ML20198C083

Safety Evaluation Supporting Defueling Canister Design.Use of Canisters Contingent on Approval of Procedures Per Tech Spec 6.8.2

ML20198C083
Person / Time
Site:	Three Mile Island
Issue date:	11/05/1985
From:	NRC
To:
Shared Package
ML20198C081	List: ML20198C076 ML20198C083 ML20198C090 ML20198C101
References
NUDOCS 8511110288
Download: ML20198C083 (9)

Text

. .

NRC STAFF SAFETY EVALUATION OF DEFUELING CANISTER DESIGN DESCRIPTION OF CANISTERS The defueling canisters are designed to accept and confine the TMI-2 core debris ranging in size from fines of about 0.5 microns in diameter up to partial length fuel assemblies of full cross section. The canisters are to be an integral part of the defueling systems and are intended to provide effective confinement for transport and long term storage of the damaged core debris. They are designed to ensure their contents remain subcritical under all postulated on-site conditions and also, when in combination with a shipping cask, to remain subcritical under both normal and accident conditions during transport. The three types of canisters (i.e., fuel, knockout, and filter canisters) are equipped with fixed neutron absorber material for criticality control, with catalytic recombiners to control the concentration of combustible gas mixtures generated from radiolytic decomposition of water, and with appropriate process connections for filling, closing, dewatering, inerting, and monitoring. All three types of canisters have a nominal overall length of 150 inches with the outer shell being fabricated of 14 inch OD 304L stainless steel pipe with a nominal 1/4 inch wall thickness. A reversed dished tank end is welded to the shell to form the lower closure head.

The fuel canister is designed as a receptacle for large pieces of core material which will be picked up and placed either directly into the canister or into other containers which will be inserted into the canister. Within the cylindrical shell is a full length approximately 9 inch square cross section shroud forming an inner cavity. The shroud is formed of stainless steel plates with Boral sheets sandwiched between them to serve as neutron absorbers. The plates are seal welded to encapsulate the boral and protect it from corrosion. The thickness of the inner plates protects the boral sheets from impacts from the canister contents. The inner cavity is sized to accept the full cross section of an intact fuel assembly. The void space outside of the shroud is filled with a light weight cement / glass bead mixture to prevent migration of fuel material to this area. The shroud assembly is welded to cad supported by a bottom support plate which is welded to the inside diameter of the shell. The bottom support plate is designed to withstand the impact od a 550 pound piece of debris dropped the full length of the canister in water.

If the drop is in air, the weight is reduced to 350 pounds. The upper end of the shroud is fitted into a recess and supported by an upper bulkhead which is welded to the shell and forms the mating surface for the upper closure head.

The removable closure head is bolted to the bulkhead and sealed with gaskets.

it has a machined socket in the center of its exterior to mate with the single point grapple on the canister lifting tool. The empty dry fuel canister weighs about 1230 pounds.

The knockout canister is designed for use in the fuel debris vacuum system.

It wTil separate debris particles ranging from about 140 microns up to full fuel pellet size or larger. The process inlet line enters the top of the canister and bends to direct the flow tangentially along the inner circumference of the shell creating a swirling action that causes the 8511110288 851105 PDR P

ADOCK 05000320 PDR

t l

entrained debris to settle out in the canister vessel. The water then exits through an 850 micron screen to a process connection in the top of the canister. The canister internal assembly is supported from a bottom support plate that is welded to the inside diameter of the canister shell. The assembly is positioned at the top by welded chock blocks. The internals consist of an array of four outer poison rods and one central poison rod. The outer rods are 1.3 inch 00 stainless steel tubes that are filled with neutron absorbing B C pellets and sealed at both ends. The center rod is a 2.875 inch 3

OD guard pipe surrounding a 2.125 inch OD tube filled with B g C pellets. The rod array is supported laterally by seven intermediate suppott plates along its length. The guard pipe around the center poison rod forms a 1/16 inch i annular gap, open to the bottom head of the canister and connected to a process fitting at the top, to provide a canister dewatering pathway. The i empty knockout canister weighs 1046 pounds in air. <

