ML19347F515

Testimony of R Williams & N Horton Re Doherty Contentions 14 & 25,on Main Steam Line Radiation Monitor & Fuel Failure/ Flow Blockage,Respectively.Prof Qualifications Encl.Related Correspondence

ML19347F515
Person / Time
Site:	Allens Creek File:Houston Lighting and Power Company icon.png
Issue date:	05/11/1981
From:	Horton N, Robert Williams GENERAL ELECTRIC CO.
To:
Shared Package
ML19347F516	List: ML19347F515 ML19347F520 ML19347F522 ML19347F524 ML19347F527 ML19347F529 ML19347F530 ML19347F531 ML19347F532 ML19347F534 ML19347F535 ML19347F536 ML19347F537 ML19347F539
References
NUDOCS 8105190504
Download: ML19347F515 (19)

Text

_-

--WTED_ CORRESPONDENCE May 11, 1981 O qf

~ a

// Dxwes

"# 3 1 UNITED STATES OF AMERICA ,,

NUCLEAR REGULATORY COMMISSION $-

, .m

, , , 1v

, 19 81 > -4 2

BEFORE THE ATOMIC SAFETY AND LICENSING BC omes of thesec.

3 e gfg asn. g

/

4 In the Matter of ) b N

)

5 HAUSTON LIGHTING & POWER COMPANY) Docket No. 50-466

)

6 (Allens Creek Nuclear Generating)

Station, Unit No. 1) )

7 )

8 9 TESTIMONY OF DR. RICHARD WILLIAMS AND NED HORTON ON BEHALF OF HOUSTON LIGHTING & POWER CO. ON 10 DOHERTY CONTENTION 25 - FUEL FAILURE / FLOW BLOCKAGE; AND DOHERTY CONTENTION 14 - MAIN STEAM 11 LINE RADIATION MONITOR 12 Q. Please state your names and place of employment.

13 A. My name is Rienard J. Williams and I am employed by General 14 Electric Company as an Engineer.

15 My name is Ned R. Horton, I am cmployed by General Electric 16 as Manager of the Chemical and Radiological Design Unit.

17 Q. Dr. Williams, please state your educational 18 background, work experience and professional qualifications.

19 A. A statement of my qualifications is attached as 20 Exhibit RJW-1 to this testimony.

21 Q. Mr. Horton, please state your educational background, 22 work experience ar.a professional qualifications.

23 A. A statement of my qualifications is attached as 24 Exhibit NRH-1 to this testimony.

25 Q. Dr. Williams. what is the purpose of your testimony?

26 A. The purpose of my testimony is to address Doherty 27 Contention 25 which asserts that analyzing the flow .)1ockage 28 of only one fuel assembly is inadequate. In support of

/I 810s100 g

1 2 this assertion, Intervenor relies on the 1966 incident at' 3> Fermi Unit 1 which involved blockage of two fuel assemblies.

4 The source of the material for blocking the flow paths is 5 alleged to be various reactor internal parts including piping, fuel assembly pieces, channel box sections and steam 6l 7 dryer sections. Intervenor is also concerned that a flow 8 blockage incident would go undetected in its early stages 9 and contends that such an incident could result in fuel 10 melting and therefore adverse safety consequences. Inter-11 venor proposes that Applicant analyze for blockage of more 12 than one fuel assembly and incorporate experimental flow 13 blockage detection instrumentation in the Allens Creek 14 design.

15 Q. Mr. Horton, what is the purpou ' of your testimony?

16 A. The purpose of my testimony is to address Doherty 17 Contentior 14 which asserts that the design of the Main 18 Steam Line Radiation Monitor (MSLRM) is not adequate to detect 19 rapid fuel failure. In support of this assertion, Intervenor 20 refers to incidents at Dresden Unit 3 in October, 1974, 21 and Three Mile Island Unit 2 in March, 1979. Intervenor con-22 tends that due to background radiation of Nitrogen-16, the 23 monitor cannot be set low enough to detect rapid fuel failure.

24 Q. Dr. Williams, let us first address the flow blockage 25 contention. Have any incidences of flow blockage resulting in 26 fuel damage occurred?

