Text

Potential Radiological Impact of Tornadoes on the Safety of Nuclear Fuel Services West Valley Fuel Reprocessing Plant
	ML20031F005
Person / Time
Site:	West Valley Demonstration Project
Issue date:	09/30/1981
From:	Andrae R, Holloway L; OAK RIDGE NATIONAL LABORATORY
To:	; NRC OFFICE OF NUCLEAR MATERIAL SAFETY & SAFEGUARDS (NMSS)
References
	CON-FIN-B-0102, CON-FIN-B-102 NUREG-CR-1530, NUREG-CR-1530-V01, NUREG-CR-1530-V1, ORNL-NUREG-80-1, NUDOCS 8110190022
	Download: ML20031F005 (100)
	v • d • e

{{#Wiki_filter:. l NUREG/CR-1530 f ORNL/NUREG-80/1 Vol.1 ry Potential Radiological Impact of Tornadoes on the Safety of l Nuclear Fuel Services' West Valley l Fuel Reprocessing Plant l-Tornado Effects on Head-end Cell Airflow i i l Prepared by L. J. Holloway, R. W. 'Andrae l Oak Ridge National Laborctory l, { Prepared for U.S. Nuclear Regulatory a w/ o) Commission e f6 fn!fl/21)i c 13 "ocr a reers.m j "iFJr:P / j veggp l l [ 8 0 2 810930 j, ca-1:30 e eos l

NOTICE This report was prepared as an account of work sponsored by an agency.of the United States Government. Neither the ~ United States Government nor any agency thereof, or any of their employees, makes any warranty, expressed or implied, or - assumes any legal liability or responsibility for any third party's a use, or the results of such use, of any information, apparatus - 4 - product or process disclosed in this report, or represents iat ' -ry

its use by such' third party would not infringe privately owned rights.

h

l

,1 4 8 Available from A sy GP0 Sales Program l Division of Technical'Information~and Document Control -U..S. Nuclear Regulatory Commission-Washington, D. C.;20555 Printed ~ copy. price:-$5.50- -and National Technical-Information Service Springfield, Virginia 22161:

1 NUREG/CR-1530 ORNL/NUREG@/1 Vol.1 Potential Radiological Impact of Tornadoes on the Safety of ,b Nuclear Fuel Services' West Valley Fuel Reprocessing Plant ~ Tornado Effects on Head-end Cell Airflow Manuscript Completed: July 1981 j Date Published: September 1981 Prepared by L. J. Holloway, Oak Ridge National Laboratory i R. W. Andrae, Los Alamos Scientific Laboratory Oak Ridge National Laboratory Oak Ridge, TN 37830 r Prepared for Division of Fuel Cycle and Material Safety i Office of Nuclear Material Safety and Safeguards U.S. Nuclear Regulatory Commission l Washington, D.C. 20555 / NRC FIN B0102 s f C' i i l r

1 TABLE OF CONTEMS Page LIST OF TABLES V LIST OF FIGURES vii ACKNOWLEDGMENTS ABSTRACT ~ 1. INTRODUCTION 1 2.

SUMMARY

2-1 3. ANALYTICAL PROCEDURES 3-1 3.1 General Procedures 3-1 3.2 Detailed Analytical Techniques 3-3 4. HEAD-END CELL AIRFLOW ANALYSIS 4 4.1 Potential Release Pathways 4-1 -- 4.2 Modeling Techniques 4-6 4.2.1 Case 1. PMC manipulator port pathway 4-9 4.2.2 Case-2. CPC-EDR pathway 4 4.2.3 Case 3. GPC to operating aisle pathway 4-9 4.2.4 Case 4. GPC-MSM building pathway 4-10 5. HEAD-END CELL AIRFLOW UNDER TORNADO CONDITIONS 5-1 5.1 Air Discharge Volumes for Cases Analyzed 5-1 5.2 Cases with Potential for Transport of 5-6 Radioactive Material i 5.2.1 Case 1. PMC. manipulator port pathway 5-7: 5.2.2 Case 2. CPC-EDR pathway 5-10 5.2.3 Case'3. GPC to operating aisle pathway 5-15 5.2.4 Case 4. GPC-MSM building pathway 5-15 6. TORNADO-INDUCED AIR VELOCITIES IN HEAD-END CELLS 6-1 6.1 Modeling Techniques 6-1 6.2 Airflow Velocity - Time Histories for ~ Cases 6-7 g Anslyzed CONCLUSIJNS 17 '7. REFERENCES' R-1_

Appendix A.

DESCRIPTION OF TVENT ' A Appendix B. DETAILED ANALYTICAL TECHNIQUES 'B-1 iii

LIST OF TABLES Table Page Number Title Number 3.1 Tornado parameters based on DBT-78 tornado model 3-5 o 4.1 Flow characteristics of NFS cell exhaust model _4-8 ' for TVENT-computer analysis et 5.1 Summary of pathways with discharges of air from 5-3 radioactively contaminated areas (head-end cells) during postulated tornado strike 5.2 Air volumes discharged from cells during 5-4 postulated tornado strike 5 O .l J J e W V i l ,,, ~.. _, _ _.

LIST OF FIGURES Figure Page Number Title Number 3.1 Tornado pressure functions for input to TVENT. Based 3-6 on tornado model DBT-78. 4.1 NFS cell configuration and cell ventilation path, plan 4-2 and sections 4.2 Dimetric cutaway of NFS head-end cells 4-3 4.3 NFS cell configuration, block diagram 4-5 4.4 Schematic model of NFS cell exhaust system fcr TVENT 4-7 computer analysis 5.1 Tornado-generated airflows in manipulator sleeve 5-8 pathway for case lAl (100-mph tornado) and case 1B1 (200-mph tornado) as predicted with TVENT. Exte rior building walls intact. 5.2 (a) Tornado-generated airflows in manipulator steeve 5-9 pathway for case ICl (300-mph tornado). (b) Tornado-generated airflows in branches leading into chemical process cell for case 1C1 (300-mph tornado). 5.3 Tornado-generated airflows in manipulator sleeve path-5-11 way with loss of exterior building walls simulated for case 1A2 (100-mph tornado) and case 1B2 (200-mph to rnado). Airflows predicted with TVENT. 5.4 Tornado-generated airflows in manipulator sleeve 5-12 . pathway as predicted with TVENT for case IC2 (300-mph to rnado). Loss of exterior building walls simulated. 5.5 Tornado-generated airficws in pathway from chemical 5-13 = process cell through equipment decontamination room to atmosphere for case 2A (100-mph tornado) and case 2B (200-mph tornado). Exterior building walls intact. Airflows predicted with TVENT. 5.6 Tornado-generated airflows in pathway from chemical 5-14 process cell through equipment decontanination room to atmosphere for case 2C (300-mph tornado) with ( exterior building walls intact. -Airflows predicted with TVENT. i I t vii l

i viii Page Figure Number Title Number 5-16 5.7 Tornade~ generated air.~1ows in pathway from general-purpose cell through ports to the operating aisle, case 3, for 100- and 200-mph tornadoes as predicted with TVENT. Exterior building walls intact. 5-17 5.8' Tornado-generated airflows in pathway from general-purpose cell through ports to operating aisle. (a) For 300-mph tornado with building intact; (b) for 100-mph tornado with building lost. 5-18 5.9 Tornado-generated airflows in pathway from general-purpose cell through ports to operating aisle, case 3, for 200- and 300-mph tornadoes as predicted with TVENT. 5-19 5.10 Tornado-generated airflows from general-purpose cell through main exhaust duct to master-slave manipulator building and atmosphere, case 4, for 100- and 200-mph to rnadoes. Airflows predicted with TVENT. 5.11 Tornado-generated airflows in pathway from general-5-20 purpose cell through main exhaust duct to master-slave manipulator building and atmosphere, case 4c, for 300-mph tornado as predicted with TVENT. 6.1 (a) Isometric drawing of head-end cells; (b) SOLA-ICE 6-2 model of general-purpose cell 6.2 (a) Isometric drawing of head-end cells; (b) SOLA-ICE 6-4 model of process mechanical cell 6.3 Velocities at boundaries of general-purpose cell for 6-5 cases 1B1, 200-mph tornado, and 1C1, 300-mph tornado, as predicted with TVENT 6.4 Velocities at boundaries of process mechanical cell 6-6 for cases 1B1, 200-mph tornado, and 1C1, 300-mph tornado, as predicted with TVENT 6.5 Streamlines through general-purpose cell for case 1Al, 6-8 100-uph tornado, as determined by SOLA-ICE code. (Normal conditions from 0 to 10 s; tornado impact at 10 s). 6.6 Streamlines through general-purpose cell for case 4C, 6-9 300-mph tornado, as determined with SOLA-ICE Code

