the enclosure, the RCS pressure increase has a beneficial effect on DNBR and is sufficient to offset the small heat flux and temperature increases. Therefore, the smaller dead volume and upper plenum volume in the steamline break analysis of record is conservative with respect to DNBR.

The effect of the volume differences on steamline break mass and energy release was also evaluated by Westinghouse and found to be inconsequential.

Based upon the Westinghouse results, we evaluated the calculated peak containment pressure resulting from a steamline break and concluded that the volume differences had no effect on the peak containment pressure for the limiting hot full power condition and that the hot full power condition remained the limiting case.

Other Non-LOCA LOFTRAN Analyses As discussed in the enclosure, the dead volume difference is of most concern for depressurization events where the dead volume acts like a pressurizer. For events like the RWAP, LONF, and FLB, a lower RCS volume is conservative since the smaller volume would cause higher RCS pressures and less total RCS volume available before a water solid condition would be reached. In addition, a smaller initial coolant inventory would be conservative with

Document Control Desk May 17, 1991 respect to ensuring the core remains covered after the pressurizer relief valves discharge during a feedline break transient.

Westinghouse completed a new feedline break analysis to confirm these conservatisms.

For the remaining non-LOCA events analyzed with LOFTRAN, the primary concern is DNB and the parameters of importance (e.g.,

power, pressure, temperature, flow) are defined independently of the RCS volumes. Therefore, the dead volume and the upper plenum volume will not significantly affect the results of the other non LOCA events analyzed with LOFTRAN. Thus, the SONGS 1 non-LOCA analyses of record that were performed with the LOFTRAN code are either conservative or were not significantly impacted by the identified volume differences.

Boron Dilution Analysis The SONGS 1 boron dilution accident analysis was performed by SCE.

As part of the RCS volume differences evaluation, we have verified that the volumes used in that accident analysis were 1) calculated independently of the RCS volumes calculated by Westinghouse, and

2) are conservative relative to the applicable NOTRUMP RCS volumes.

CONCLUSION We have completed our evaluation of the impact of volume differences on the SONGS 1 design basis accident analyses. Our evaluation has concluded that existing administrative controls restricting incore axial offset, and operation with a reduced Tavg provide sufficient margin to compensate for all of the identified volume differences and assure the plant continues to be operated within its design basis.

Very truly yours, Enclosure cc: George Kalman, NRC Project Manager, San Onofre Unit 1 J. B. Martin, Regional Administrator, NRC Region V C. Caldwell, NRC Senior Resident Inspector, San Onofre Units 1, 2 and 3 C. D. Townsend, NRC Resident Inspector, San Onofre Unit 1

MA 17 '91 18:44 FROM LICENSING TO SPE IRUINE PAGE.002 Westinghouse Energy Systems Energy CenTef Electric Corporation Business Unit P

P 35 3

CL Cas General Manager Nuclear & Advanced Technology Division NS-OPLS-OPL-1-91-286 Mr. J. T. Reilly Manager Nuclear Engineering & Construction Southern California Edison Company Irvine Operations Center 23 Parker Street IrvineW CA 92718 SOUTHERN CALIFORNIA EDISON COMPANY SAN ONOFRE NUCLEAR GENERATING STATION UNIT 1 Safety Analysis Volume Differences Ref. 1) SCE-91-528, "Large Break LOCA Lower Plenum Volume Discrepancy",

Letter From H. C. Calton (Westinghouse) to J. T. Reilly (Southern California Edison Company), Dated March 28, 1991.

Ref. 2) SCE-91-533, "Integrated Assessment for Refill Volume Discrepancy",

Letter From S. A. Pujadas (Westinghouse) to B. Carlisle (Southern California Edison Company), Dated April 24, 1991.

Ref. 3) SCE-91-550, "Assessment of Safety Analysis Volume Differences",

Letter from C. Caso (Westinghouse) to J. T. Reilly (Southern California Edison Company), Dated May 9, 1991.

Dear Mr. Reilly:

Attached is the Westinghouse assessmentof the volume differences found in the safety analyses for the San Onofre Nuclear Generating Station Unit I (SONGS-1). The differences result from a comparison of the volume inputs used in the existing safety analyses, and the volumes which would be calculated using current analysis methods and standards. Previous discussions on this subject were provided in the references. The attached provides more detailed information and an assessment of the effect of these volume differences on the safety analyses within Westinghouse scope. In addition, Westinghouse will be providing the results of a large break LOCA reactor coolant system volume sensitivity as a separate transmittal.

If you have any questions or comments, please contact me or Mr. Michael Young of the Nuclear Safety Department.

Very truly yours, DPD/jmb Attachment

MA ' 17 '91 18:45 FROj ICENSING TO E

IRUINE PAGE.003 J.

May 16, 1991 SCE-91-551 cc: M. Wharton 1L, LA T. Yackle 1L, 1A A. J. Eckhart IL, LA D. Frey 1L, 1A R. M. Rosenblum IL, 1A L. K. Carlisle 1L, 1A B. Craig IL, LA M. Stickel 1L, 1A

MA1Y 17 '91 18:45 FROM LICENSING TO SE IRUINE PAGE.004 SAN ONOFRE NUCLEAR GENERATING STATION UNIT 1 (SONGS-1)

ASSESSMENT OF SAFETY ANALYSIS VOLUME DIFFERENCES INITRODUCLON Recently a difference was discovered between the refill volume used in the Westinghouse IAC (Interim Acceptance Criteria) large break loss-of-coolant accident (LOCA) analysis for the San Onofre Nuclear Generating Station Unit 1 (SONGS-1) and the volume which would be calculated using current analysis methods and processes.