The filter canister is designed for use in the fuel debris vacuum system, the ,

defueling water cleanup system, and the canister dewatering system. It will remove fuel fines larger than 0.5 microns from the process streams. The canister internals are attached to a bottom support plate which is welded to inner diameter of the canister shell. The internals are comprised of a bundle of 17 filter modules, a drain line, and a centrally located poison rod containing B 4 C pellets. The central poison rod is similar to that in the knockout cantster. The filter modules consist of elements which are a pleated sintered stainless steel media around a center support tube. The media and support tube are induction brazed to stainless steel end caps. Eleven elements are stacked end to end around a perforated drain tube and seal welded at the end caps to form a module. The drain tube is plugged at the top and open at the bottom. The process flow enters the top of the filter canister and flows around the filter bundle. The process liquid flows through the filter elements depositing the entrained particles larger than 0.5 microns on the outside of the media. The liquid enters the perforated tube and flows i,

downward into the bottom plenum of the canister. The effluent exits through the top of the canister via an effluent fitting connected by the internal drain tube to the bottom plenum. The empty filter canister weighs about 1440 pounds in air.

All types of canisters are designed with suitable process connections for their intended use. The top head of each type is provided with a 1/4 inch inert gas purge connection and a 3/8 inch drain fitting which is connected to t

the internal dewatering pipe. Each of these connections is fitted with a l Hansen quick disconnect coupling. The filter canister head has 21 inch inlet and outlet process connections, and the knockout canister has 2 inch inlet and outlet process connections. The process connections are provided with cam and groove type fittings which will be closed with expanding mandrel plugs af ter the canisters are filled. Welded to the top of each canister is a cylindrical skirt to protect the penetration fittings during normal handling and postulated handling accidents. All types of canisters have a machined recess

in the outside surface of the upper head to acconnodate the single point i lifting grapple used for normal handling operations. The bottom support  ;

l plates in all three types of canisters forms a fuel free " sump" in the bottom

head. This is connected to the drain fitting at the top head for canister dewa tering. In the fuel canister, the dewatering path is a 3/8 inch tube running from the lower head through the area outside the boral shroud. In the

. knockout canister, the annular gap between the center poison rod and its

I t

strongback pipe forms the dewatering pathway. In the filter canister, the process effluent pipe runs from the lower head region to the upper head penetration. A dewatering pathway is machined internally in the upper head from the effluent pipe to the drain fitting. All three types of canisters are designed with catalytic recombiner cartridges in the lower and upper heads.

These are described in more detail in the gas management section of this report.

STRUCTURAL EVALUATION The defueling canisters are designed to the requirements of the 1983 edition of the ASME Pressure Vessel Code Section Vill, Division 1, Part UW (lethal).

They have design pressures of 150 psig internal and 30 psig external.

Fabrication, inspection, and testing of the canisters is performed to the standards of the ASME Code. The canisters are Nuclear Safety Related and the licensee's procurement specifications require that they be manufactured under i the controls of a Quality Assurance program meeting the requirements of 10 CFR 50 Appendix B and ANSI N45.2. Structural analysis by the canister designers included evaluations of the loads imposed on the canisters during normal operations as well as postulated load drops and shipping accidents.

Acceptance criteria for normal operations was based on the ASME Pressure Vessel Code. In addition, analysis was performed to show acceptable safety margins when applying the specified stress factors of NUREG-0612 and ANSI l N14.6 for the normal handling condition. The design criteria for postulated

! accident conditions is that for the predicted deformed geometry following an accident, the canisters and their contents must remain subcritical,although leakage of material is permissible.

Canister structural analysis for the normal operation and handling condition  !

was performed using standard analytic techniques. This analysis demonstrated '

! acceptable design margins and met the requirements of the ASME Code and other  !

applicable regulatory requirements and industry codes and standards.

! The approach used in demonstrating that the canister design met the

, specification for the postulated accident conditions used a combination of l analytical methods and component testing. The design specifications for the shipping cask intended for use in transporting the filled defueling canisters is that it shall limit the loads imposed on the canisters to no more than  !

40 g's axial and 100 g's lateral during hypothetical transportation accidents l per 10 CFR 71. A detailed evaluation of the proposed cask's conformance to l this specification has been performed and included both analysis and impact testing of a scale model. This evaluation is presently under review by the NRC Transportation Certification Branch as part of the licensing process for the cask. Analysis and supporting drop tests of the canister was performed to demonstrate that the fixed pokons installed in the canister remain intact and capable of performing their intended criticality control function when i.

subjected to toese loads, or that subcriticality could be maintained by other geometrical constraints.