ix Figure Page Number Title Number 6.7 Transient. velocities along floor of general-purpose 6-10 cell for case 1C1, 300-mph tornado. (a) Composite of all velocity vs time curves; (jl-d) velocity-time histories at specific locations along the floor, o 6.8 Transient velocities along floor of general-purpose 6-11 q cell for case 4C, 300-mph tornado. (3) Composite of all velocity vs time curves; (b_-d_) velocity-time histories at specific locations along the floor. 6.9 Plots of velocity vectors through process mechanical 6-12 cell for case 1A1,100-mph tornado, as determined with SOLA-ICE. (Normal conditions from 0 to 10 s; tornado impact at 10 s). 6.10 Velocity vectors through process mechanical cell for 6-13 case 1B, 200-mph tornado, as determined with SOLA-ICE. i 6.11 Transient velocities along floor of process mechanical 6-15 cell for case 1B1, 200-mph tornado. (a) Ccmposite of all velocity vs time curves; (b-d) velocity-time histories at specific locations.along the floor. l 6.12 Transient velocities along floor of process mechanical 6-16 cell'for case 101, 300-mph tornado. (a) Composite of all velocity vs time curves; (jr-d) velocity-time histories at specific locations along the floor. B.1 Schematic model of NFS cell exhaust system for TVENT B-2 computer analysis d f r

l ACKNOWLEDGMENTS We are very pleased to acknowledge the continued advice of l A. T. Clark, L. S. Person, and C. J. Haughney, of the Office of Nuclear Material Safety and Safeguards, and of R. F. Abbey, Jr., of the Office of Nuclear Regulatory Research, U.S. Nuclear Regulatory Commission,. ' during the development of this report. Calculations to determine quantities of radioactive materials released from the cells during a tornado were performed by W. Davis, Jr., Chemical Technology Division, author of Part II of this report. Finally, we thank Alden Pierce, Clyde i ( Alday, and their colleagues at Nuclear Fuel Services for providing us with details of the head end cells of the plant and for the new data on t dose rates in these cells. b I-l ? m-e 9 xi' i i

POTENTIAL RADIOLOGICAL IMPACT OF TORNADOES ON THE SAFETY OF NUCLEAR FUEL SERVICES' WEST VALLEY FUEL REPROCESSING PLANT I. TORNADO EFFECTS ON HEAD-END CELL AIRFLOW L. J. Holloway R. W. Andrae .s l l ABSTRACT l l This report describes results of a paratetric study of the impacts of a tornado-generated depressurization on i airflow in the contaminated process cells within the l l presently inoperative Nuclear Fuel Services fuel re- ) processing facility near West Valley, N.Y. The study involved the following tasks: (1) mathematical modeling of installed ventilation and abnormal exhaust pathways from the cells and prediction of tornado-induced airflows in these pathways; (2) mathematical modeling of individual cell flow characteristics and prediction of in-cell velocities induced by flows from step 1; and (3) evaluation of the results of steps 1 and 2 to determine whether any of the pathways investigated have the potential fo'r releasing quantities of radioactively contaminated air from the main process cells. The study has concluded that in the event of a torna lo strike, certain pathways from the cells have the potential to release radioactive materials to the atmosphere. Determination of the quantities of radioactive material released from the cells through pathways identified in step 3 is presented in volume 2 of this report. 1. INTRODUCTION The primary objective of this study is the determination of octential radiological impacts of a tornado strike at the presently inoperative Nuclear Fuci Services' (NFS) fuel reprocessing plant located near West Valley, N.Y. Considerable quantities of radioactive materials remain in the head-end cells [60 t o 600 kg in the process mechanical cell (PMC) and 50 to 500 kg in the general-purpose cell (GPC)] (see Part II of this reportl) as a result of previous reprecessing operations. There is concern about the ability of the cells and ventilation system to adequately provide containment of this material when subjected to tornado-induced forces. The present report involves analysis of the t 1-1

1-2 f actors that could lead to transport of radioactive material f rom the cells to the atmosphere during a tornado. In Part II of this report,I the determination of quantities of radioactive material that could be discharged from the cells under tornado accident conditions is described. Nuclear Fuel Services, Inc., established the first commercial facility for processing spent nuclear fucis at West Valley, N.Y., on property of the Western New York Nuclear Services' Center (WNYNSC) in 1966. The plant was operated until 1972, during which time 640 metric tons of irradiated fuel was processed. The plant was originally shut down in 1972 because of plans to expand capacity; on September 22, 1976, NFS announced its decision to withdraw f rom the nuclear fuel reprocessing business. The history of the WNYNSC and the interectione of local residents and local, state, and federal agencies have been docu me n ted. 2 The Nuclear Regulatory Comm'ssion (NRC) has requested an analysis of the effect of a tornado strike at the reprocessing plant.3 G mama-ray dose rate measurements have enabled estimates of the quantities of radioactive materials in the PMC, the GPC, and the chemical process cell (CPC). Because of the massive thickness of cell walls and floors, no significant structural damage to the head-end cells is expected during a tornado str'ike. The transient depressurization associated with the vortex motion of air within a tornado could function to draw radioactive materials from the head-end cells through insufficiently scaled or unfiltered pa thways. Several pathways of this type were lef t in the cell walls when equipment was removed during plant shutdown. Backflow pathways could potentially be the most severe because no filtration would be provided. This study involves determination cf the potential for release of radioactive materials through the identified pathways during a tornado (Part I) and estimation of the quantities of material that would be released (Part II). The following tasks were involved in this study. 1. identification of all pathways through which air could flow f rom the cells (Part I); 2. prediction of tornado-induced airflows through these pathways (step

1) from the cells (Part I);

3. prediction of tornado-induced air velocities over surf aces within the cells where deposited radioactive material is likely (Part I); 4 estimation of quantities of radioactive materials and size distribution of these materials within the cells (Part II);

1-3 5. determination of quantities of radioactive materials that would be resuspended f rom the cell floor by higher than normal tornado-induced air velocities (from step 3) in the cells (Part II); 6. determination of quantities of radioactive materials that would be drawn f rom the cells to the atmosphere through unfiltered pathways by tornado-induced airflows (combination of steps 2 and 5) (Part 11). In an effort to learn how radioactive materials might be transferred through ventilation systems under tornado conditions, the U.S. Department of Energy sponsored a study by Los Alamos Scientific Laboratory of tornado-induced airflow in nuclear facilities.4 The LASL study led to the development of the TVENT5 computer code designed to predict one-dimensional tornado-induced flows and pressures in the ventilation system of a nuclear f acility. The NRC has used TVENT to determine the potential for radioactive particulate release f rom the NFS facility. TVENT was used in step 2 of the present analysis to predict the dynamic characteristics of air movement within the NFS plant caused by a range of hypothetical transient conditions. Under the same DOE sponsored study, another computer code, SOLA.CE,6 was modified by LASL to utilize the TVENT output of airflows in the cell ventilation system. The output served as boundary conditions for calculating the detailed two-dimensional, tornado-induced air velocities within the cells over surfaces where deposited radioactive material is likely. Analysis of the NFS facility was performed using the SOLA-ICE code to predict air velocities in the cells (step 3 of this analysis). The analysis was performed as a parametric atudy, using a range of tornadoes postulated to strike the NFS facility. Tornadoes with maximum wind speeds of 100, 200, and 300 mph were considered in this analysis. No attempt was made in this study to assess the likelihood of occurrence of any particular tornado. One topic not analyzed in this study concerns the structural integrity of ptocess building walls and doors under tornado-induced loadings. The analysis was performed assuming for one condition that exterior walls and doors would remain intact and for the other condition that they would he destroyed by the tornado. The following section is a Lummary of the basic analysis involved in Part I, that is, the estimation of tornado-induced head end cell airflows for each pathway. Conclusions concerning the pathways with greatest potential for release of materials are also presented.

2.

SUMMARY

Analysis of the f actors that could lead to transport of radioactive materials f rom the NFS head end cells to the atmosphere during a tornado strike involved the development, implementation, and integration of a variety of analytical tools. Investigation of a aunber of distinct phenomena was required in areas including site meteorology, mechanical design of airflow systems, compressible and incompressible flow analysis, radiation shielding analysis, particle size analysis, aerosol science, atmospheric dispersion, and uptake and dose-rate modeling. An unusual aspect of this type of analysis, involving consideration of airflows from the facility, is that unless it can be determined at the start that no pathways for airflow from contaminated areas exist, all phenomena must be investigated in order to reach conclusions that can be substantiated. No single factor will determine whether a release potential exis ts. The analytical procedures developed for the entire study are described in Section 3. In Section 4, the determination of potential airflow pathways f rom the cells and the modeling of these paths for computer analysis are described. Two visits to the site as well as examination of detailed facility drawings were required in order to identify all possible pathways. All potential pathways were investigated including those involving reverse flow. Preliminary analysis indicated that some pathways did not have significant release potential because of substantial airflow restrictions or filters, and those were eliminated from further consideration. Several pathways were discovered which are unique to this f acility because of its inoperative status. In the PMC and the GPC, a number of manipulatora had been removed from the cell walls af ter plant shutdown and replaced by substantial plugs or covered with plywood or in a few cases with cardboard. Forces accompanying a tornado either f rom high winds or f rca depressarization would be sufficient to pull covers f rom some of the ports and possibly draw contaminated air f rom the cells. Another unusual pathway involved an open 8-in.-diam duct in the GPC wall through which a hose to provide carbon dioxide for fire protection pa aed into the cell. The normal cell fire protection system had been shut down in order to reduce personnel and equipment usage requirements for the inoperative f acility. Modeling of these pathways f or computer analysis with the TVENT code required slight modifications in establiched procedures for utilizing the code in order to model loss of covers over ports in the cell walls during a tornado-induced pressure transient. These procedures are described in Section 3. 2-1