The significance of the refill volume to the large break LOCA analysis results prompted a safety assessment of the effect of the difference.

As a result, constraints on the axial offset limits were recommended to provide reasonable assurance of safe operation within the limits of the Interim Acceptance Criteria.

The purpose of this communication is to provide Southern California Edison with the findings of the Westinghouse evaluation of the safety analysis volume differences resulting from the comparison of the current small break LOCA, and the various safety analyses within Westinghouse scope of supply. It was noted in References I and 2 that only the large break LOCA analysis was significantly affected by the refill volume difference. Since that time, Westinghouse has continued to work with SCE to determine the significance of the differences and to review other SONGS-1 accident analyses within Westinghouse scope of supply to determine what other differences exist between the reactor coolant system (RCS) volumes in the existing safety analyses and the volumes as they would be calculated by current methods and practices.

Page 1 SCEi;5/9/91:MYY

MAY 17 '91 18:46 FROM LICENSING TO crE IRUINE PAGE.005 BACKGROUND On March 26, 1991, a discrepancy was discovered between the SONGS-1 refill volume assumed in the Westinghouse IAC (Interim Acceptance Criteria) large break LOCA analysis and the recently completed small break LOCA analysis using the NOTRUMP computer code.

The original input calculations for the SONGS-1 large break LOCA analysis were performed in 1970. As part of the.NOTRUMP small break LOCA analysis performed to support increases in safety injection (SI)

miniflow, the SONGS-1 plant data was collected and entered into the Westinghouse plant component data base used in the LOCA analysis process. This data base automatically generates the volume inputs for the LOCA analysis from the basic geometric plant information.

Generation of the small break LOCA analysis volumes indicated a lower'plenum volume of 611 ft3 to the bottom of the active fuel inside the core barrel.

An additional 61 ft3 represents the downcomer volume from the bottom of the core barrel to the bottom of the active fuel elevation.

Together, this 672 ft3 volume comprises what can be termed "refill volume" with respect to the IAC large break evaluation model (EM).

Review of the SONGS-I vessel diagrams indicates that, using the current methods and processes, the refill volume used in the large break LOCA analysis was underestimated.

Use of the refill volume as calculated using current methods and practices will result in an increase in calculated Peak Cladding Temperature (PCT). A review of available documentation has determined that the inputs used in the currrent SONGS-1 analysis were calculated by Westinghouse in

1971, and submitted to SCE for review and approval (Reference 3). However, the review was not successful in determining the basis for the original calculation of the lower plenum volume.

It is possible that, due to the unique safety injection system (no accumulators), credit was taken for some of the initial inventory of primary water remaining in the vessel during blowdown, collecting in the lower plenum, and reducing the volume which had to be refilled prior to core recovery. Recent data from such tests as LOFT and calculations with advanced codes such as TRAC also indicate that the amount of original primary water retained may be significant.

A report written shortly before the original SONGS-1 analysis (Reference 4) indicated that calculations performed at the time predicted a significant amount of residual liquid volume left In Page 2 SCE1:5/9/91:MYY

M9Y 17 '91 18:47 FROM LICENSING TO E IRUINE PAGE.008 the reactor vessel.

The reference shows that approximately 28 percent of the lower plenum volume contains liquid, a percentage which is similar to the identified refill volume difference.

Based on this information, a smaller effective volume may have been assumed for the SONGS-1 analysis, to take credit for this accumulated water.

However, documentation indicating that this assumption was in fact made could not be found, and it is explicitly stated in the current FSAR for SONGS-1 that the lower plenum is assumed empty at the end of blowdown (Section 15.16.2.1.2).

As a result of this conflicting information, and in view of the importance of the refill volume to the results of the SONGS-1 analysis, it was concluded that the lower plenum volume used for the refill period should be made consistent with current methods and standards. The lack of documentation in the original SONGS-1 analysis, which would not meet the current standards, was determined to result from inadequate design documentation procedures for application of the Interim Acceptance Criteria LOCA codes in the late 1960's. Current practices would require documentation and independent review of the calculations.

Since the new regulations in 10CFR50.46 and Appendix K exempted PWR's which used stainless steel clad fuel, the evaluation model developed for the Interim Acceptance Criteria continued to be used for SONGS-1.

LOCA analyses for SONGS-1 were repeated in 1980 and 1987 using the Westinghouse Interim Acceptance Criteria ECCS Evaluation Model.

The calculations were performed using the original input values, except for the inputs which were affected by the plant change for which the analysis was required (e.g., increased tube plugging resulted in modification of the steam generator tube flow area and volume).

The input data unaffected by the plant change was assumed to be valid.

While the lack of documentation for these inputs was recognized, there was no reason to believe the inputs to be incorrect.

Page 3 SCE1:5/9/91:MYY

MAY 17 '91 18:48 FRjLICENSING TOCE IRUINE PAGE.007 The approach used in the 1980 and 1987 analyses was consistent with the approach used by Westinghouse for other plant analyses. However, for all other plants, the prior input has been documented and independently verified.

Current quality assurance procedures (Reference 5) require independent verification of the analysis inputs and assumptions, with an identification of those which must be verified. This procedure will be clarified and training will be provided to assure that assumptions concerning the validity of prior input are verified in future analyses.