For onsite handling accidents, canister drops of 6 fect-1) inches in air

. followed by 19 fect-6 inches in water, or 11 feet-7 inches in air were considered to be credible. This does not include a potential drop in the Fuel Handling Building Truck Bay during cask loading. This potential canister drop will be evaluated in the fuel shipping Safety Evaluation Report. Combinations

of vertical and horizontal drops were considered. These drops were determined to impart loads on the canisters in excess of those for the transportation accident. Structural analyses were performed to determine the extent of the canister shell and internals deformation resulting from these loads.

Deformation of the canisters due to a vertical drop was determined by analysis of data from a drop test program and was found to be shell dependent. The predicted deformation in this case was a bulging of the canister shell below the lower support plate. No significant deformation of the canister internals, significant to the criticality analysis, is expected to occur from a pure vertical drop. This was demonstrated during actual drop tests for a bottom end impact. This also bounds the top end impact and for purposes of criticality analysis the deformed shape was assumed to exist at both ends of the canister.

For the horizontal drop case, the filter and knockout canister's internals were analyzed with finite element methods using the ANSYS computer code. It incorporated the actual non-linear properties of the material and accounted for geometric constraints imposed by the canister shells. The deformations predicted by these analyses with additional conservatisms on poison structure locations were used in the criticality calculations. The defarmed geometry for the fuel canister was determined by a 30 foot drop of a simulated partial lengtn unit. The testing showed insignificant deformation of the boral shroud from the lateral loads imposed.

Vector combinations of the vertical and horizontal load components were used to predict the effect of a drop in any orientation, and the conservatively modeled worst case deformed geometry for each type of canister was factored into the criticality analysis.

The NRC staff review of the licensee's structural analysis has determined that proper codes and standards were employed in the design of the defueling canisters. The structural analysis shows sufficient margins of safety when applying the maximum predicted loads expected during normal onsite operations and handling and subsequent transportation. The structural analysis for accident conditions used industry standard and NRC accepted analytic techniques and provides reasonable assurance that the maximum expected deformation has been predicted for factoring into the criticality analysis.

CRITICALITY EVALUATION The defueling canisters are designed to ensure their contents remain subcritical under all normal operational conditions and during all postulated accident conditions. The conditions analyzed included both a single canister configuration and an array of canisters on a 17.3 inch center to center spacing, which is the minimum spacing for all onsite storage rack locations.

Both an intact canister and a canister deformed by the worst case drop accident were modeled. The deformed geometry used in the calculations was that predicted by the structural analysis with additional conservatism for poison structure location. The canisters were modeled using computer codes generally recognized as acceptable by the NRC staff. The calculational model used the following conservative assumptions:

1. The canister's contents consist of batch 3 fuel only with the average batch 3 enrichment plus 2 standard deviatiens. Batch 3 fuel is in the highest enriched region of the core and has an average enrichment of 2.93 percent. It assumed no fissile burnup or fission product inventory that would contribute negative reactivity.

2. The canisters contents are assumed to contain no cladding or core

structural material and no soluble poison or control material (i.e. ,

control rod debris or burnable poison) from the core.

3. The contents are assumed to be fuel in the optimal lump size and to contain the optimal fuel to moderator ratio with no boration of the entrained water.

4. All void regions af the canister are assumed to be filled with fuel without regard to the weight restrictions on a loaded container. All l

three types of canisters contain catalytic recombiners in the upper and i

lower heads. The criticality analysis assumed that the regions occupied by the recombiners was filled with fuel.

5. The 6nalysis assumed the lowest possible loading of fixed poison material.

1 The canister geometry was conservatively modeled to account for the internal configuration and the structural members of canister internals and closure heads, i

The fuel canisters were analyzed for a single canister infinitely reflected by j water, an infinite array of canisters in unborated water, and a canister deformed by the bounding drop case. The deformed case assumed fuel had

!, migrated into the bulged lower and upper heads. All cases yielded a maximum Keff of 0.877.

l Two knockout canister configurations were considered. These included the standard undamaged configuration and the damaged configuration in which the i worst deformed geometry was used. The damaged configuration for the knockout canister did not assume that fuel had migrated into the upper and lower head regions as in the other types of canisters, and did not assume loss of the 8 4C pellets as in the filter canister. A drop test of an as-built knockout

canister was performed by Oak Ridge National Laboratory and demonstrated that the bottom support plate and the poison rods remained intact following the maximum predicted impact loads from a drop accident. These configurations were analyzed as a single canister infinitely reflected by water, an infinite array of undamaged canisters in unborated water, and a single dropped canister. The maximum calculated Keff was 0.915.