2-2 In Section 5, results of the TVENT analysis are presented in the - form of airflow rates and total volumes of air discharged through each pathway f rom the cells. The predictions of airflow for 100, 200, and 300 mph tornadoes were examined to determine whether air f rom the cells would flow to the atmocphere through unfiltered pathways. This t'.rflow would be opposite to the normal' direction of flow, which is f rom corridors and rooms adjoining the cells into and through the cells to a filtration system, exhaust blowers, and stack discharge. Conclusions f rom this analysis are that for certain of the pathways and tornado wind speeds examined, the tornado depressurization would not be suf ficient to overcome the induced draf t f rom exhaust blowers, which acts to pull air into and through the cells; for these cases, no air could be drawn f rom the cells to the atmosphere through unfiltered pathways, and no potential for release of radioactive materials exists. For a f ew' other cases examined, analysis indicated that tornado-induced forces have the potential to draw unfiltered air f rom the cells. The pathways with the greatest potential for release of unfiltered air were identified as seven unsealed manipulator ports in the walls of the PMC.- The analysis also indicated that several other pathways had minor release potential. Af ter identification of pathways and tornado ' conditions with release potential (by TVENT analysis), it was necessary-to determine whether air velocities generated within the cells for these cases would be sufficient to reentrain radioactive particles f rom contaminated surf aces and transport this material f rom the cells. The SOLA-ICE computer code was used for this analysis by LASL, and the results are presented in Section 6. SOLA-ICE predicted velocities over the cell floors were used as described in Part II to predict quantities of material released from the cells. O 9

l 3. ANALYTICAL PROCEDURES A primary purpose of this report is the documentation of analytical techniques developed for evaluating tornado effects on the NFS facility. Techniques for analyzing the effects of natural phenomena on nuclear fuel cycle f acilities are still evolving. In this section an attempt is made to document the step-by step procedures developed for this task in order that the techniques may be utilized in future evaluations. 3.1 General Procedures A structural evaluation to determine loss of facility integrity and damage under extreme conditions generated by natural phenomena is normally tha first step in an analysis of this type.7,8,9 For the NFS f acility, the assumption was made at the outset that the 5-to 6-f t-thick reinforced concrete head end cells containing radioactive materials would not be damaged by tornado-induced forces. This is a reasonable assumption f or reinforced concrete enclosures of this size. It was assumed that the cells would remain intact, with only small penetrations to the cells unprotected f rom tornado forces (i.e., ports for manipulators, fire protection ports, etc.). It was also assumed that the exterior building walls constructed of concrete block and sheet metal would sustain damage under tornado conditions such that the head end cells within would be exposed to tornado-induced forces. The assumed level of damage to the exterior walls and doors ranged from loss of portions of the sheet metal or exterior doors under lesser tornadoes to complete collapse of the exterior building when subjected to the highest tornado considered. In the ongoing NRC evaluations of tornado effects on plutoniur facilities,7,8 analysis indicates that the primary containment structures (glove boxes) would not remain intact, and the radioactive materials within would be dispersed to the surroundings by tornadic winds. In the case of the NFS facility, primary containment is provided by the massive reinforced concrete cells, which would remain intact e during a tornado. The only mechanism which could induce a release of radioactive materials under tornado conditions is the transient depressurization associated with a tornado. The depressurization created by tLe vortex motion of air within a tornado could induce higher than normal airflows, flow reversals, and pressure dif ferences in the cells and draw air containing radioactive particles out of the cells through existing unfiltered penetrations in the cell walls. 3-1

3-2 The following is a summary of the steps involved in the process of determining the potential for discharge and the quantities of cadioactive materials which could be discharged from the NFS facility under tornado accident conditions. 1.- Examine the f acility to determine pathways through which unfiltered air containing radioactive materials could be drawn f rom the main process cells to the surrourdings by a depressurization associated with a tornado. 2. Model the cell ventilation pathways and other identified airflow pathways from step 1 for computer analysis with the TVENT code. 3. Characterize a range of tornadoes for the. 3eographic location of the facility by the following parameters: maximum wind speed, rotational and translational speeds, radius of maximum winds. total pressure drop, and rate of pressure drop. 4. Utilizing the TVENT code, predict transient flows and pressures in cells and airflow pathways induced by the transient depressurizatior, accompanying a tornado. 5. Examine tornado-induced airflows in the system (step 4) to determine if air potentially containing radioactive materials could flow f rom the head end cells through unfiltered pathways to the atmosphere. The volumes of air flowing through each pathway over a period of time and the flow velocities in each pathway are output from TVENT (step 4).

6.. Model the individual cells f rom which contaminated air could-flow (step 5) and analyze the model with the SOLA-ICE-computer code.

SOLA-ICE utilizes airflow velocities through each path f rom the cell (TVENT output) as input for predicting two-dimensicnal ~ transient velocities within the cell. The velocities near surfaces where deposited radioactive material is likely are the quantities of interest. 7. Estimate quantities of radioactive materials deposited on surfaces of head end cells, utilizing dose rate measurements f rom the cells, an estimate of the typical or average fuel processed at NFS, the 10 gamma-ray spectrum of the reference fuel calculated by the ORIGEN code, and dose conversion factors determined from' the SDCII gamma-ray shielding code > (Part II of this report). 8. Estimate the particle size distribution of radioactive materials on surf aces of cells, using particle size information f rom processing of similar fuels (Part II of this report). D m J

3-3 9. Determine terminal settling ar.d Phreshold f riction speeds (air speeds required to 11f c and cor.y, reentrain, particles from a surface) for the size distribution of radioactive materials in the cells (Part II of this report). 10. Using the tornado-induced near floor air speeds in the cells predicted with the SOLA-ICE code (step 6), the particle size distribution of materials in the cell (step 8), and the terminal settling and threshold friction speeds (step 9), estimate the quantities of radioactive material that would become airborne by tornado-induced airflows (Part II of this report). 11. Using volumes of air discharged from the cells as predicted by the TVE!;T analysis (step 4), determine whether radioactive particles reentrained from the cell floors would be discharged to the atmosphere. The amount of material discharged is the source term (Part II of this report). The final step in this type of analysis would be determination of the ef fects of the release predicted in step 11 on operating personnel, the public, and the surrounding environment, using appropriate atmospheric dispersion, uptake, and dosage models. This evaluation was not included in the scope of the present study and will not be addressed in this report. Part I of this report addresses in detail the analysis and results of steps 1 to 6 involving the effects of a tornado on airflow in the l head-end cells. Part II of the report addresses steps 7 to 11 of the analysis, including predictions of the quantities of material released during a tornado strike. In the following section, details involved in utilizing the analytical methods described in steps 1 to 6 are presented. Detailed techniques utilized in steps 7 to 11 are presented in Part II of this report.1

3. 2 Detailed Analytical Techniques The computer ccde selected by the NRC staff for use in this evaluation is TVENT,5 developed at IASL.

This code was devel.oped by R. W. Andrae, K. H. Duerre, and W. S. Gregory of LASI as a tool for predicting responses of ventilation systems to tornado-induced pressure transients. A description of TVFNT along with its capabilities and limitations is included in Appendix A. E

3-4 In modeling a ventilation system for analysis with TVENT, step 2, it is necessary to determine the flow characteristics of the system under normal operating conditions. Operatit g pressures and flows and the location of ventilation components such es valves, filters, and blowers are required for input to TVENT and were obtained for the NFS facility f rom drawings of the head end ventilation system and f rom instrumentation located at the plant. Flow characteristics of the system under normsi operating conditions are presented in Sect. 4.2. In modeling the ventilation system and other exhaust pathways at the NFS f acility, several alterations to specific flow characteristics of the normal operating system were required in order to simulate behavior of certain portions of the system during a tornado. Since TVENT performs a steady state analysis of a modeled system to establish system resistances before the transient analysis, it was necessary to g develop several techniques for modeling the transient behavior of parts of the system while not altering resistances in the remainder of the system. TVENT assumes a fixed geometry in the analysis of a ventilation system; thus, techniques were required for modeling partial and complete f ailure of the exterior walls of the NFS building. The simulation of partial wall failure was accomplished by including a small resistance to airflow for the exterior walls it the system model. Modeling complete exterior wall failure was accomplished by utilizing the " restart" option of TVENT. A steady state analysis of the system model with normal operating flow characteristics is performed assuming the exterior walls are intact; that is, resiscances are included in the model for the exterior walls. Using the system resistances established in the steady state analysis, a transient analysis is performed but with resistances provided by the exterior walls deleted. Thus the ef fect of loss of exterior walls on the ventilation system can be determined without altering resistances to airflow within the system. More specific details involved in this technique are discu sed in Appendix B. A 'similar technique was used to model the removal of plywood and cardboard covers f rom exhaust pathways in the cells when pressure gradients across these covers reached a certain point. Calculations (see Appendix B) indicate that a pressure dif ferential of 0.4 in. H O 2 (0.014 psi, cell pressure positive relative to carroundings) would be sufficient to dislodge plywood covers over manipulator ports in the PMC and cardboard covers over ports in the GPC. This pressure differential was predicted by TVENT to occur for the 100, 200- and 300 mph tornadoes at 2.7, 0.9, and 0.6, respectively, af ter the start of the tornado-induced transient. At these times, each case was restarted with altered resistances to reflect loss of the covers, and the transient analysis was continued.