IMPACT OF VOLUME DISCREPANCIES ON ACCIDENT ANALYSES To determine if other potentially significant differences exist between the volumes used in the safety analysis and the volumes which would be calculated using current methods and practices, a review of other safety analysis volume inputs was performed. These differences were then evaluated to determine whether they would have a significant impact on the safety analysis results.

I. Large Break LOCA:

The current licensing basis large break LOCA analysis assumes a refill volume of 490 ft3, which is underestimated by 182 ft3 when current input calculation methods and processes are used. A comparison of the total volume of the RCS, as originally calculated, with more recent, verified values, indicates an additional difference of approximately 370 cubic feet, for a total difference of approximately 8.6 percent of the total volume. A large fraction of the additional difference (approximately 150 cubic feet) appears to result from exclusion from the analysis of the region between the core barrel and the core baffle plates.

Page 4 SCE :S/9/91:MYY

MPY 17

'91 18:49 FROM LICENSING TO 1CE IRUINE PAGE.008 Westinghouse has concluded that given the simplicity of the Interim Acceptance Criteria LOCA models, and the level of knowledge of the LOCA at the time, it would have been reasonable to ignore the barrel/baffle region, which is connected to the other regions in the reactor vessel by high resistance flowpaths. Other differences appear to result from differences in calculational approach.

Due to the complexity of the reactor vessel internals, several geometrical approximations must be made which could result in different calculated volumes. As a result, it has been concluded that these other differences should not be considered errors.

Assessment of Overall Validity of the SONGS-1 LOCA Analysis As noted above, there are other differences between the volumes used for the Interim Acceptance Criteria large break LOCA analysis and the volumes which would be calculated using current methods and practices. While some of these differences appear to be the result of different assumptions used, rather than errors, an assessment of the impact of these differences on the overall LOCA results was performed.

The approach taken is to examine each phase of the transient as currently calculated, and to compare key results with verified calculations, or with test data.

The overall validity of the calculated results can then be determined based on this review. The three phases to be examined are the blowdown phase, the refill phase, and the reflood phase. In the blowdown phase, the key result is the hot rod temperature at end of blowdown, and the end of blowdown time.

During the refill phase, the key result is the cladding temperature at the end of refill.

During the reflood phase, the key result is the cladding temperature rise and turnaround time.

Page 5 SCE1:5/9/91:MYY

MPY 17 '91 18:50 FRO LICENSING TO 1CE IRUINE PAGE.009 A comparison of the total volume of the RCS, as originally calculated, with more recent, verified values, indicates an overall difference of approximately 550 cubic feet, or approximately 8.6 percent of the total volume. Differences exist between individual volumes which may reflect differences in calculational or modeling assumptions. To determine whether the differences in volume could have led to significant differences in the calculated results during blowdown, conditions near the end of blowdown for SONGS-1 were compared to the results for another plant, which also used the same evaluation model and whose peak linear heat rate was similar.

Although this plant is larger, with four loops, it exhibits similar overall behavior because it does not have accumulators. This calculation was used to determine if a larger vessel volume significantly affected blowdown results.

The geometric input for this plant was independently calculated by the utility in 1985, and again calculated by Westinghouse in 1989. A check of the input for this other plant confirmed that they are consistent with values.obtained using current methods and standards. The results near the end of the blowdown calculation are shown below:

SONGS-1 OTHER PLANT CD=0.8 CD=1.0 Transient Time(sec) approx.

20 20 PCT at Transient Time 1460 1446 The end of blowdown time for the SONGS-1 analysis is approximately 17 seconds while the end of blowdown time for the four loop plant is approximately 20 seconds. Since SONGS-1 is a smaller plant, it would be expected to blowdown more quickly and the earlier end of blowdown is expected. Consequently, the blowdown transient for the reference four loop plant which is most similar to SONGS-1 has a larger break discharge coefficient. The blowdown transient for SONGS-1 may have also been shortened by the reduced RCS volume. If a point at 20 seconds in the SONGS-1 transient is examined, it is seen that the cladding temperature is similar to the cladding temperature in the four loop plant. Thus the volume discrepancy is judged not to have had a strong effect on the calculated cladding temperature during blowdown.

Page 6 SCE1:5/9/91:MYY

MAY 17 '91 18:51 FROM LICENSING TO SCE IRUINE PAGE.010 The portion of the LOCA transient which dominates the SONGS-1 calculation is the refill transient.

A temperature rise of over 800*F is calculated during this period. The critical parameters for this calculation are the refill volume, the safety injection flow, and the adiabatic heatup rate. Assuming the lower plenum below the active fuel is empty at the end of blowdown, the refill volume has been confirmed to be equal to 672 cubic feet.

A second critical aspect of the refill calculation is the adiabatic heatup rate.

This heatup rate is dependent on the fuel and cladding heat capacities and the decay heat level.

These quantities were checked for the SONGS-1 calculation and found to be consistent with values calculated with current standards and methods.

It is concluded that the cladding heatup during the refill period which was calculated for SONGS-1 is correct, when the revised refill volume is used.

The reflood transient begins with a flooding rate of approximately 1.8 inches per second, prior to the onset of significant entrainment.

This valUe represents the rate at which the core, barrel/baffle region, and downcomer volumes fill, with the given safety injection flow. Following the onset of entrainment, the flooding rate is approximately one inch per second, typical of the values calculated for other plants.