Two filter canister configurations were also considered. They assumed fuel above the lower support plate and a second configuration with fuel in the lower head plus fuel filling the filter element drain tubes (i.e., ruptured filters). The maximum calculated Keff when considering the single canister, the array of canisters, and the single dropped canister was 0.892.

The NRC staff performed independent calculations to verify the licensee's criticality analysis. These included computer code analysis of several test

.- r - , , . , . - . , _ _ - , . .

_ _ ._ _._ . _ _ _ _ _ _ _ . _ _ _ _ _. . _ . ~

p

~

t i

cases as well as an evaluation of the assumptions and the computer codes used by the licensee. The NRC results were in agreement with the licensee's.

Sumaries of the NRC's analyses are included as appendices 1 and 2 to this SER.

b '

~

The_ licensee presented additional analysis to determine the effects on i criticality by the canister transfer shielding. The staff determined that the analysis used acceptable analytic techniques with appropriate levels of i conservatism. This analysis showed that handling a filled undamaged canister in the proposed transfer shield will not result in a Keff of greater than 0.95. Analysis of a damaged canister will be performed on a case by case basis as needed.

CANISTER GAS MANAGEMENT --

i After filling a canister with fuel debris, water will remain in the canister.

1 Prior to dewatering, the canister will be completely flooded with RCS water.

L Following dewatering, the canister will contain residual water entrained in j the fuel debris as well-as a certain amount of free " slosh" water not removed l by the dewatering system. Since this water will be in direct contact with i fuel and fission product containing debris without benefit of the fuel

cladding to provide shielding from alpha and beta radiation, there could be j significant amounts of hydrogen and oxygen generation from radiolytic 1 decomposition of the water. Gas generation could result in internal pressure l build up and production of combustible gas mixtures inside the canisters.

Studies were performed by Rockwell Hanford Operations (RHO) to predict the j rate of gas generation and to develop suitable catalytic recombiners to j control the gas concentrations, j The rate of gas generation has been shown to be a function of 1) the amount

! of ionizing radiation emitted by debris in a canister, 2) the fraction of the energy absorbed in the water, 3) the ratio of peak to average decay heat j energy in the fuel debris, and 4) the amount of gas produced per unit of ,

j energy. Using the empirical relationship which has been confirmed

experimentally, the maximum theoretical gas generation rate has been predicted as 0.114 liters per hour of hydrogen plus oxygen in stoichiometric l proportions. The licensee's evaluation states that there is significant i conservatism in this calculation and provided what was considered a " maximum i realistic generation" rate based on what is considered a more probable condition in the core debris. The licensee's predicted maximum realistic gas  !

j generation rate is 0.0075 liters per hour. The conservatisms used in the theoretical predictions are as follows

1) the maximum quantity of fuel in a canister used in the calculations (800kg) did not include allowances for .

! residual water or for weighing accuracy. This quantity was reduced in the j " realistic" prediction, 2) the fraction of energy absorbed in the water i conservatively assumed that large amounts of water were present for absorption -

rather than using the maximum amount'of water that could possibly be present i in a filled canister, 3) the amount of gas produced per unit of absorbed

' energy assumed no oxygen scavenging (i.e. , chemical removal)' that would i produce excess hydrogen and resultant back-reactions, 4) the ratio of peak to i average decay heat energy is based on the most active region of an undamaged j core and does not account for possible dispersal of the material from this core region during the accident. The NRC staff reviewed the basis for the gas i generation rates' and concurs that there is significant conservatism in the

_ _ ~ _'

_ _ _ _ . _ _ _ _ - , - - ,~,.__,; _ _ . _ _ . _ _, _ . _ _,_,_ . _ _ . - . _

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

1 .