3-5 Boundary points are present in the f acility such as exterior doors or opeaings to the atmosphere through which air would flow out of the building if a positive pressure dif ferential were applied at these points. Normally, pressure dif ferentials across doors and vents in exterior walls of the building are negative with respect to the atmosphere (interior corridors of the building are held at -0.05 to -0.1 in. H O with res; et to atmospheric pressure). A low pressure area such 2 as that occurring within a tornado could create a positive pressure dif ferential across vents and doors, causing air to flow out of the building through unfiltered pathways in a direction opposite to normal flow. When using TVENT, it is necessary to define a time-dependent pressure function representing depressurization effects which would occur as a tornado crosses a boundary point (step 3). This pressure function is used as a forcing function by TVENT in predicting pressure and flow transient responses within the melaled system ar a result of a low pressure function at the boundary points. An analytical tornado model developed by Fujita,12 DBT-78, wsu selected by the NRC staff as appropriate for determining pressure functions for 100, 200, and 300-mph tornadoes. Parameters necessary for determining pressure functions based on tornado model DBT-78 are given in Table 3.1. Table 3.1. Tornado parameters based on DBT-78 tornado modela Maximum Radius Total Pressure tangential velocity (m) pressure drop drop rate (mph) (psi). (psi /s) 100 56 0.18 0.03 (4.9 in. H O) (0.9 in. H 0/s) 2 2 200 108 0.63 0.12 (17.6 in. H O) (3.4 in. H 0/s) 2 2 300 157 1.3 0.26 (36.3 in. H O) (7.2 in. H 0/s) 2 2 aSee Ref. 12. Pressure transients derived from the parameters given in Table 3.1 are shown in Fig. 3.1 for the 100, 200, and 300 mph tornadoes. An assumption made in deriving these pressure functions was that the maximum total pressure drop within the tornado would last for 4 s as the

3-6 ORNL-DWG 81-13177 TIME (s) 5.o io.o i5.0 00 i i ,,,,, 3 -10.0 Q -15.0 I b 200 mph tu -20.0 T D U) U) Id T' -25.0 -30.0 l M AXIMUM TOTAL l TAN GE N TIAL PRES $URE l VELOCITY OROP 100 mph 4 9 in H2O -350 \\ g o,, ,,,3 2 00 mph 17 6 in H O 2 300 mph (o.63 esi) 300 mph 36.3 in. H O 2 (13 psi) Fig. 3.1. Tornado pressure functions for input to TVENT. Based on tornado model DBT-78 (ref. 10). i

3-7 tornado crossed the facility. Calculations supporting this assumption based on the translational velocity of the tornado, the tornado diameter, and the building length are described in Appendix B. A primary concern in analyzing a ventilation system for tornado effects is the effect of internally generated pressure transients and pressure dif ferentials on HEPA illters within the system. High pressure dif ferentials or tornado-induced pressure shocks could cause f ailure of the filters, resulting in a release of unfiltered air through the normal ventilation exhaust pathway. Tornado-induced pressure dif ferentials across HEPA filters in the NFS ventilation system as predicted by TVENT were not suf ficient to cause filter f ailure. The highest pressure dif ferential predicted across a HEPA filter for all the cases and conditions analyzed was 17.5 in. H 0, which would occur across a filter in the air supply path for 2 13 of HEPA filter f ailure the CPC. An investigation conducted by LASL during overpressurization has indicated that the average break pressure for the weakest filter tested was 1.32 psi (36.5 in. H O) with a 2 standard deviation of 0.22 psi. With two standard deviations, the break pressure is still 0.88 psi (24.4 in. H O), which is above the highest 2 pressure dif ferential occurring in the NFS analysis. Thus it was determined that all HEPA filters within the NFS facility would remain intact under tornado accident conditions, provided the structure housing the filter remained intact and any releases of air through these paths (in the normal or reverse directions) would be filtered. Emphasis in analyzing release pathways from the head-end cells was thus placed on unfiltered patnways. In analyzing the NFS facility for tornado effects (step 4), two different modes of tornado impact were postulated, and both were applied to each casc being considered. In the first mode, the tornado would simultaneously impact all facility portals (boundary point.) through which air is both supplied to and exhausted f rom the ventilation system. In the second mode, the tornado would strike the portion of the building in which boundary points for air supply to the ventilation system are located while not striking the point of ventilation exhaust located approximately 100 ft away. In the main process building, air is supplied to the head end ventilation system by stairwells located at the building perimeter. Air intakes for the building are supplied by outside air vents in the stairwells. Since the stairwells supplying air are interconnected, ventilation air for the CPC would originate from the same source as that for ventilation air for the PMC or GPC; therefore, the low pressure area within a torr. ado would af fect all three supply points equally and simultaneously. Under normal operating conditions, ventilation air f rom the process cells is exhausted through an underground duct to a filter house containing a bank of roughing filters and HEPA filters. Two 7000 cfm t

3-8 f ans then f orce the filtered air to an exhaust stack located on the building roof, where it is released to the atmosphere. Since the filter house is located approximately 100 f t f rom the stairwells which supply air to the ventilation system, it is possible, because of the random behavior and movement of tornadoes, for the ventilation supply points to be af fected by the tornado, while the exhaust point is not. Analysis of all postulated release pathways for the two modes of tornado impact indicated that the second mcde predicted higher volum s of unfiltered air released from the cells than did the first mode (step 5). Higher volumes would be released during the second mode, because in the first mode a lower back pressure f rom the tornado is applied at the normal exhaust point and eventually lowers the cell pressures, creating less of a pressure dif ferential to force air out of the unfiltered exhaust pathways. To simplify presentation of results, only the second impact mode will be considered, since it presents the " worst case" for this analysis The procedures outlined in this section may be useful in evaluating tornado effects on other nuclear facilities and should be modified to suit the particular f acility being analyzed. More detailed modeling techniques and analysis results are described in the renainder of this report and in Part II.I

h N N 4 + n

N ',*. ~ '

CHEMICAL PROCESS CELL ~T g y' .., :..q., zTHtSCWG. i / [a 44 s t / /., l "f f )1 l l e curLisE oF GENERAL b MdIurTv

Stafit ATioN,

re-Pulif 'E CELL BELow surety uuct + gom (uoce.atTuaw 3 ratus) j ci actu suerty 6aouso rtoon LEVEL y t, N N t PROCF 55 MECHANIC AL CEt1*, i:: .[h Tuts OWG. 3 '" 3' ' t.... ' ( E,,,ggy pecy s UROM GPC BELOW) T u t *> OW Ge. PARTlAL PL AN - GROUND Floor EL.831 0* M l u c- . p 4 n..#. EL ll4*- 6* c ae 4 PMC ^ \\ -_. s110.... = s. f* EL.100'-O* GRND. FL. fi,, 1 E - 20* # NENTIL ATION [ T j Ent*AOST DOCT

.y-. -.

GPC .JE0"t g,j EL. 7 5'-O' 25. SECT 10P4.

DIMENSIONS ARE APPR@

Tis $ Dwde.' Fig. 4.1. NFS cell configuration and cell ventilat: . (

i 4-2 - ~7- . :3 CPC CRANE ROOM . 4:. ~- } .J l 7 C PC [ys" ; I 9Y 0" E.L. 10 0 - O

_a

!). L_

J (2PC

-- 4 I SECTIOh! M Twas pura. h f EL.14 8*- O*

v. c...

._t J. ' T;j k .'h , ; e.*, Agggg-10"W u E.NTI-rLATIOM ,{PMC f] C PC, SuPPIT Ud[T .g r_.p.p ,o O m .!9 N t

y;

, it'- O"e, s '., : 2 2'- O' e M f..e,.