27 28

i f

1 2 A. The most significant occurrence of flow blockage occurred at Fermi Unit 1 in 1966. In this incident, two fuel sub-3 4 asser..ailes experienced significant flow blockage that re-5 sulted in damage and some degree of fuel melting. The 6 blockage was caused by a zirconium segmene of a flow guide 7 that had torn loose during operation.

8 In addition, flow blockage occurred at the Materials 9.

Test Reactor (MTR) in 1962 and at the Engineering Test 10 Reactor (ETR) in 1963. The MTR blockage was due to dis-11 lodged gasket material from the primary system that partially blocked one assembly inlet. The ETR blockage resulted from 12 13 parts of a lucite sight box.

14 It is important to note that the consequences of all 15 three incidents were minimal and were confined to shutting 16 down the reactor, cleaning up the primary system and re-17 placing damaged assemblies.

18 Q. Is the Allens Creek design similar to the design of 19 the reactors which you have just mentionec?

20 A. No. The design of Fermi Unit 1 is a liquid metal 21 cooled fast breeder test reactor, the design and operation 22 of which are significantly different from a modern GE BWR.

23 simil arly, the MTR and ETR differ significantly ' rom the 24 Allens Creek design and thus the referenced flow blockage 25 events are not indicative of the probability of a flow 26 blockage occurrence in a BWR.

27 Q. Can a flow blockage occur in a BWR/6 such as Allens Creek?

28

1 2, A. Possible flow blockage mechanisms including a foreign 3 object, crud buildup, and fuel clad swelling have been studied.

4 Foreign objects are the most likely method of causing 5 significant flow blockage in a BWR. Crud buildup is a 6 less severe blockage than that which could occur for a 7 foreign object because it would.be a core wide restriction 8 causing an increase in the core pressure drop. This g increase would easily be detected before the severe flow

~

10 reductions necessary to induce fuel failures coald occur.

11 Fuel cladding swelling could result from a nuclear 12 excursion incident (the most severe excursion would be a 13 rod drop accident) or a loss of coolant accident (LocA) 14 resulting in a flow reduction. General Electric has con-15 ducted experiments using heated rods to simulate the 16 conditions during a loss of coolant accident. The results 17 from these experiments confirm that partial flow blockage 18 does occur due to the swelling and perforation of the fuel 19 cladding, but failure propagation does not occur and the 20 resulting reduction in flow area does not impair the 21 coolability of the fuel bundle. In addition actual transient 22 irradiation tests were performed on BWR type fuel pins.

23 Results of these tests confirm that a nuclear excursion, 24 limited to GE design basis energy depositions, 25 does not cause sufficient flow blockage to increase the 26 fuel damage beyond that induced by the transient itself, 27 and certainly does not induce fuel melting.

28 Q. How could a foreign object cause a flow blockage at

f 1

2 ACNGS?

3 A. A foreign object could cause flow reduction by 4 blocking the orifice in the fuel support piece, by 5 entering the fuel support piece and becoming lodged 6 between the fuel support piece and the lower tie plate 7 e a fuel assembly, or by going through the lower tie plate 8 and becoming lodged in the fuel assembly.

9 An object entering the fuel assembly would be of less 10 consequence than one blocking the orifice or lower tie plate 11 because it would have to be of much smaller size to 12 enter the bundle and would therefore block a much smaller percentage of the flow area. It would, therefore, take 13 14 the extremely low probability event of a large number of 15 small objects entering and Icaging in one fuel assembly 16 to block flow sufficiently to cause extensive fuel damage 17 of the type postulated by this contention. For this reason the more likely mechanism of flow blockage is an 18 19 object blocking the fuel support piece orifice'or lower f

20 tie plate.

f

! 21 Q. Please discuss the flow blockage process at the fuel l

22 support piece orifice and lower tie plate.

l 23 A. Possible foreign objects that could block flow include 24 reactor internal pieces that break loose or objects in-l 25 advertently left in the Reactor Pressure Vessel (RPV) after l

26 fuel loading and not detected during the extensive pre-start-l 27 up inspections. In order for an object to block a fuel 28 assembly it must reach the lower plenum area of the RPV

I 1 2 and be swept up by the coolant flow to the fuel support 3 P i ece. To get to the lower plenum region, an object would 4 have to make its way through the jet pumps which 5 have 5 nozzles with a diameter of 1.3 inches and a throat diameter of 6.5 inches. After reaching the lower plenum, 6

7 the object would have to travel a tortuous path between 8 the control rod guide tubes to reach an inlet orifice of 9 a fuel support piece. The guide tubes are arranged with 10 a maximum gap (diagonally) of 6.096 inches and a minimum 11 gap of 1.125 inch.