Because SONGS-1 uses stainless steel cladding, the FLECHT tests can be used directly to establish the validity of the reflood calculation. The cladding temperature rise during reflood is about 70 degrees F, prior to peak cladding temperature turnaround.

This temperature rise is reasonable because the cladding is already at an elevated temperature, and the fuel rod power is low when reflood begins in the SONGS-1 calculation. FLECHT tests performed under similar conditions support this conclusion. These tests indicate a substantial reduction in temperature rise with increasing initial cladding temperature, and with decreasing initial peak linear power. For conditions approximating those found in the SONGS-1 analysis, the measured temperature rise from these tests is consistent with the value calculated.

Page 7 SCE1:S/9/91:MYY

MPY 17 '91 18:52 FR LICENSING TO E IRUINE PAGE.011 These analyses provide some assurance that the SONGS-1 large break LOCA Interim Acceptance Criteria evaluation model produces results which are reasonable and conservative, and in compliance with requirements set forth in the IAC.

A review was performed of model conservatisms and past plant operation to determine the effect of the volume discrepancies on past operation.

Actual Plant Peaking Factor, Fa The effect of an increase in the refill volume is to prolong the refill time between the end of blowdown and the beginning of reflood in the SONGS-1 LOCA analysis.

To maintain the calculated results within the Interim Acceptance Criteria limit of 23000F, the maximum allowed peak linear heat rate was reduced from 13.2 kw/ft to 11.3 kw/ft, which corresponds to an allowable total peaking factor of 2.38.. This had been done by applying a more restrictive Axial Offset (AO) band of + 15%

at full power. Applying the + 15% AO 'criteria to past operation (cycles 8 through 10) results in a maximum Fq during that period of 2.52, which provides Fq margin of 0.26, to the current analyzed value of 2.78.

This is equivalent to approximately 154 0F of PCT margin. It is Westinghouse understanding that a review by SCE of'the available plant data for cycles 8 through 10 verified that the actual recorded AO was maintained within this power dependent AO band during normal operation and all power maneuvers.

For cycles 8 through 10, the data relating the maximum Fq*P to axial offset was obtained from flyspeck plots available in the respective cycle's Nuclear Design Report or Reload Safety Evaluation. For cycles 1 through 4, the flyspeck plots do not exist in these documents because Fq was evaluated in a different manner.

Prior to cycle 5, the rodded and unrodded radial power distributions (Fxy) were combined with axial power distributions (Fz) which were calculated at the rod insertion limits. At that time, Technical Specification 3.11 did not

exist, and core power distribution control was maintained through compliance with the rod insertion limits alone. Therefore, the data necessary to evaluate the core at + 15% AO does not exist for these cycles, and cannot be generated without doing extensive load follow simulation analysis.

Page 8 SCE1:5/9/91 :MYY

MAY 17 '91 18:53 FR LICENSING TO E IRUINE PAGE.012 Calculated design peaking factors for the flyspeck plots are increased by very large conservatisms for San Onofre relative to other plants.

Current measurement and engineering uncertainties applied to calculated Fq's are 10.6%

and 4%, respectively, compared to 5% and 3% for most other Westinghouse plants.

The measurement uncertainty is-larger due to the design of the incore instrumentation system, and not due to any additional calculation uncertainty in the design models. In addition, the calculated peaks include an allowance for radial xenon oscillations (3%).

All of these uncertainty factors are conservatively assumed to apply at the worst case core location at all times in life.

Conservatism in Analysis Models Even if SONGS-1 had operated at its Tech Spec limit of 13.2 kw/ft rather than 11.3 kw/ft, it can be demonstrated that there is reasonable assurance that the plant did not operate in a manner which posed a threat to the safety of the nuclear power plant or to the public health and safety.

This may be demonstrated as follows:

The decay heat model used in the LOCA analysis is based on a preliminary ANS standard issued in 1971.

The more recent 1979 standard specifies decay heat levels (at 95% confidence levels) approximately 21 percent lower than the 1971 standard. This translates directly to reduced peak linear heat rate during the blowdown, refill, and reflood periods. A conservative estimate of the potential reduction in PCT from this source of margin can be made by applying a 21 percent lower adiabatic heatup rate during the refill period.

This approach does not take credit for reductions during blowdown or reflood.

The refill time consists of delays in initiating pumped injection, a delay in delivery to the vessel, and the time required to fill the lower plenum below the active fuel. These delays were previously calculated for the increased safety injection miniflow evaluation-provided in Reference 6.

The time to refill the vessel to the bottom of the fuel (refill volume after the end of blowdown (EOB) may be calculated as follows:

Page 9 SCE1:5/9/91:MYY

MtY 17

'91 19:55 FRON' LICENSING TO E IRUINE PAGE.013 Refill Time = injection delay after EOB + refill volume fill time

='22.4 seconds + (490 ft3)(62.3 lb/ft3)/(6551b/sec) 69 sec DELTA PCT (refill) refill time x adiabatic heatup rate (69 seconds)

(12.1OF/sec) 835'F This calculation assumed a safety injection miniflow value of 1712 gpm or less.

The heatup rate is the average for the refill period, assuming a peak linear heat rate of 13.2 kw/ft.

Since the 1979 decay heat is approximately 21% lower, the adiabatic heatup rate would be approximately (12.1OF/sec)(0.79) n 9.56'F/sec.