' s,

,I U ' 00$~ _01__. T L [T. fj P e. __j k _.k Y GPC ~~~ C3 P C &~ 4 W o* e CRANE RM. ..,,c. . 9.. SECTION /3T iWE Q MSDW6. H3D 11781 G004 on path, plan and sections. s e _1

4. HEAD-END CELL AIRFLOW ANALYSIS Examination of the NFS facility in its present shutdown status revealed several potential pathways for release of radioactive material from the head-end cells to the environment in the event of a tornado i strike. A schematic model of the head end ventilation system and the l potential release pathways was developed for use with the TVENT computer 1 program. Analysis with TVENT provided an estimate of maximum airflows to be exsected within the head end cells and adjoining pathways and maximus volumes of potentially contaminated air that could be released from facility (main process building) boundaries during a postulated tornado strike. The following subsections describe identified pathways and techniques utilized in modeling these pathways for analysis. } l 4.1 Potential Release Pathways i, A primary concern of the NRC staff investigating safety aspects of. the NFS facility is the ability of the head end cells to contain radioactive material in the presence of tornado-induced forces. Working in conjuncti~on with the NRC staff, a number of pathways with potential for releasc _ f rom the head-end cells were identified. Pathways from the GPC and the PMC,were given major consideration because of high radiation levels in these cells. ( l The b-ad end cell configuration and cell ventilation pathways are shosa ir, .l. Arrows - through ducts and hatches indicate the di rec ti: aormal ventilation airflow through these paths. Ventilation air is' supplied to the PMC and the CPC by infiltration through gaps around doors f rom adjoining rooms. Flow rates for ventilation supply air were obtained from drawings of the existing ventilation system and from actual measurements at the site. Using measured pr.cssure dif ferentials and airflow rates between the rooms, calculations were made _t;o determine the equivalent diameter ducts which would supply. required ventilation air to the PMC and the CPC from the surrounding corridor at the existing pressure dif ferentials. These ~ equivalent ducCs are indicated in Fig. 4.1 by the assumed 36-in.-diam supply duct to-the CPC and the assumed 10-in.-diam supply duct to the PMC. The ducts represent the same resistance to airflow as presented by the actual air supply pathway. A dimetric cutaway of the three head end cells,with adjoining rooms ynd corridors is-shown in Fig. 4.2. Potential airflow pathways in the g. foJu of open hatches and ducts and partially sealed ducts are indicated a-in this drawing. 3 .. ~ s i t b 4-1 o

v

~ + ,g i _3 _e

7 w ~ 4-3 yx \\ \\ / <A s x N \\ $l-[iJ ',i: ?.' tt Mo.. ~1 ..,j ;g . * * ~ ~ ' *

s N

Y ~ ( %:b Q'.9 / L~ ti N s s \\ t .-,.s W -) t

i s i ens e.r, to ass ,FN ? N LEGEND s x m- . -w 'N A W,Ni & T URE CELL N B-CENTR AL PURPOSE Cf LL (JPC) -s\\ s-GPC OPE R ATING AtSLE 'N 0- PROCESS WECH ANtC AL C E L i. W 4e s ) E - P M C OPE R AfiNG AiSL E s \\ \\ F-CHEpiCAL PROCES$tNG CELL (C P CI \\. G-GPC Ch ANC M*A'N ROOW '-s s H. C PC Cli A%( R( PA R R 0tsW 's N N t-W AN,Put ATOR 0 E NiNG \\.: :,. ' \\ twa=+vt a tos,a eL aca l \\ \\ i, 2-W AhiPUL At M C E 8elNG teivoooo covee s ~ 3-W ANiPUL AYOR OPE N'NG N iainst o. Loces sos cowtai m 4-W ANiPUL A f GR N y ,g N staeco-cm canoesano tovsei N x .o \\ [ 's.i' 5 - DUCT SE T

E E N GPC AND W4Af uK. CELL tort ai s \\,

,\\ l# 6-T R ANSF E R PORT tortte 7-MATCH - PWC TO GPC soetti t-M ATCH - GPC YO CPC eoetmi p"gif 9-MATCH - GPC YO SCRAP REWOvaL AREA I i i g taaTCu et PLuGGEo) g %, g N z to - VERYiC ALLY WOviNG CC*R - ( J Get to GPC caans esta e aet a ,' 3 1\\ 'j v;%( k f it - VE RTiCALLv WoviNG 000R - p ewC to PwC caant espais enga 4 ~7 h,-Tg. ~ 5 y-N* N ,,,o, vt R TIC ALLv WoviNG. DOOR -

2-

.p .o.....a.. ..Pa.a...a g) s 13-V E NTIL ATiON E MM AUST DUCT 's g; N-f s N i :, ,x .s / ice < = > 14-CO2 TRANSFE R OPE NiNG w-l 3 ~ 'I d e #. f ,s c. I "I , a, s-U gL / ~i v /ru ,w e. 3 (/ / y 'y N,f" 9 to Fo

9...es_

x a. (... Fig. 4.2. Dimetric cutaway of NFS head-end cells. )

1 4-4 ~ e.. The most obvious and, as later established, most critical pathway is through seven manipulator ports in the PMC cell walls. Manipulators j have been removed from these ports and the openings covered with either plywood or cardboard held in place by duct tape to the cell walls. If a tornado passed across the f acility in the vicinity of the PMC, these manipulator covers would be removed by either tornado-induced winds or the differential pressure produced by the tornado. Locations of the seven manipulator ports are indicated in Fig. 4.2 by labels 2 or 4. Also shown in Fig. 4.2 is an open hatch connecting the PMC and the GPC. The presence of this hatch is significant because a release of air from the PMC to the atmosphere has the potential for containing radioactively contaminated air f rom the GPC. The second identified release pathway is from the CPC through a partially open butterfly valve (operating position open) and gap around a shield door to an equipment decontamination room (EDR) with accesses i to the outside of the building. A block diagram indicating cell d configurations and adjoining rooms and connections is shown in Fig. 4.3. The CPC-EDR pathway can be seen on this diagram. The CPC is located above the CPC, with an open hatchway joining the two so that any release of air from the CPC has the potential for containing radioactive material f rom the much more highly contaminated GPC. The third pathway considered involves three ducts temporarily sealed with cardboard covers in the walls of the GPC and the adjoining minicell. If the cardboard covers cou?d be removad as a result of pressure dif ferentials created by tornadic wf ads or by the force of the winds, air potentially contaminated with radioactive perticulate matter could flow through these ducts to the surrounding general purpose operating aisle (GOA) and from this aisle to an adjoining stairway to the outside. These ducts and room configurations are shown in Figs. 4.2 and 4.3. To arrive at the fourth possible pathway, a conservative assumption aas maJe that the outer walls or roof of the master slave manipulator j (MSM) building adjacent to the main f acility building poses no restriction to airflow. As indicated in Fig. 4.3, the main exhaust duct f rom the CPC connects with the main exhaust duct f rom the MSM building before reaching HEPA filters and exhaust blowers. It is conceivable that with a collapse of the MSM building walls or roof, this exhaust duct could be penetrated thus providing a path for contaminated air to be vented f rom the GPC to the atmosphere. A fif th pathwa-; through several rooms adjoining the PMC was identified and evaluated, and its contribution to a release f rom the building was found to be negligible. This pathway is from the PMC through a labyrinth joint between a sliding shield door and fixed valls to the mechanictl crane room. From this room, str could flow through an .g open hatch to the manipulator reoair room and tarough an air lock with a [

Stamotu fT5GT - -s Ci catuca6 o,tmarwes a.et mesta (PM fS0000 hh 1 C( catuc< SG (see9 SMdLS 1152 1 CF catucai 64 t@,9 enaast. sep% e 1P MRa uCR PM raattss W waas.e.vt aton. u.teman.ao w,,2m, . =, t ct. .E,a c aw (14 30 hh (5420 hh { f f ses, - ^ db MRR PuCR asWL0ca a WL(Tet 8 (200 hh (MM 1 estasass 986408 h 1 .g -.g l a.~.g l o o l'A.ewitL sta.nettL GJTSM OUTSM l 4 k,

I -5 0.est Dwii 79 20073 Out t towf.f OffbO

f. JY OUTStit to g

wiTM woDD 1 0 <- e.a.c. ise -.4 ll l Ouf5M O s.ete t-t.t e Osam 000 ..EDR LOCA (200.'l TO Castu O.L2tfiesr,.et.L iSLI CC., CA. pet ) s ---S DECOset.apen.t t rou, w foonocu e.aag.,pm i' s?'

\\'

i 00*% =, +w .. S.R 9 to.c*v. ve s.o e

== +' a

enm-- i

! 07. -~ ~~--- ~ a, g,.

oo,,,,,,,.

,,,,,,, _c,c n - "e% ua y, ~ ' ',, + ',4' *-.'x=*-Pt- ..uiu

* =. * >

o.., i =o.s .._.m.,,,, i - n i... ~., 9- ..a>w _,,,, ~, 31j m n 4 F!:Gl+ =_.:e u m. coa s.=*L ~ m..L,uv , m......; ..a o _t_ oursa mooo.., ..,v,,, $.MR ....M LELL na uu m a,.. re,, 7----+ 123920_.I_ _. _ l g, I - - -. - - + - - - - - > orCo.r.amnaro.,s w j r., .on.i uceso mv..gc.L i ~* m s,.,.,,....,..

o..so.u,,,..y..

