ML20149G950
| ML20149G950 | |
| Person / Time | |
|---|---|
| Site: | Calvert Cliffs  |
| Issue date: | 11/01/1994 |
| From: | Denton R BALTIMORE GAS & ELECTRIC CO. |
| To: | NRC OFFICE OF INFORMATION RESOURCES MANAGEMENT (IRM) |
| References | |
| NUDOCS 9411080252 | |
| Download: ML20149G950 (5) | |
Text
.
f 1
Ronn;RT V. Un'.NiON llaltirnure Gas and Electric Cor*1pany Calvert Clijp Nuclear Power Plant y;(,p,,,sident
'03" ("I" CI'll' Y"'N""?
Nuclear Energ'y l.ushy. Afaryland 20657 4 to 586-2200 lht. 4455 liwal 410 260.1455 llaltirnore November 1,1994 0
U. S. Nuclear Regulatory Commission Washington, DC 20555 ATTENTION:
Document Control Desk SUl? JECT:
Calvert Cliffs Nuclear Power Plant Unit Nos.1 & 2; Docket Nos. 50-317 & 50-318 l
Request for Approval to Use Convolution Technique in Main Steam Line Break An_abyis
REFERENCES:
(a)
Letter from Mr. A. E. Lundvall, Jr. (BGE), to Mr. R. A. Clark (NRC),
dated September 1,1983, Supplement to Unit 1 Seventh Cycle License Application (b)
CENPD-183-A, Loss of Flow, CE Methods for Loss of Flow Analysis, l
l June 1975 I
(c)
Letter from Ms. C. M. Thompson (NRC) to Mr. W. F. Conway (Anzona Public Senice), dated May 20,1991, Issuance of Amendment No. 26 to Facility Operating License, Palo Verde Nuclear Generating Station, Unit No. 3 The Baltimore Gas and Electric Company (BGE) requests Nuclear Regulatory Commission (NRC) approval to use the convolution technique for the analysis of the Prc-Trip Main Steam Line Break event at Calvert Cliffs Unit Nos. I and 2.
DESCRIPTION The proposed amendment would allow the use of the ABB/ Combustion Engineering (ADB/CE) convolution technique in the Calvert Cliffs Main Steam Line Break Analysis. The convolution technique utilizes fuel damage probability distributions in estimating the number of failed fuel rods resulting from a main steam 080133 94110G0252 941101 PDR ADOCK 050003t7 P
PDR k
"p\\'
o (WV 2 9M3274 t
1 l
Document Control Desk November 1,1994 Page 2 line break (MSLB). This technique has been previously approved for the analysis of similar postulated accidents for Calvert Cliffs.
BACKGROUND At Calvert Cliffs Nuclear Power Plant, the Main Steam System carries steam from the two steam generators to the main turbine. A rupture in the Main Steam System, known as an MSLB accident, increases the rate of heat removal by the steam generators and causes a cooldown of the Reactor Coolant System (RCS). The Calvert Cliffs reactors are designed with a negative or neutral Moderator Temperature Coefficient. Therefore, a cooldown of the RCS is assumed to result in an increase in reactor power. The decrease in main steam pressure will also result in a reactor trip and the closure of the main steam isolation i
valves. Further, the analysis assumes that a concurrent loss of offsite power occurs and the reactor coolant pumps are dcenergized and coast down. Calvert Cliffs Updated Final Safety Analysis Report (UFSAR),
Section 14.14, describes the analysis of an MSLB in detail.
The MSLB accident is analyzed for a range of break sizes inside and outside of the containment, initiated from tmth full-power and zero-power conditions. In practice, the analysis is divided into two components:
pre-trip effects and post-trip effects. The pre-trip analysis evaluates the effects on the reactor fuel caused by the event before the reactor is shut down by the control rods released by a reactor trip signal. The primary concern in this analysis is fuel failure which could result in offsite dose corsequences due to high power and low flow. The post-trip analysis evaluates a return to power caused by the continuing RCS cooldown after the reactor has tripped.