12 It is clear therefore that the probability of a 13 piece of 12 actor internals blocking the inlet would be 14 very low because of the path through small openings 15 it must take. In addition, in contrast to allegations by 16 Intervenor, the reactor internals for ACNGS are very 17 similar in design to the internals used in many operating 18 BWR's that have accumulated hundreds of reactor-years 19 of service without experiencing a single case of flow blockage 20 due to parts that have broken loose.

21 The second possibility for a foreign object to block 22 flow to a fuel assembly would be if the object were already in the lower plenum. This would be very unlikely because 23 24 of the procedures taken to ensure the cleanliness of the 25 vessel when the internals are installed, and the fact that 26 any object not in the plenum at initial reactor startup 27 would have to travel the path through the jet pumps 28 described previously.

1 2 If an object were in the lower plenum, it would have 3 to be swept upward off the vessel bottom by the coolant 4 flow and reach the fuel support piece orifice. The following 5 factors would reduce the likelihood of this occurring and 6 the consequences if it did:

7 a. There are very few locations where the radial 8 velocity would be high enough to sweep the piece off 9 the floor through the narrow 1.125 inch gap between 10 guide tubes.

11 b. If an object fell to the bottom of the vessel, 12 it would tend to drift toward the vessel centerline 13 where horizontal velocities are low and the boundary 14 layers on the vessel may be thicker than the object.

15 Thus, the boundary-layer effect would reduce the 16 capability of the fluid to sweep the piece up off 17 the floor of the vessel so that the vertical com-18 ponents could carry it upward.

19 c. Even if the object were somehow swept upward, 20 it is unlikely that it could completely block the 21 orifice holes, which are vertically oriented.

22 d. If the object were small enough to pass through l

23 the orifice, it would have to pass through the lower 24 tie plate nosepiece and the lower tie plate to enter l 25 into the fuel channel, which would require very 26 unlikely alignment of the object and the passage 27 through holes of only 0.410 inch diameter.

28 e. If the object were to make it through the lower

1 2 tie plate, it would be stopped by the first spacer, 3 which would probably cause local boiling transition 4 and overheating. Depending on the size and shape, 5 the object will most likely remain in a vertical 6 position since the maximum distance between fuel rods 7 is only 0.422 inch. The object would not significantly g reduce the flow in one bundle or cause serious 9 degradation of the heat transfer conditions in other 10 areas of the fuel assembly.

11 Therefore, even though it is possible for minor blockages 12 to occur by small objects entering the fuel bundle and 13 affecting the life of the fuel, it is very unlikely that 14 a blockage which would induce a significant flow reduction 15 will occur.

16 Q. If a flow blockage incident occurred, what would be 17 the consequences?

18 A. As previously indicated, a BWR fuel assembly can with-19 stand severe flow blockage without losing adequate cooling.

20 Analysis of the flow blockage situation using the General l

l 21 Electric thermal hydraulic computer model has shown that l

22 for limiting high power bundles, blockage from more than 23 98% of the flow area would result in melting of the clad l

l 24 and fuel. Flow would be reduced to approximately 5% which l

25 would come from various leakage paths into the assembly.

26 The resultant release of fission gas products will cause a 1

27 reactor scram to be initiated by the MSLRM in approximately 28 15-20 seconds. The reactor scram will terminate the

1 2 transient by shutting down the generation of heat in the 3 fuel assembly.

4 Q. Intervenor contends that it is necessary tc consider 5 blockage of more than one assembly. Please discuss the 6 mechanism, if any, of blocking the flow to multiple fuel 7 assemblies.

8 A. It is impossible for one object to block flow to more 9 than one assembly because of the design of the fuel support 10 piece used in a BWR. Each support piece supports four fuel 11 assemblies, and has four separate vertical orifices spaced 12 at 90 degree intervals to allow coolant to enter each bundle.