-. ~. ~ .L .n o s, DEc.ces a,<ces .a. .....--..e,m,....m.,.,,,,yo,t..... .,c, u, coo n>> -. ~ '.. ,, coo,,,, CLO$f D ,M..W C PA'oe 70 CA,75M t.'t.f, pee,.t, o e ... e ,t.ewt LL ,T.a.rtt outsot OUTSG OufteDE de'.8 E 8 47f. Fig. 4.3. NFS cell configuration, block diagram. f I L

5^ I pr' 4-6 ( plastic-covered door to the outside. This configuration is also shown in Fig. 4. 3. 6 Schematic models representing the four major pathways and modeling L' techniques are discussed in the f ollowing section. I 4.2 Modeling Techniques A single schematic model was developed to represent the mathematical flow characteristics of the head end ventilation systcm and four potential release pathways described in Sect,. 4.1. This model is shown in Fig. 4.4, with flows indicated for normal operating conditions. In modeling the ventilation system, a lumped parameter approach was used in which flow characteristics of cocponents such as dampers, filters, and ducts are lumped within oystem elements or branches, and spatial distribution of variables is neglected. The connection points of the system elements or branches are denoted as nodes, and the interconnecting system of nodes and branches forms a network which can be analyzed mathematically to determine system performance. In Fig. 4.4 the major components of the ventilation system, valves, filters, ducts, and blowers, are located on branches as denoted by the numbers in parentheses. These component 9 are connected by numbered nodes. Flow characteristics of components within the cell ventilation and alternate exhaust pathways under normal operating conditions are given in Table 4.1. The filters in branches 2 and 6 represent infiltration to adjacent cells f rom surrounding rooms. The relationship between pressure drop and airflow rate through an infiltration path is approximately linear, as is the pressure drop flow relation across a filter. The corridor surrounding the PMC, represented by a room at node 2, actually adjoins other rooms in addition to the PMC; however, the pathway was modeled to utilize the TVENT capability of allowing fluid storage at nodes, which would represent damping ef f ects of a large volume surrounding the PMC on flow out of the cell to the atmosphere. When compared with the ef fect on flow f rom the PMC, connection of the e corridor to other rooms would have an insignificant ef fect on results and would only complicate the modeling unnecessarily. The four potential pa chways were actually analyzed separately in order to provide information in the event that one or more pathways were removed by f acility modifications. Alterations as described below for each case were made in pressure and flow or resistance characteristics of the model in Fig. 4.4 according to which pathway was being considered. In this way, the necessity for developing four separate models was avoided by simple modifications to a single model. The four ,T canes annlyzed are described below.

e ORNL-DWG 81-13178 e, e, 2S 28 3500 l 3600 C' " till CFM (2e, RE P#ESE NTS usFeLTeate0m ~~ esses one ~h reone Comespam. CME esiCat ( >26 EDe men ne (; CRANE Acoes $setLD Doom 14J00 ce n I ( 52.000 se M1-- 6900 a40 Scear REuovat noces ^ = C F is .500 Ca es ' '[t29) (25 ~~ 4F00 IIOI C } 25 o e c ' ' " o 6 F @ 125) S ,g>. ....C,. .. e.. S .e..,... CC 1821 S00 se h b gy e, g, mtPatSENTS issFatimailDie Faces CometDon 60520 [_@ gCF (283 15920 Cres 15920 Cres Coma Oca PesC 5 300 2000 CF_, (IS900enh) ce n 3 (25 000 se M ) g t er 20 2 3 . 5, e. _ 3.... e6..e. . 9. .22, .2, . 5, wa. n e., wra w ra ,,g,, ~ 4 s20s 4 iAA "9 (se or ,o M CMi so 5 C F '8 ga },, L EGE AD ,,g,, I C'" VA = ouer,w r.o. . ~,c, q c,, i o 0, -p are. ion =ai ii i .S.o0o.o h i inic 34 we m1]=::r, e, , e, -+ -.ao e necedery NFS CELL EXHAUST MODEL FOR TVENT COh0PUTER ANALYSIS does 1-25-79 doetLJM dra DAT UCC-led dag. H30ll781-GOOS Fig. 4.4. Schematic model of NFS cell exhaust system for TVENT computer analysis.

4 4-8 4 5 i ' 7 .ie . he E.a....h.e.e h h h ww h .m W w . be vge g.. ge r = 2 = 3 me = = = = =- w me

== w se em .e W w w==== = se== g= 1 g ig3 2 1 g==== g== 1 1 1 1 s.e s= 9 y > h =J >e g =h>>>w CC>>hhha>mm>>>h>>bch i W 4 I h== $.$ e $ c ek k b 3 $ k.k k k k k k. 3 k. 2 5.Nb h $. 4. x b h k h s.e .e t. A .a.. N = 1 a** N.ie Nw4P

== e. e. e P'.M ra 4 4 m e as e = m m m ] C f g ..= me = ee me i 'I he i w en ga

== T 9 =.* ^C d ua o a,, m 2g

== c e== A e== e 4 e ee e =

w.

3 w a w ", . e. e. C. C. W.== C.==. #. N. O. w. P=.a. C. e. P* e C. O. 4.==. m. M 4 N e. 4. m. 2 A W .= a 6e 1 H 1 =A 2. =. 2 C C C C C C C C C C C C e O C M ao me C 4 C C C e C e C C C C .a a Z a w.2

== a N *e a h.e " 1 h li. I

== 3 a.

J C

Q, to me be 0 30 ,5 - en S 4 C S(J C ,== J.s .O y e =

3

= ort

== e = e se e = -e.C. C. C. C. @. C. C. N. O. W. G. e. N.P6 Pm. N. e. 4 4 4. C. =. a. SG.10. e. A. @. C. C. e C 6e N g =3 j C -. = c O - C g.= g me 4 P= P. p a p p g C g g a e e 4 g g== g 3. 4 .s 2

a 6 e i a T t 6 a 6 e t i 1 e e== -s =*

4 I i 8 6 4. I 3 9.g 4 9 2 4 1 8 ] 3 hs== .S ** S $is 48' .8 n 8= 3 a = 3 .=, c, I. 3 =e 5 e e =.=s W 3 es

am W 3 83 ya es *@

N M M 4 4 Pm W S C 4 = W N =M 4 e 4 Pm a e.@.N N N N N== NC *e M 4 4 N 4 Pm 4

== 3

J=

e a g I = 3

== . = = = = = = =. = = = = g 3 3 .a W 3 a w

e 3

4 $$ g G

ac he==

W W ed

== w .* 1 ee 1 he 4 .a 3 I le a el m . s e me 3C e e e= y O..ie.==. C. C. C. e. C. C. N. N. P. C. e. N. **. Pm. N. e. 4 4. w. C. =e.e

1.==. 2. C. M. N.

e .e Q L he hs me N O 3 y y 3= =. I l2 O C C== C S S C e C o O ** 4 N ** S. cD*mmOCSeCeCaC 2 1 s e a TT T e T 4 i a e # = = = = = - - Ts e # T oe 3 1 u 2 e. 1 8e M 4 4 6 8 4 6= a

3 w

he = w ga T SW } U J = 4 W he = 0 i S D ha 1# 3 ='aJ O > 0 =. .e a e5 i =. 4 m. ?! is 16 9 3 3 3- .hs "C

== N N M W's 4 P= W p C C== W N M W e 4 f's.e =83 @ em N M N @ c4C 2 hs 3 3 8 =e== se ,a========== === N N N N N N N =* A 1 = es 2 C D we 2 se. a ha 3 3 4 3 4 Gs. 21 4 u s . to3.. k"" 4 W he f. n = A 3 he M i 4 b a X be J G k4 U mmea e. w m em m es m a e== m a n e m a e m e s .=is s e m a e ee==m a,,C N g @ E N M 4.c %se w-r =st %.a smeS G4 C 88 N M 4 e @ Pm 3 @ geNMQ@ O e 4 Pt w"8 8 ww% r== ==========.======= N N N N N N N N N N M se 4 IS %EF %uf W %uiF %u# %s# insP V iWP %5F tm# %ulP MuuF tur %ur %ud %mF W 81mir %5F 9usf due O 4 4 ) n.

4-9

4. 2.1 Case 1.