Using the more accurate 1979 ANS decay heat standard at 95% upper bound confidence limits with the revised refill volume would result in an adiabatic cladding heatup as follows:

Refill Time = 22.4 second + (672 ft3)(62.3 lb/ft3)/(655 lb/sec)

= 86.3 seconds DELTA PCT (refill) = (86.3 second)(12.1*F)(.79) w 825'F This is 100F lower than the previous value, and would result in a lower PCT during reflood.

Thus taking credit for a more realistic value of decay heat would have resulted in an acceptable calculated PCT,. even if SONGS-1 had operated at the previous Technical Specification limit.

The above calculation does not take into account the effect of decay heat on blowdown, nor does it account for several other conservative assumptions in the evaluation model, some of which are listed below:

Page 10 SCE1:5/9/91:MYY

MAY 17 '91 18:55 FRO LICENSING TO CE IRUINE PAGE.014 No rewet during blowdown.

Rewet during blowdown would significantly reduce peak cladding temperatures at the end of blowdown. Evidence of this phenomenon has been found in several experiments performed during the 1970's and 1980'.s.

Lower plenum empty at end of blowdown. Accumulation of water from the initial inventory of primary coolant in the lower plenum of the reactor would reduce the volume which had to be refilled prior to core recovery. Again, tests performed on large scale integral tests such as LOFT indicated that this water accumulation could be significant.

Adiabatic heatup during refill.

The prolonged refill period would allow ample time for radiation to steam and lower power fuel rods to reduce the total hot rod heatup.

This phenomena is included in current Final Acceptance Criteria (FAC) models.

The margins described above are among several that are available through the use of advanced, Best Estimate LOCA methods. Westinghouse has successfully applied these methods to other plants (reference 7), allowing for significant increases in allowable peaking factor.

These results are supported by NRC sponsored studies (reference 8),

which indicate potential margins of approximately 500*F, compared to older evaluation models.

Conclusion It is Westinghouse judgement that operation of SONGS-1 did not pose a threat to the safety of the nuclear power plant or to the public health and safety as a result of the refill volume discrepancy. In addition, Westinghouse has concluded that, if the effect of the lower plenum refill volume is accounted for, the existing SONGS-1 analysis remains acceptable. These conclusions are based upon the following:

1) The plant specific measurements for axial offset indicate that the total peaking factor would have maintained compliance with the Interim Acceptance Criteria limit of 2300'F Page 11 SCEI:5/9/91:MYY

Mf1Y 17 '91 18:56 FROK LICENSING TO E IRUINE PAGE.015

2) Reduction of a more realistic, yet conservative decay heat model in the ECCS analysis would demonstrate compliance with the Interim Acceptance Criteria limit of 2300*F, even if SONGS-1 had operated continuously at past Technical Specification limits.

3) Comparison of the SONGS-1 analysis results to the results of calculations performed for another plant which used the Westinghouse Interim Acceptance Criteria ECCS Evaluation Model indicate consistent results during the blowdown phase.

4) Comparison of the analysis results to applicable full length emergency core cooling heat transfer (FLECHT) tests indicate consistent results during the reflood phase.

5) A sensitivity study for the effect of volume increases on the IAC ECCS Evaluation Model large break LOCA analysis was performed. Increasing only the volumes in the IAC large break LOCA analysis to correspond to valus which would be calculated using current methods and processes results in no adverse effect on the peak cladding temperature resulting from blowdown.

Increasing the refill volume in the IAC Refill calculation.results in a substantial penalty on the peak cladding temperature, as discussed above.

Increasing the volume related inputs to the IAC Reflood calculation results in only a small effect on peak cladding temperature calculation. Since it appears that the barrel baffle volume is not modeled in the IAC reflood methodology, the volume discrepancies resulting from a comparison to current methods and processes do not affect the SONGS-1 IAC peak cladding temperature calculation during reflood.

Based on the volume sensitivity study, and the comparison with test results and the LOCA calculation for another plant, it has been shown that no penalty is expected in blowdown or reflood due to the volume differences between the IAC calculation and current methods.

However, a penalty may be assigned to the blowdown and reflood portions of the transient for conservatism. SCE has calculated penalties of 21oF for blowdown and 36oF for reflood, using the differences in the blowdown and reflood times and assuming conservative heatup Page 12 SCF1:5/9/91:MYY

MAY 17 '91 18:57 FROM LICENSING TO E IRUINE PAGE.01S rates from the volume sensitivity study.

Consequently, a total penalty of 60oF may be assigned to the large break LOCA analysis for the blowdown and reflood period.

Based upon the volume sensitivity study, assigning a PCT penalty of 600F provides reasonable assurance of safe operation of SONGS-1.

II NON-LOCA A review of the current SONGS-1 non-LOCA safety analyses (which use the non-LOCA safety analysis LOFTRAN code) shows that all the RCS volumes currently in use stem from 1980 RCS volume calculations performed in support of the SONGS-1 Cycle 8 restart. These calculations were performed using dimensional and geometrical configurations as specified by appropriate Westinghouse drawings for the various SONGS-1 RCS components.

A review of the calculations was performed and a comparison of the resulting volumes used in the LOFTRAN computer program was made with the equivalent volumes generated using available volumes contained in the plant component data base used in the LOCA analysis process (i.e., IMP data).

Based on this review, it was determined that on an.equivalent basis (i.e., cold volumes), a difference of 365.8 ft3 in total RCS volume exists between that used, in the LOFTRAN analyses (LOFTRAN being lower) and that determined using the available volumes contained in the LOCA analysis plant component data base. This represents a 5.0% difference in the total RCS volume.