13 The orifice is typically approximately 1.50 inch in diameter.

14 Fuel support pieces for some fuel assemblies on the perimeter 15 of the core only support one assembly and thus only have one 16 orifice of approximately 2.50 inches in diameter.

17 Because of t'e design where each fuel assembly has a 18 unique flow path for coolant to enter it is impossible for 19 a single object to block multiple fuel assemblies. The 20 only possible way to block coolant flow to more than one 21 assembly is for more than one object to block separate flow 22 paths at virtually the same instant. This extremely low 23 probability event would have to occur almost simultaneously 24 because a flow blockage severe enough to cause extensive 25 fuel camage to one assembly would cause a reactor scram 26 almost immediately as previously discussed.

27 The ACNGS design therefore is in stark contrast to the 28 design for Fermi Unit 1 referenced by Intervenor, where the

I 1 2 coolant entrance orifices for each assembly were in a 3 horizontal plane such that one object could block more 4 than one assembly.

5 O. Let us now address detection of fuel failures if a 6 flow blockage occurs. Mr. Horton, the Main Steam Line 7 Radiation Monitor (MSLRM) and Offgass Radiation Monitor 8g (GRM) have been discussed as devices that will detect fuel 1

9 failures. Please describe their operation.

10 A. Fuel damage, ranging from cladding failure to fuel 11 melting, will result in direct release of fission products 12 from the fuel, with the rapidity and quantity of release 13 generally directly related to the rate of occurrence and 14 overall severity of fuel rod damage. Fission prod.0t 15 release from the fuel would initially result in the release 16 of noble gases and halogens, and in the worst damage cases 17 by the release of other fission products (particulates).

18 The transport of the noble gases (Xenon, Krypton) in the 19 steam provides the primary method of detecting fuel failures.

20 The detection systems that will respond to fuel failures are 21 the MSLBM and ORM.

22 The purpose of the Main Steam Line Radiation Monitors 23 (MSLRM) is to provide prompt detection of any core damage 24 event which results in the significant release of noble gases.

25 As a part of the Reactor Protection System, the MSLRM provides 26 prompt reactor scram and main steam line isolation.

27 The MSLRM operates in the presence of a normal high 28 level Nitrogen-16 activity in the steam, which is on the

1 2' order of 100 curies per second at rated power for large BWRs.

3 The trip point is normally set at about 3 times the N-16 4 background radiation level. Due to the lower average gamma 5 decay energy of noble gases released from failed fuel, 6 releases on the order of a few kilocuries per second are 7 necessary to cause MSLRM trip and initiation of the protection 8 system. Thus the MSLRM would not react to the noble gas 9 releases associated with expected cladding defects of a 10 single rod. These releases are expected to be on the order 11 of a few curies per second per fuel rod. However, the MSLRM 12 would promptly react to a massive noble gas release which 13 would occur coincident with fuel Irelting.

14 The detection and shutdown sequences has been evaluated 15 for the low probability event of complete orifice flow 16 blockage. This degree of flow blockage is predicted to lead 17 to melting of fuel. For this event, once fuel melting in Ib the hottest rod in the bundle starts, the melting will increase 19 across the bundle as well as axially. Considering the action 20 in the initially hotest six inch axial node, the fuel node 21 melting from start to completion vill occur within approxi-22 mately ten seconds. This six inch node corresponds to about N 4 percent of a total rod length. Considering the noble gas N inventory in this node of the fuel assembly only, and N ignoring melting in other axial nodes which would soon follow, N the average release rate over the first ten second period would be 7.5 x 10 3 curies per second. When the monitor trip 28 is set at three times normal N-16 background, a noble gas

a 1 2 release of about 1.1 x 10 curies per second is required 3 for a trip signal. Therefore the release of the fission gases 4 in the highest powered axial node is sufficient for the 5 MsLRM to trip and initiate scram and isolation.

6 At rated power, the noble gas transit time from core 7 midplane to the MSLRM is within 9 seconds for the typical 8' large BWR. Detection, scram, and steam line isolation 9 follows within approximately 6 seconds. These time estimates 10 are based on the transit time for steam to travel from the 11 reactor core to the location of the MSLRM on the steam lines.