In the pre-trip analysis, the combination of the increase in reactor power and the decrease in RCS flow may cause some of the fuel rods to experience Departure from Nucleate Boiling and these rods are assumed to fail. The margin to Departure from Nucleate Boiling is measured using the Departure from Nucleate Boiling Ratio (DNBR). Given a population of fuel rods and their associated DNBR, there are two techniques used to determine if those fuel rods are experiencing Departure from Nucleate Boiling. He deterministic approach is to assume that all fuel rods with a DNBR less than the limit fail. This is the technique described in UFSAR Section 14.14.4.2.b and documented in Reference (a), our MSLB analysis i
of record.
The second technigre, called convolution, uses a statistical approach in determining the number of fuel rods
)
that fail. Because the DNBR limit is based on a probability that fuel rods with this DNBR are not actually experiencing Departure from Nucleate Boiling, not every fuel rod with a DNBR less than the limit will actually fail. This probability is applied to the population of fuel rods with a given DNBR to determine the i
number of fuel rods in that population that are actually experiencing Departure from Nucleate Boiling. As j
in the deterministic approach, all fuel rods predicted to experience Departure from Nucleate Boiling are still assumed to fail. The convolution technique was described in detail in Reference (b) and generically approved by the NRC for Seized Rotor Events in a Safety Evaluation, dated May 12, 1982. He current Calvert Cliffs Seized Rotor Event analysis, UFSAR Section 14.16, utilizes this convolution technique.
It has recently come to our attention that previous MSLB analyses performed for Calvert Cliffs by ABB/CE used the convolution technique without explicit NRC approval. ABB/ Combustion Engineering has provided core design and accident analysis ser ices for BGE since initial startup of the units. For each
Document Control Desk November 1,1994 Page 3 reload cycle, ABB/CE provided BGE with a Reload Analysis Report which summarized the analyses I
performed. For those reload cycles which required NRC approval, BGE prepand a license amendment request based on ABB/CE analyses.
i e
During a 1994 review of accident analyses performed by ABB/CE for Unit 1 Cycle 12 reload (and prior to the approval and use of the information), BGE discovered a difference between the description of an 4
analysis in the Reload Analysis Report and the analysis performed. Therefore, we requested that ABB/CE verify the descriptions given in the Reload Analysis Report of all the analyses. During that review, ABB/CE discovered that the MSLB discussion in the Reload Analysis Report incorrectly stated that the deterministic fuel rod failure estimation technique was used when, in fact, ABS /CE had used the convolution technique to determine the number of failed fuel rods in the MSLB analysis. After reviewing past reload analyses, ABB/CE informed BGE that the convolution technique had always been used for the Calvert Cliffs MSLB Analysis even though the analysis of record, Reference (a), stated the deterministic technique was used. This analysis was the basis for assessing the acceptability of subsequent operating cycles. Baltimore Gas and Electric Company directed ABB/CE to immediately perform a new MSLB analysis using the technique described in Reference (a) (i.e., the deterministic fuel failure estimation technique). The reanalysis resulted in an increase in the fuel pin failures from less than 2% to 5.6%.
Baltimore Gas and Electric Company then adjusted the operating parameters in the Calvert Cliffs Core Operating Limits Report to maintain the fuel damage fraction to within that previously reported to the NRC.
The use of the deterministic technique in the MSLB accident and the corresponding revision of the Core Operating Limits Report imposes operational restrictions on Calvert Cliffs Units 1 and 2..it is possible that future cycles could require a reduction in reactor power as the reactor peaking factors approach their Technical Specification limits. Because the convolution methodology has been approved by the NRC for use by BGE in a similar analysis, BGE desires to obtain approval to use the convolution methodology in future pre-trip MSLB analyses in order to regain our operating margin.