PMC manipulator port pathway The pathway through seven manipulator ports in the process mechanical cell was considered for the first case. The seven ports were Icared into one duct with an equivalent resistance as represented by branch 3 in Fig. 4.4. It was assumed that when a certain pressure gradient across the manipulator covers was reached, all covers would dislodge simultaneously, thus lowering resistance and allowing higher air flow through the pathway. Techniques for modeling cover dislodgement are discussed in Appendix B. As discussed in Sect. 3, all cases were analyzed for two postulated tornado imroc t modes: 1. the tornado simultaneously imlacts all building supply and exhaust

points, 2.

the tornado impacts only boundary points providing air supply to the ventilation system. Results of the analysis indicated that during the second mode of impact, larger volumes of unfiltered air with the potential for containing radioactive materials would be released from the cells to the atmosphere than during the first impac t mode. For case 1, the second impact mode Wds simulated by applying tornadic effects at boundary nodes 1, 5, and 8 in the mathematical mods 1 while nct affecting boundary nodes 20, 21, and 25. 4.2.2 Case 2. CPC-FDR pa thway This case involved a pathway f rom the CPC through the adjolaing FDR to the atmosphere. Access from the atmosphere to the FDR (represented by boundary node 25) is at a point on an opposite side of the building approximately 100 ft from other boundary points. It is possible during the second postulated tornado impact mede for the tornado to impact accesses to the FDR while not affecting the other portals. For the second impac t mode, the tornado pressure function was thus applied only at node 25 in the analysis of this pathway. Covers over manipulator ports and over openings into the CPC were left intact for case 2 and thus eliminated these pathways from consideration. 4.2.3 Case 3. GPC to operating aisle pathway Pathways through a cardboard-covered fire protection duct in the GPC wall and through two cardboard-covered manipulator ports in the adjoining minicell walls were analyzed in case 3, with manipulator ports

4-10 in the PMC walls sealed. For the second tornado impact mode, the tornado pressure func tion was applied to nodes 1, 5, and 8 as in case 1. As discussed in Sect. 3.2, the physical locations of boundary points 1, 5, and 8 in the building are in the rain building stairwell, and all three points would be simultaneously and equally af fected by a tornado. Techniques for simulating loss of cardboard covers during the tornado transient are similar to that used in case 1 for the PMC manipulator covers and are described in Appendix B. 4.2.4 Case 4. GPC-MSM building rathway In the final case the pathway considered is from the CPC through its main exhaust duct to the MSM building exhaust duct, which has been renetrated and exposed to the atmosphere by collapse of the MSM building walls. The MSM building and exhaust blower and filter house are closely located and could be exclusively af fected by the tornado. The technique for simulating loss of MSM building walls is discussed in detail in Appendix B. In summary, resistances to airflow provided by branches 23, 24, 25, and 26 in the model are deleted, and the tornado is applied at node 24. At the same instant, the tornado is applied to node 20, representing the boundary f rom the exhaust blower and filter housing. Covers over manipulator ports in PMC and CPC walls remain intact for this case. The following section describes in detail results of analysis of the four potential release pathways described above. e

5. HEAD-END CELL AIRFLOW UNDER TORNADO CONDITIONS In interpreting results of the analysis of head end cell airflow, the TVENT output of airflows and integrated volumes of air passing through each pathway was examined to determine whether air would flow from the cells to the atmosphere through unfiltered pathways. If a case were found to have unfiltered airflow to the atmosphere, volumes of air released were computed. For most of the postulated pathways, because radioactive material is deposited on the cell floors and not dispersed within the cell volume, a substantial quantity of air must be discharged from the cells before any contaninated air is released. A discussion of the basis for the required discharge volumes is given in Part II. 5.1 Air Discharge Volumes for Cases Analyzed The pathway with highest potential for release of radioactive materials was found to involve the PMC manipulator sleeve ports considered in case 1. Volumes of air released through this pathway were the highest for any of the pathways considered, and the air released could contain radioactive materials f rom both the PMC and the GPC. Some unfiltered volumes of air were predicted to be released to the atmosphere through all pathways investigated, but a release of air would not adversely af fect the environment unless it contained radioactive materials. Part II describes the methods used to determine whether tornado-induced airflows within the cells could reentrain and release radioactive materials to the atmosphere. As discussed in Part II, the air discharged f rom the contaminated process cells will always contain small quantities of rador ' rom the decay chains of uranium and plutonium isotopes still in the...ad end cells; however, the discharged air may not contain any particulate matter. In determining whether further analysis would be required f or cases in which air is discharged from the cells, a volumetric shape for the air space within the cells prior to its discharge was assumed. If this air space intersected surfaces on which particulate matter is contained, then some calculable amounts of particulate radioactivity could be included in the release. If a predicted volume of released air produced a geometry which did not intersect contaminated surf aces, then this case was eliminated from further consideration because no radioactive particulate would be included in the release. For cases involving a release f rom the PMC, the volumetric shape of air discharged thrcagh the manipulator ports is assumed to be quarter-spherical or a truncated cone centered at ports in the northwest corner of the cell (see Sect. 4.2.1, ref. 1). In order for this volume 5-1

5-2 to intersect the floor where contaminated material is deposited, the volume must be at least 1050 ft3 Cases involving discharges less than this amount are deleted from further consideration. For cases involving discharges of air f rom the CPC through a port 3 to the operating aisle, a hemispherical volume of 2100 f t is required before radioactive particulate can be released. This covers all of case 3. For cases 4 and 5, in which air is discharged f rom the CPC through the ventilation exhaust located near the ceiling of the cell, and for case 2, in which air passes through the GPC hatch in the ceiling to the CPC, a volume of 7700 f t3 is required before radioactive material is released. A further discussion of the basis for these estimates is presented in Part II. Table 5.1 presents a summary of pathways through which a discharge of air from radioactively contaminated areas (head end cells) was predicted during a postulated tornado strike. These predictions were made by analyzing the four cases described previously plus an additional case described below for ef fects of 100, 200, and 300 mph tornadoes, utilizing TVENT. For case 4A, analysis indicated that no releases f rom any of the contaminated cells could occer through this pathway. A summary of estimated volumes of air discharged from the head end cells during the three postulated tornado accident conditions is presented in Table 5.2. The branch designation above each column indicates through which branch in the NFS schematic model (Fig. 4.4) the estimated discharge occurs. Air discharged f rom the PMC, GPC, and CPC has the potential to contain radioactive materials from these cells. The air released from these cells would not necessarily reach the outside but would be mixed with uncontaminated air in surrounding corridors before release to the ambient; therefore, the total diccharge to the ambient f rom the release pathway (column 6) is a combination of uncontaminated and potentially contaminated air. 4 For case 1, the PMC manipulator sleeve pathway, substantial volumes of unfiltered air are predicted to be released from both the PMC and the GPC for 200 and 300 mph tornadoes for the conditions of both exterior walls intact and removed. Predicted volumes released for cases lAl and l A2 were less than the required 1050 f t3, and-these cases were deleted from further consideration. Unfiltered discharges of air through this pathway are caused by reversal of airflow in the PMC-GPC ventilation supply path and removal of plywood covers over the PMC manipulator ports due to the presence of a low pressure area within the tornado at the supply end of the ventilation system. Further analysis is required for cases 1B1, ICl, 1B2, and IC2 to determine the amount of radioactive material contained in air released through this pathway.

5-3 Table 5.1. Summary of pathways _with discharges of air from radioactively contaminated areas (head-end cells) during postulated tornado strikea Pathway of Tornado Cell (s) from which release wind speed potentially contaminated analyzed Case (mph) air is released e 1 1A1 Outer 100 PMC, GPC PMC 1B1 Walls 200 PMC, GPC manipulator 1C1 Intact 300 PMC, GPC sleeves to operating 1A2 Outer 100 PMC, GPC aisled 1B2 Walls 200 PMC, GPC IC2 Removed 300 PMC, GPC 2 2A 100 CPC CPC EDR 2B 200 CPC 2C 300 GPC, CPC 3 3Al Outer 100 GPC GPC to 3B1 Walls 200 GPC GPC 301 Intact 300 GPC operat(ng aisle-3A2 Outer 100 GPC 3B2 Walls 200 GPC 3C2 Removed' 300 GPC ^ 4A 100 none MSM 4B 200 GPC building 4C 300 GPC 5 SA 100 GPC Blowers 5B 200 GPC inoperative SC 300 GPC

As determined from TVENT computer analysis of potential release pathways.

bA release to the operating aisles was considered to be a release to the atmosphere because of the uncertainties regarding structural integrity of the exterior building walls and doors. u

~~ .~ Tal,la 5.2. Air volumes discl.arped f rom cells during postulated tornada st rike Estimated lay TVENT coquter cmle b W almuus Voliamed Volume Volisme" Total Discharge d i tornado released released released to ame>1ent Came wind, tamed t u om INC from CPC from CPC f rom pathway (mph) (ft ) (it ) (ft3) (ft ) Brancli 3 Beaucis 4 Braach ! IAI Dut e r 100 362 275 0 570 151 walls 200 1273 !!42 0 2300 ICI intact 300 1983 1922 0 3953 Branches 2, 3 1A2 Outer 100 484 4 39 0 568 182 valls 200 1947 2222 0 2792 1C2 removed 300 3057 3912 0 5098 Brancle 7 Branch 29 Branch 28 2A 100 0 0 896 919 23 200 0 0 2522 2584 2C 300 0 45 39 34 3976 y S~ B rancties to, Il Branch 8 3Al Outc. 100 0 0.4 0 149 381 valls 200 0 11.6 0 570 3Cl intact 300 n 4.0 0 1001 Branches 10, 11 3A2 Outer 100 0 289 0 300 352 walls 200 0 928 0 970 3C2 removed 300 0 1428 0 1515 Brancte 13 Brancti 27 4A 100 0 0 0 0 I 43 200 0 2344 0 57 '00 0 4004 0 810 4C J 5A 100 0 1251 0 1026 5B '200 0 2515 0 1843 SC 300 0 3657 0 2544 ' Air potentially contaminated witta raJtoactive materials. b ondelneJ potentially contaminated air and uncontaminated air. C