The review further indicated that of this 365.8 ft3 difference, 245.7 ft3 (or better than 2/3 of the total difference) is contained in two volumes; the dead volume in the reactor vessel head and the total upper core plenum volume, and that the calculations of these volumes contained computational errors that resulted in the lower values for thesevolumes. The total computational error of these two volumes is 147.6 ft3.

The corrected values for these volumes reduce the total difference to 178.1 ft3 or to within 3.0 % of the total RCS volume on an equivalent basis. Of this 178.1 ft3 difference, 98.1 ft3 is from the calculated differences of the dead and upper core plenum volumes.

Page 13 SCE1:5/9/91:MYY

MAY 17 '91 18:58 FRO LICENSING TORE IRUINE PAGE.017 Based on the corrected LOFTRAN RCS volumes, it is concluded that the volumes computed using the dimensional and geometrical configurations as specified by the appropriate Westinghouse drawings for the various SONGS-1 RCS components are essentially equivalent to those determined using available volumes contained in the plant component data base used in the LOCA analysis process.

The small differences in volumes that do exist are the result of differences in the computational methods employed.

The information contained in the plant component data base used in the LOCA analysis process are conservatively calculated with respect to the LOCA results.

In some instances, the volume calculation process conservatively (with respect to LOCA) overlooks small details which could lead to a more precise volume calculation.

With respect to the computation errors discovered in the LOFTRAN RCS volume calculations for the dead volume in the reactor vessel head and the total upper core plenum volume, it is concluded that for the safety analyses which are most concerned with these volumes (steamline break),

the use of smaller volumes is either conservative or does not significantly-affect the results.

For steamline break, the volume of particular concern is the dead volume since most of the calculational error is associated with this volume.

For steamline break, the dead volume acts like a pressurizer. When the RCS pressure drops below the saturation pressure of the dead volume, the water in the dead volume starts flashing to steam. The flashing tends to keep the primary pressure up.

Thus, the largest effect of assuming a smaller dead volume is a lower primary pressure, and, with all else unchanged, a lower primary pressure is conservative with respect to the calculated minimum DNSR.

Page 14 SCe1 :S/9/91:MYY

MAY 17 '91 19:59 FR LICENSING TORE IRUINE PAGE.018 However, the increase in primary pressure resulting from an increased dead volume may slightly reduce the corresponding safety injection flow rate (which is injected as a function of RCS pressure).

To quantify this effect, steamline break analyses with the corrected dead volume and upper plenum volume have been performed.

For the most limiting core response (i.e., DNB) SLB case, at the time of peak heat flux the effect of increased volumes is 1) a 36.6 psi increase in RCS pressure, 2) an increase in the peak heat flux from 0.403 to 0.407 (fraction of nominal), and 3) a change of -1.0 'F and +0.5 "F in the hottest and coldest core inlet temperatures, respectively (a temperature gradient 'across the core inlet exists due to the steamline break in one of the three loops).

For the applicable W-3 DNB correlation, the increase in RCS pressure is beneficial and is of sufficient magnitude to offset the small increase in peak heat flux and identified changes in inlet temperatures.

Thus, the use of the smaller dead volume and upper plenum volume in the current licensing basis Core Response SLB analysis is conservative and the results and conclusions of the existing licensing basis analysis remains valid.

For SLB mass and energy (M&E) releases, two inside containment SLB cases were analyzed to assess the effect of the changes in the dead volume and upper plenum volume. These two cases reflect the current licensing basis SLB M&E analysis for SONGS-1 (i.e., HFP and HZP initial conditions and assuming 1500 ppm boron in the SI lines and considering SI flow diversion flow).

The results of the SLB M&E analyses with the revised volumes are summarized in Table 1 in comparison to the current licensing basis Inside Containment SLB M&E analysis.

The results indicate that the changes in the dead volume and upper plenum volume do not have a significant effect on the transient results.

However, since SCE is responsible for the containment integrity portion of the analysis, this information is provided for SCE's use in making the final assessment with respect to peak containment pressure.

Page 15

$CEl:519/91:MYY

4MPY 17 '91 19:00 FROM LICENSING TO *E IRUINE PAGE.019 With. respect.to other non-LOCA events analyzed using LOFTRAN, in general, the LOFTRAN results are not very sensitive to small changes in RCS volumes (in either direction) since RCS volumes are defined independently from the more pertinent parameters such as power, pressure, temperature, flow, steam generator secondary side conditions, heat transfer coefficients, etc.

These pertinent parameters are the driving forces for the transients, and thus determine the results.

For the current SONGS-1 non-LOCA licensing basis, the events analyzed by Westinghouse using LOFTRAN include; 1) Rod Withdrawal at Power, 2) Dropped Rod, 3) Loss of Flow, 4) Locked Rotor, 5) Steamline Break (addressed above), 6) Loss of Normal Feedwater (including Loss of Offsite Power), and

7) Feedline Break.

The Rod Withdrawal at Power (RWAP) and Loss of Normal Feedwater (LONF) events are analyzed to demonstrate that a water solid condition in the pressurizer is precluded. Precluding pressurizer filling and water relief from the RCS assures that the core remains covered which is the acceptance criterion applied in the analysis of the Feedline Break (FLB) event. In each of these events, the RCS pressure will increase (after an initial decrease in the FLB event) and this increase is evaluated below to ensure RCS integrity is maintained.