l- theoretical generation rate. However, there is insufficient data presented in

the Technical Evaluation Report to justify the staff's use of the licensee's

! lower predicted " realistic" rate or to accurately quantify the conservatisms in the theoretical calculations. Therefore, the staff's safety evaluation is based on the 0.114 liter per hour maximum theoretical gas generation rate.

i Following a series of tests by RHO, the catalyst chosen for use in the

defueling canisters was a mixture of 80 percent Engelhard Deoxo-D nuclear grade catalyst and 20 percent AECL silicone-coated catalyst. Details of the catalyst test program are documented in GEND-051, " Evaluation of Special Safety Issues Associated with Handling the Three Mile Island Unit 2 Core Debris", dated June 1985. The test program involved a catalyst bed similar to that in the canisters. It was installed in a test chamber into which hydrogen and oxygen were admitted at a controlled rate. The test chamber's pressure ano temperatures were monitored and its internal atmosphere was sampled and analyzed. The tests demonstrated that the designed catalyst beds containing

! 100 grams of catalyst in the required proportions were capable of maintaining

the chamber atmosphere below 1.2 percent hydrogen and 0.6 percent oxygen while 4

recombining the gases at a rate of 0.3 liters per hour of hydrogen plus oxygen

) in stoichiometric proportions. This shows significant margins of safety from

, the lower flammability limits of 5 percent oxygen and 4 percent hydrogen, and

! from the maximum theoretical gas generation rate of 0.114 liters per hour.

The testing demonstrated, though that the catalysts do not function when i

ininersed in water. After immersion and being " drip dried" in a 100 percent relative humidity atmosphere, they will begin recombination at a reduced rate.

j The rate increases and reaches full capacity within a short period of time as the heat generated by the recombination reaction causes further drying of the catalyst. Further testing was performed to demonstrate that the chemical j

species expected to come in contact with the catalyst from the RCS or during

canister fabrication will have no deleterious effects on the catalyst perfonnance. Additionally, tests were performed to demonstrate that freezing conditions during transportation will not stop the recombination reaction once started.

J The catalyst teds installed in the defueling canisters are designed so that as

- long as the canister is no more than half full of free wat'er, at least 100 grams of the catalyst will not be ininersed in water regardless of canister orientation. Four recombiner packages, each containing 25 grams of catalyst, are attached synnetrically about the axis of the inner surface of the lower canister head in all types of canisters. The upper head of the fuel canister has one large diameter flat catalyst bed containing 100 grams of catalyst on the inner surface. The knockout and filter canisters have two syninetrically located beds containing 50 grams each of catalyst in the upper heads. All catalyst cartridges are welded in place and structurally designed to remain intact and functional, provided they are not immersed, during any postulated drop accident. The catalyst material is covered by a retainer screen that holds it in place but allows free diffusion of gas to the catalyst surface and diffusion of water vapor away from the catalyst.

Based on a review of the licensee's evaluation and available literature on radiolytic decomposition, the NRC staff has determined that the maximum

theoretical gas generation rate has been predicted with considerable

, conservatism. The staff has further determined that the designed ~ catalytic

{ recombiners have acceptable margins of safety and provide reasonable assurance I

that combustible gas mixtures will not develop in the filled canisters af ter dewatering.

CANISTER OPERATIONS The fuel canisters are designed to be inserted into the reactor vessel where they are supported by either the' canister positioning system or the single canister support bracket. Pieces of fuel debris are picked up by various types of defueling tools and placed into the canisters. Methods of debris placement will be controlled by procedures approved by the NRC staff and will ensure that dropped debris will not impose impact loads on the bottom support plate in excess of those designed. The knockout canisters are inserted in the canister positioning system where they are connected to the fuel debris vacuuming system. The filter canisters are installed in either the defueling water cleanup system where they are supported by the storage racks in the fuel transfer canal and spent fuel pool, or they are installed in the fuel debris vacuuming system in the reactor vessel. They can also be used in the final canister dewatering system in the spent fuel pool. The canisters will be filled with core debris in their respective processing systems. They are dcsigned to be filled to a maximum dewatered weight of 2800 poun6 with an allowance of 5 percent of the canisters to be 5 percent ovemeight or 2940 pcunds. The worst case loaded and flooded canister could weigh 3500 pounds.