12 Thus reactor scram and isolation will occur within ap-13 proximately 15-20 seconds after fuel melting begins.

14 The offgas radiation monitoring system, which also serves 15 as the pre-treatment monitor on augmented offgas treatment 16 systems, provides routine core surveillance by detection of 17 low-level emissions of noble gases. A two minute holdup 18 period is provided to allow for the decay of N-16 and other 19 half-life fission products and activation products.

20 In normal operation, an alarm trip setting of several 21 times the offgas level is established to provide prompt 22 operator warning of any significant change. With expected 23 fuel performance, any change associated with an increase 24 of the order of 2-10 curies per second would be promptly 25 alarmed.

26 For the event where the assumed flow blockage provides 1

27 bundle flow sufficient to prevent fuel melting, but not 28 enough to prevent boiling transition, cladding oxidation l

1 2 and fragmentation is expected to occur after some period 3 of time. The noble gas inventory of an average fuel rod is 4 about 30,000 curies. In terms of reference to the two 5 minute decay at the offgas monitor the noble gas inventory 6 would be at least 10,000 curies per rod. At two minutes 7 decay, release of 1% of the inventory of a single rod is 8 equivalent to 100 curies. As indicated previously, the 9 offgas monitor normally would respond sharply to a sudden 10 increase of 2-10 curies per second (and perhaps 0.1 curies 11 per second most of the time) . Therefore if the 100 curies 12 did in fact leave the fuel rod in a period less than 10 13 seconds, the offgas monitor provides the desired alarm to 14 operating personnel. The failura of more than one fuel rod 15 would provide additional activity for monitor response.

16 Upon the occurrence of an alarm, operator action would be 17 initiated which would involve if necessary, power reduction 18 or reactor shutdoun, depending upon the level of the increase 19 indicated.

20 Thus for any flow blockage event which results in

, 21 significant cladding failure, alarm and control is available 1

22 from the offgas monitoring system. Noble gases reaching the l 23 condenser would enter the offgas treatment system as long as 1

24 steam is available. If reactor shutdown occurs, offgas i 25 system isolation would retain the stored gases for decay.

1 26 Halogens and other fission products reaching the condenser 27 would flow to the offgas system, or would be initially 28 retained in the condenser air space or in the condensate.

l' 1 '

2 Q. Intervenor has cited occurrences at Dresden Unit 3 and 3 Three Mile Island Unit 2 as being indicative of the inability 4 of detection systems to detect rapid fuel failures. Please 5 ! discuss the incidents with regard to detection of fuel failure 1

6 by the MSLRM and/or ORM.

7 A. Three Mile Island Unit 2 is a pressurized water reactor 8 (PWR) which makes use of steam generators to isolate the 9 primary coolant from the steam cycle. This design is sub-10- stantially different from the boiling water reactor (BWR) 11 design used at Allens Creek. Unless steam generator tube 12 leaks exist, coolant from the reactor does not mix with the 13 steam in a PWR; therefore, a radiation monitor on the steam 14 lines (MSLRM) or on the condenser effgas system (ORM) would not 15 detect fuel failures as for a BWR, This incident is obviously 16 not applicable.

17 The fuel failures that occurred at Dresden resulted 18 from a failure mechanism known as pellet-clad interaction 19, (PCI). PCI type failures are typically in the form of a 20 small cladding crack or split. These tightly formed cracks 21 would release fission products at a relatively slow rate 22 compared to molten fuel. The rod inventory of fission gas 23 therefore would be released over an extended time period.

24 The defects at Dresden were observed to occur over a five 25 hour period. Although a namber of rods failed, this 26 incident was not a rapid fuel failure as would occur for a 27 complete flow blockage where clad and fuel melting would 28 occur within seconds of its occurrence.

1 2 Q. Intervenor also relies on a statement in NUREG-0401 3 which concludes that a postulated flow blockage might pro-4 coed undetected in its early stages, to indicate the in-5 adequacy of the MSLRM and possible adverse consequences of 6 flow blockages. Is the capability of the MSLRM adequate 7 for mitigating the consequences of a flow blockage?

8 A. A flow blockage incident that would cause fuel failures 9 is an unlikely event in a BWR. Radiation detection systems 10 such as the MSLRM or ORM are not designed to prevent the 11 occurrence of fuel damage but to mitigate the damage once 12 it has occurred. By their very design the MSLRM and ORM 13 depund upon the release of fission products to detect fuel 14 failure. As previously discussed, these systems will detect 15 the range of fuel 'ailures that can result from a fuel blockage 16 incident.