As indicated earlier for steamline break, the subject discrepancy in the dead volume is mostly a concern for depressurization events where the dead volume acts like a pressurizer. For events like the RWAP, LONF, and FLB where RCS pressure and RCS coolant volume ultimately increase, modeling a lower RCS volume is inherently conservative since it will lead to higher RCS pressures and less total RCS volume available before a water solid pressurizer condition would be reached. Likewise, less initial coolant mass exists in the RCS which is conservative with respect to ensuring that the core remains covered in the event of any water relief from the RCS through the pressurizer pressure relief valves (e.g., FLB). To confirm the latter, an analysis of the FL8 event with the correct volumes was performed.

The results confirmed that less initial coolant mass is conservative with respect to core uncovery concerns. In fact, the Page 16 scel:5/9/91 :MY

, MPY 17 '91 19:01 FROM LICENSING TOCE IRUINE PAGE.020 increase in the dead volume and upper plenum volume results in more margin to core uncovery in the FLB event. Hence, the existing licensing basis analysis for the FL8 event is conservative and remains valid.

For the remaining SONGS-1 licensing basis events analyzed by Westinghouse using LOFTRAN (excluding SLB which was addressed earlier), the primary concern is DNB.

For these type events, the parameters of importance (eg., power, pressure, temperature, flow) are defined independently from the RCS volumes and will not change as the result of a change in the defined RCS volumes. Hence, the subject discrepancies in the dead volume and the upper plenum volume will not have any significant effect on the results of these events.

Thus, the SONGS-1 non-LOCA safety analyses performed using LOFTRAN are either conservative or will not be significantly affected by the identified computational errors in the dead volume and upper plenum volume.

Therefore, the corresponding SONGS-1 FSAR licensing basis for these events remain valid (pending SCE verification of containment integrity for SLB M&E).

All other SONGS-1 licensing basis non-LOCA events analyzed by Westinghouse are not affected by the subject volume discrepancies since they are either not analyzed by LOFTRAN (i.e.,

Rod Withdrawal from Subcritical, Rod Ejection) or are events analyzed using analog computer simulation techniques which pre-date the use of LOFTRAN on SONGS-1 (i.e., Startup of an Inactive Coolant Loop, Excessive Feedwater Flow, Excessive Load Increase, Loss of Load/Turbine Trip).

The RWFS and Rod Ejection events are both analyzed using the digital computer codes TWINKLE, FACTRAN, and THINC (RWFS only). These codes do not model RCS volumes.

The remaining events which were analyzed using analog computer simulation techniques are from the original SONGS-1 FERSA licensing basis (1966 vintage).

Due to the vintage of these analyses, available documentation is limited to that included in the FERSA and updated FSAR. While it is Page 17 SCE1:5/9/97:MYY

MAY 17 '91 19:03 FR LICENSING TORE IRUINE PAGE.021 impossible to confirm the appropriateness of the RCS volumes use in these analog simulations, there is no reason to suspect that these analyses are incorrect.

Furthermore, these events are all Condition II transients which typically do not result in transient conditions more limiting than the more severe SONGS-1 non-LOCA transient analyzed using LOFTRAN.

All other SONGS-1 non-LOCA events not specifically mentioned within are outside the scope of Westinghouse. This includes the.Boron Dilution event which is one event (not analyzed using LOFTRAN) which is most critically affected by changes in RCS volume.

III CONTAINMENT MASS AND ENERGY RELEASE The RCS volumetric differences between the current licensing basis LOCA Mass and Energy release model and the small break LOCA NOTRUMP analysis, which is based upon the latest RCS,system geometric data, were compared.

The major difference was determined to be-in the lower plenum region, which would potentially affect the total energy available for release from the RCS, and could also affect the time of refill.

For mass and energy release analysis, the conservative assumption that the bottom of core recovery occurs immediately after blowdown is made.

Therefore, the refill concern can be eliminated because the refill period is conservatively neglected in the LOCA mass and energy release calculations.

When the volumetric data of the whole RCS system was compared, it was determined that the latest volumetric data results in approximately 300 ft3 of additional RCS volume, which translates into additional energy releases (less than 8 million BTUs).

However, the original analysis utilized an average RCS temperature of 580 OF, as opposed to the actual operating Tav of 551.5 OF.

This reduced Tav translates into approximately 11 million extra BTUs of energy in the current licensing basis analysis.

This surplus of energy more than offsets the energy associated with the additional 300 ft3 of RCS inventory. Additionally, a conservatively high pressurizer water volume was used in the current Page 18 SCE1:5/9/91:MY Y.

MAY 17 '91 19:04 FR LICENSING T

E IRUINE PAGE.022 release calculations.

When the pressurizer volume based upon the current 100%

power data is considered it is seen that the current analysis used a conservatively higher pressurizer volume by approximately 295 ft3.

Additionally, the current release calculations assume 0% tube plugging.

At SONGS-1, approximately 15%

of the tubes are plugged, which would further reduce the volume discrepancy. Therefore, in summary, from an RCS initial mass.and energy content, the current releases remain bounding.

In summary, the LOCA mass and energy release evaluation was performed with realistic, but still conservative geometric data, and assumptions consistent with the current plant operating conditions. The results of the evaluation demonstrate that the current licensing basis LOCA Mass and Energy release analysis results remain bounding when the analysis conservatisms are compared to the minor perturbation associated with the RCS volume differences.

All applicable safety criteria for the containment mass and energy release analysis continue to be met.

CONCLUSIONS In addition to the LOCA refill volume, an error in the Non-LOCA upper vessel region volumes was found.