The canisters will be weighed during processing to ensure they are maintained within the design weight limits. When filling is complete, the upper head is bolted onto the fuel canister. The process connections are plugged on the filter and knockout canisters. They may then be partially dewatered in the reactor vessel to expose suf ficient catalyst to control the gas buildup. Two relief valves will then be installed. A 25 psig relief is installed on the inert gas purge connection and a 150 psig ASME code relief valve is iastalled on the dewatering connection. These relief valves are to protect the canisters from overpressurization in the unlikely event of catalyst failure or in the event of canister storage prior to dewatering. The canisters will then be transferred to the ' A' spent fuel pool for storage, final dewatering, and preparation for shipment. Both initial dewatering in the reactor vessel and final dewatering in the spent fuel pool will involve water removal by purging the canisters with argon, an inert cover gas. They will be lef t pressurized to about 13 psig with the inert gas. After final dewatering and purging, the canisters will be monitored for a sufficient period of time to verify that the catalytic recombiners are functioning.

The staff has evaluated the consequences of several situations in which gas generation may occur in a canister.

If a canister is filled solid with debris and water, the recombiners will be ineffective. This will result in pressure buildup and periodic lifting of the relief valves. This will occur in a short period of time (about 40 hours4.62963e-4 days <br />0.0111 hours <br />6.613757e-5 weeks <br />1.522e-5 months <br />) with the maximum theoretical gas generation rate. If the 25 psig relief valve fails to operate, the internal pressure will reach the setpoint of the 150 psig relief valve in about nine days. Lif ting of the relief valves is considered to be acceptable since the canisters are stored underwater. The quantity of flammable gas mixtures vented by relief valve actuation will be small and readily dispersed by venting into the water and diluted by the surrounding atmosphere. Thus, no fire hazard should exist. Activity released to the water by relief valve lif ting is readily removed by the defueling water I

cleanup system. Failure of both relief valves is considered very unlikely since they are independent.of one another and installed in such a manner that they are not subject to a common mode failure. If, however, both were to fail, it would take nearly one year for canister internal pressures to reach the yield stress on the canister shell. This is not considered credible since canister dewatering should take place before this time has elapsed.

Following dewatering and inerting of a canister, its internal pressure should remain stable. If, however, the recombiners fail to operate, the pressure will increase. Assuming failure of the recombiners, it will take about one week to achieve a flammable mixture in the canister. Ignition of this mixture is unlikely, but if it were to occur the canister yield stresses would not be exceeded. It will take about one month to reach the set point of the 25 psig relief valve and about one year to reach 150 psig relief valve setpoint. This is assuming the minimum canister void space of 96 liters and a gas generation rate of 0.114 liters per hour. Lifting of the relief valves in these cases is I of no safety consequence as previously discussed above.

The licensee's evaluation presented an analysis of the consequences of ignition of the vented gases if relief valve actuation were to occur while a canister is in the transfer shield. The staff review of that evaluation concurs that the consequences of such an event pose no significant risk.

The staff has determined that the canister design is compatible with the scope of operations discussed in the licensee's Technical Evaluation Report.

CONCLUSION The NRC staff has performed a safety review of the design of the proposed defueling canisters. This review consisted of evaluation of the canister structural design, evaluation of the licensee's criticality analysis, evaluation of the canister's combustible gas control features, and evaluation of the affects of postulated accidents and abnormal conditions. Based on the review, the canister design and their proposed operations do not pose a significant risk to the occupational work force or the public. The defueling canisters, which are necessary to support planned defueling activities, do not i

present the possibility of any accident not previously analyzed nor do they change the consequences of, or likelihood of any previously analyzed accident.

Margins of safety as previously analyzed are not reduced. The staff concludes that the canister design does not necessitate additional changes to the plant Technical Specifications and does not constitute an unreviewed safety question. The scope of activities and the associated environmental impact of the defueling canisters as discussed in Defueling Canister Technical Evaluation Report are within those previously considered in the PEIS. We therefore approve the design of the defueling canisters. Use of the canisters is contingent upon our approval of those procedures subject to Technical Specification 6.8.2. Operations to fill the canisters with core debris will also be contingent upon our approval of the Early Defueling Safety Evaluation Report.

-, ._y , , _ _ . . _ .