17 Q. What are your conclusions?

18 A. If blockage does occur, resulting fuel damage 1 19 will be detected by either the Main Steam Line Radiation 20 Monitor or Offgas Radiation Monitor, depending upon the 21 severity of fuel failure.

22 23 24 25 26 27 28

I' Exhibit RJW-1

( 2 EDUCATION AND PROFESSIONAL QUALIFICATIONS 16 3'

0 Dr. Richard J. Williams l

4d Dr. Williams is an Engineer working in the Fuel Rod 5- Thermal and Mechanical Analysis Unit at General Electric's l California. His 6 Nuclear Energy Business Group in San Jose, Dr.

7' employment with General Electric began in January 1979.

Williams is responsible for nuclear fuel rod integrity under 8lI normal, off-nor- 1, transient and accident conditions. In >

9' 10 this capacity he has performed analyses of fuel integrity under Ili the loss of coolant accident (LOCA), the reactivity initiated 12l; accident (RIA) and the flow blockage event. He has also 13 . participated in the Three Mile Island Utility Support Program 14 providing analyses of the reactor fuel condition following the 15- incident.

Dr. Williams is a major General Electric engineering 16l' interface with government and regulatory agencies on the Fuel 17 Rod Research Programs being carried out in the United States, 18 Europe and Japan. He is currently involved with the OPTRAN l

19 test series being carried out in Idaho, the NRU LOCA Program 20 being carried out in Canada and the Super-Ramp Project 21 Committee, which governs the fuei rod testing being carried 1

22! out in the Studsvic reactor in Sweden.

L 23 Prior to working at General Electric, Dr. Williams held 24 a National Research Council, (National Academy of Sciences) 25 resident research fellowship at NASA Ames Research Center, 26 Moffett Field, California where he directed cryogenic diode 27 heat pipe research. In this capacity he was responsible for 28 the thermal diodes to be used on the Long Duration Exposure

- -,-,*y ,-r w - -

1 Facility to be launched from the space shuttle, 2l Dr. Williams has published many papers in International

?

3' Technical Journals and holds both a BSc (1973) and Phd (1976) 4 in Mechanical Engineering from Swansea University (UK).

5

6l. l 7

8 I

i 9

10 11 12 13- .

14 15 16 17 18 19 20 I 21 l

22 23 24 25 26 27 28

1 Exhibit NRH-1 2 EDUCATION AND PROFESSIONAL QUALIFICATIONS 3 Ned R. Horton 4 My name is Ned R. Horton.and I am employed by the General 5 Electric Company - San Jose, California. My present position 6 with General Electric is Manager of the Chemical and Radio-7 logical Design Unit.

8 I graduated from the University of Wyoming in 1961, with 9 a Bachelor of Science degree in Mechanical Engineering and 10 accepted employment with Phillips Petroleum Company at the 11 National Reactor Testing Station, Arco, Idaho. I was employed 12 as a Reactor Test Engineer until the inception of the Loss 13 of Fluid Test (LOFT) program in 1962. At that time I trans-14 ferred to the area of nuclear safety and had responsibility 15 for performing radiological evaluations for the LOFT and 16 SNAPTRAN programs.

17 In 1965, I transferred to the Oak Ridge National Laboratory l

18 (ORNL), Oak Ridge, Tennessee, and had responsibility for 19 coordinating the LOFT Fission Product Sampling Program with 20 ORNL.

21 In June 1967, I accepted employment with the General 22 Electric Company in San Jose, California, in the area of 23 Radiological Evaluations. Since joining General Electric, 24 I have been responsible for developing analytical models to i 25 be used in safety evaluations as well as performing detailed l 26 radiological analysis with regards to design bases accidents, 27 normal operational releases, and hypothetical abnormal l 28 releases.

1 In the course of my work at the NRTS, ORNL, and GE I I

2 3 have written and contributed to numerous publications con-4 cerned with safeguards evaluations and model development.

5 6

7 8

i 9 t 10l 11 12 13

> 14 15 16 i

17 18 19 20 21 22 23 .,

24 25 1

i 26 i 27 28 I _