This error was determined to not be significant.

Additional differences have been observed between the volumes used in the SONGS-1 safety analyses and the volumes which would be calculated using current methods and practices. The remaining differences result from different methods which have been used to calculate the volumes, and assumptions made in the safety analysis calculations. For example, the volumes in the large break LOCA analyses are typically increased above the values which would be calculated by cold dimensions to account for thermal expansion and because larger volumes in general are conservative for the LOCA analysis calculations. Differing assumptions and methods may result in differences in the reactor coolant system volumes for the various safety analyses.

This does not mean that the safety analyses are in error. It is Westinghouse judgement that with the exception of the LOCA refill volume, the differences between the safety analysis values and the values which would be calculated using the current methods and processes in the LOCA input generation primarily reflect different calculational approaches and are not significant. It is Page 19 sCa1:5/9/91;MYY

MAY 17 '91 19:05 FR LICENSING T

E IRUINE PAGE.023 4

'k 40 Westinghouse judgement that the difference in the refill volume in the Interim Acceptance Criteria large break LOCA analysis is the only significant difference in the safety analysis volume inputs for which a correction is necessary.

REFERENCES

1. SCE-91-528, "Large Break LOCA Lower Plenum Volume Discrepancy", Letter From H.C.Carlton (Westinghouse) to J.T.Reilly (Southern California Edison Company), Dated March 28, 1991.

2. SCE-91-533, "Integrated Assessment for Refill Volume Discrepancy",

Letter From S.A.Pujadas (Westinghouse) to B.Carlisle (Southern California Edison Company), Dated March 28, 1991.

3. SCS-71-29, Letter from R.L.Steele (Westinghouse) to L.D.

Hamlin (Southern California Edison), Dated October 15, 1971

4. "Westinghouse PWR core Behavior Following a Loss of Coolant Accident",

WCAP 7422-L, 1970.

5. WCAP 9565, "NATD Quality Assurance Program"

6. Letter from Carlton to Carlisle, "Increased Safety Injection Miniflow", SCE-91-528.

7. "Application of Realistic Thermal-Hydraulic Methods for Pressurized Water Reactors with Upper Plenum Injection", Nuclear Technology, Vol.

85, May 1989.

8. "Quantifying Reactor Safety Margins", NUREG/CR-5249, December, 1989.

Page 20 SCET:5/9/91:myy

.'M.Y 17 '91 19:06 FR LICENSING TO#E IRUINE PAGE.024 Table 1 SONGS-1 SLB M&E Release Analysis Case HZP with HZP without Change vol. errors vol. errors Peak Nuclear Power

.0887

.0960

+.0073 (fraction of nom.)

Time (t) of peak (sec) 68.2 68.2 0

Tavg at t (F) 364.1 365.0

+.9 Tin at t (F) 361.9 362.7

+.8 RCS Press (psia) 888.9 894.2

+ 5.3 Intg. 5ass at t 117.0 117.1

+.1 (X 10 bm)

Intg. inergy at t 140.3 140.5

+.2 (X 10 Btu)

Case HFP with HFP without Change vol. errors vol. errors Peak.Nuclear Power*

.2022 /.2019

.2095

+.0073 /.0076 (fraction of nom.)

Time (t) of peak (sec) 76.2 / 78.2 78.2

+ 2.0 / 0.0 Tavg at t (F) 377.4 / 374.9 375.7

-1.7 / +0.8 Tin at t (F) 372.6 / 370.1 370.8

-1.8 / +0.7 RCS Press (psia) 887.9 / 890.6 892.6

+4.7 / +2.0 Intg 4ass at t 120.8 / 122.3 122.5

+1.7 / +0.2 (X 10 bm)

Intg. Energy at t 145.0 / 146.7 146.9

+1.9 / +0.2 (X 10 8tu)

• Note: For HFP case with vol errors, values given at 76.2 seconds (time of peak return to power) and 78.2 seconds (time of peak return to power for case without volume errors) to permit comparison of effect on integrated mass & energy values, Since peak containment pressure may occur later, additional integrated M&E values for other times (for SCE use) are provided on the following page.

Page 21 SCE1:5/9/91 :MYY

1 MPY 17 '91 19:06 FR LICENSING T

E IRUINE PAGE.025 Table I (continued)

SONGS-I SLB M&E Release Analysis Time Integra ed Mass Integgated Energy (sec)

(X 10 ibm)

(X 10 Btu)

HZP with HZP without HZP with HZP without vol. errors vol. errors vol. errors vol. errors 100.

134.7 135.1 161.3 161.7 150.

154.0 154.8 184.0 185.0 200.

167.7 169.1 200.0 201.6 250.

178.0 179.7 211.9 213.9 300.

186.1 188.0 221.3 223.6 350.

192.4 194.7 228.6 231.3 400.

197.1 199.6 234.0 237.0 Time Integra ed Mass Integ ated Energy (sec)

(X 10 ibm)

(X 10 Btu)

HFP with HFP without HFP with HFP without vol. errors vol. errors vol. errors vol. errors 100.

136.6 136.9 163.7 164.1 150.

161.7 162.2 193.3 194.0

200, 176.1 176.1 210.2 210.2 250.

179.4 179.3 214.0 213.9

.300.

182.8 182.9 218.0 218.1 350.

186.4 186.2 222.1 222.0 400.

189.8 189.8 226.1 226.1 Page 22 SCE1:5/9/91:MYY