JUSTIFICATION AND SAFETY ANALYSIS ABB/ Combustion Engineering Topical Report CENPD-183-A, "C-E Methods for Loss of Flow Analysis,"
Reference (b), describes the ABB/CE generic method for analyzing loss-of-flow transients for ABB/CE plants. Tnese transients consider the normal coastdown of one or more reactor coolant pumps or the shaft seizure of one pump. Of particular interest is the minimum value of the DNBR reached in the hot channel during the transients exacerbated by an early power increase caused by an assumed positive moderator temperature coefficient. These events are very similar to the pre-trip MSLB in that all consider loss of flow, early increases in power, and the departure from nucleate boiling.
The convolution technique involves the following steps:
a.
The minimum DNBR for an event is calculated for fuel rods with various radial peaking factors.
In the calculation of minimum DNBR versus radial peak, all hot channel factors are applied conservatively to maximize the radial peaks. This yields a table of minimum DNBR versus radial peaking factor for the event.
l i
e Document Control Desk November 1,1994 Page 4 b.
From the resu'lts of neutron flux calculations (for allowabic rod configurations and various bumups), a census is generated of the number of fuel rods having any given radial peak. The census is chosen to maximize the number of fuel rods with large radial peaks. This radial peak census is combined with the radial peak to DNBR data determined in Step a. to determine the number of fuel rods experiencing any given minimum DNBR.
c.
The number of fuel rods with a given minimum DNBR is multiplied by the probability of experiencing departure from nucleate boiling for that DNBR. This is done for all rods and the results are summed to yield the total number of fuel rods that fait during the event being analyzed.
This process is describ:xiin detail in Section 3.2.2.1 of Reference (b). The NRC Staffs Safety Evaluation Report of CENPD-181-A stated, "We further conclude: the statistical convo!ution technique is acceptable for fuel rod failure calculations." The use of the convolution technique to evaluate fuel failures should be applicable to pre-trip MSLB analyses.
In a safety evaluation dated May 20,1991 (Reference c), the NRC has given approval for the use of the convolution technique for the inadvertent opening of an atmospheric dump valve with Loss of AC power analysis for Palo Verde Unit 3. This event is very similar to the Pre-Trip MSLB cvent. An initial power rise due to an incicased steam flow is exacerbated by reactor coolant pump coastdown due to the loss of AC power. As in the Pre-Trip MSLB analysis, a limited number of rods experience Departure from Nucleate Boiling before the control rods shutdown the reactor. In the safety evaluation, the NRC staff stated, "This approach [ convolution technique] has been found acceptable by the staff for analysis of limiting faults such as the locked rotor, sheared shaft and CEA ejection accident at PVNGS. These are occurrences that are not expected to occur but are postulated because their consequences would include the potential for the release of significant amounts of radioactive material." At Cahrrt Cliffs, the MSLB event (as well as seized rotor and control element assembly ejection) is classified as a " Postulated Accident" which is a limiting fault that is not expected to occur (UFSAR Section 14.1.1.1.b and Table 14.1-1).
In conclusion, the convolution technique has been approved for the analysis of similar events at other plants, including Calvert Cliffs. Using the more conservative, traditional fuel rod failure estimation method would have significant operational impacts on Calvert Cliffs. Therefore, we request approval to employ the convolution technique for the analysis of the Pre-Trip MSLB cvent at Calvert Cliffs Units 1 and 2.
SCI 1EDULE This change is requested to be approved and issued by January 31,1995. Small power reductions (in the order of a few percent power) are considered possible during the current Unit 2 fuel cycle until this request is approved and the core operating limits are adjusted accordingly.
l
- l
Document Control Desk
~
November 1,1994 Page 5 Should you have any questions regarding this matter, we will be pleased to discuss them with you.
Very truly yours, m
F D_sv3e
{
RED /BDM/d!m cc:
D. A. Brune, Esquire J. E. Silberg, Esquire L B. Marsh, NRC D. G. Mcdonald, Jr., NRC T. T. Martin, NRC P. R. Wilson, NRC R. I. McLean, DNR J. II. Walter, PSC