5-5 e In analyzing case 2, which involved the pathway from the CPC through the adjoining EDR to the atmosphere, it was inappropriate to simulate removal of the exterior building walls and doors, since the 3-f t 3-in.-thick walls of the EDR are the exterior building walls at this point and would not be expected to f ail under any tornado loading.9. Accesses to the EDR from the outside are provided by a door and by a plywood-covered duct, which would be expected to f ail under tornado-induced forces, and are modeled as removed for purposes of e analysis. l As indicated in Table 5.2 for case 2, substantial volumes of air are predicted by TVENT to be released from the CPC through the unfiltered EDR pathway for all three tornadoes postulated. However, as will be discussed in Part II, although significant radiation levels have been measured in the CPC, these readings are caused by radioactive materials in equipment in the cell, and very little resuspendable or dispersible radioactive material is present in the CPC. Based on this conclusion, it was determined that unless a predicted release f ron the CPC could contain airflow f rom the GPC, the cases involving air discharges f rom the CPC alone were not considered to have a radioactive release potential, that is, cases 2A and 2B. In case 2C, 45 ft3 of air from the GPC could be included in the discharge f rom the CPC; however, this volume was not considered to have significant potential for release from the GPC to require further analysis based on the minimum criteria of 7700 ft3 of air. For case 3, involving the pathway through manipulator and fire protection ports in the GPC and adjoining minicell, the case was analyzed with the exterior walls and doors both intact 'and removed. Tornado-induced forces would remove the covers over ports in the cell wall, creating an unfiltered pathway for air to flow into the operating aisle of the GPC. The operating aisle adjoins the main building stairwell, which has doors and openings to the outside, and any release to the operating aisle is considered to be a release to' the atmosphere. As shown in Table 5.2, releases through the GPC to operating aisle pathway to the atmosphere were predicted by TVENT to be less than the required 2100 ft3 for all the subcases considered. This is due primarily to the f act that openings f rom the cell to the operating aisle are very small compared with the sizes of other airflow pathways f rom the GPC; thus, resistance to airflow through the manipulator and fire protection ports is much higher than through the ' exhaust duct and the hatch to the PMC, and air tends to take the path of least resistance. Based on the low predicted volumes of air released through the CPC to operating aisle pathway,_it was determined that case 3 (all subcases) did not present a hazard potential to the environment, and no.further analysis was required. For case 4 which involved a relcase of air f rom the GPC exhaust duct through the adjoining MSM building exhaust duct to the atmosphere, in postulating this release pathway, the MSM building walls and roof are

5-6 assumed to collapse upon tornado impact, causing damage to the FiSM exhaust duct, which provides a pathway for air to be vented to the atmosphere. Without collapse of the building walls and damage to the exhaust duct, no pathway with potential for releasing air f rom the cells to the atmosphere is present; thus it is inappropriate to analyze case 4 with the exterior building walls and doors intact. As shown in Table 5.2, the TVENT predicted volumes of air released a 3 from the GPC for cases 4A-4C are substantially less than the 7700-f t minimum; thus these subcases do not present a hazard potential, and no further analysis was required. At the request of the NRC, an additional case was analyzed which considered the f ailure of the blower system for the exhaust blowers and of the emergency backup power system. The ventilation system was analyzed with all of the manipulator and other ports in the cell walls sealed, with the IGM building walls removed because of the building's proximity to the blowers and power supply, and with the blowers inoperative. Analysis of this case, denoted case 5, with TVENT, for 100, 200, and 300-mph tornadoes predicted that signiticant volumes of ~ unfiltered cir would be released f rom the GPC through the MSM building exhaust duct to the atmosphere. The volumes released from the GPC through the ventilation exhaust duct (represented by branch 13 in the schematic model) are as follows: case SA, 1281 ft3; case SB, 2515 ft3; case SC, 3657 ft3 These volumes are less than the required 7700 f t3; therefore, further analysis was not required for this case for radioactive materials. Another case was also analyzed at the request of the NRC to predict performance of the ventilation system under tornado accident conditions if all the ports in the cell walls were permanently sealed. Securing the ports is a method proposed to ensure containment provided by the head-end cells during a tornado strike. Analysis of this case with TVENT indicated that with all ports into the cells sealed to resist a pressure dif ferential of at least 36.3 in. H O (1.3 psi), the total ( 2 pressure drop of the largest tornado considered, there would be no discharge of air to the atmosphere through unfiltered pathways. 5.2 Cases with Potential for Transport of Radioactive Material As discussed in Sect. 5.1, four of the twenty-one cases analyzed have potential for releasing significant amounts of radioactive materials f rom the head end cells: Cases IB1, ICl, IB2, and IC2. Releases from the CPC alone are not considered to have a significant potential.

5-7 The cases identified as having a potential for_ releasing significant quantities of radioactive materials are the following: 1. case 1B1, PMc manipulator pathway for 200-mph tornado with exterior walls intac t; 2. case IC1, PMC manipulator pathway for 300 mph tornado with exterior walls intact; a 3. case 1B2, PMC manipulator pathway for 200 mph tornado with exterior i walls destroyed; 4. case IC2, PMC manipulator pathway for 300 mph tornado with exterior walls destroyed. These cases were analyzed further to determine the amounts of radioactive materials contained in postulated releases f rom the facility. The next step required in decermining quantities of radioactive materials contained within the air released from the cells is the prediction of tornado-induced air velocities within the cells. This procedure, discussed in Sect. 6, utilizes input f rom the TVENT analysis in the form of transient airflows and average velocities at cell boundaries. Pathways through which air can flow into and out of each cell are designated as branches in the cell exhaust model, shown previously in Fig. 4.4. Flows through these branches are considered to be flows at the boundaries of each cell, since TVENT does not model airflows within the cells. Summaries of transient airflows generated within the facility under - tornado accident conditions as estimated by TVENT are presented in the following subsections as plots of flow rates as a function of time. 5.2.1 Case 1. PMC manipulator port pathway Airflows in branches 1, 2, 3, and 4 representing the PMC manipulator port pathway with the building walls intact, are shown in F ig s. 5.1 and 5.2 for 100, 200, and 300 mph tornadoes, cases 1A1, IB1 and ICl. Sharp peaks in the flow rates at 2.0 s or less indicate times at which dif ferential pressures across the manipulator covers reach a specified point, causing cover dislodgement. Negative values for the flows indicate reversal of direction f rom the directions of normal airi.'ow before the tornado strikes. Normal flow directions are indicated in Fig. 4.4, the schematic model of cell exhaust pathways. In case 1, during the 10 s duration of the tornado at the facility, the flow of air reverses direction in branches 1, 2, and 3 and flows out the_ ventilation supply path. Note that during this reversal, flow also reverses in branch 4, the open hatch between the PMC and the GPC, w w +w

yoo T ) C o 0 A d 0 T a 3 I

i: 4 3,4 i!!!, ); > ;

3 7 N n 8 9 I r 1 S r jgI1 1234 o 8 GW

}
,iy t i

G t D NO e s, L h D 0 I ODF ,t

j!

a.m j hhhh 5 L p i} + 4.! Ncccc te 2 DL Ennnn AIL m a O p. ,.v '{}kl! Gaaaa NUA 0 Errrr RBN 0 f.f tj1 t mi} LBBBB O R 1 T .E ( t

==

0 YT c 7Ji. : I], . i-9i

i{

0 A y i; ii i oa+< 2 h N l a W p I At e H l n mTD i,

1%t,

.tjj; { n' i ,I,:, AE e 0 s PT ss 0 0 2 A al j 5) ER cl +j +. -q e jt M 1_ I t _t s VE a 1 E / j / ( N rw s,_ EBEE o y

[

I /./ - - _I M ILG f g ,/' j 3l r,, S n IT ESRO yi j: A (. I iIJii0 AO wl 0 D ad - ;. /, !1 i ,!;f i', 1 CTA hi AN t u LR ab rI l i T

fI4(i*l

, (f ! d UO p s <.l / PT r /- eo I 0 ,t f .j!; i' 3hi .j, .f. 5 N vi i. g A er M ee l', l t t sx q tIjt; Il

!.l l4l l,

~ . A' _ \\Y .0 E r g( t 0 o t. o8 ,c o. {.

T gi aT l N b7

],,.s i T uE s C pV 0 A iT 1j,' ,!4 0 T n [_f t

I l;

}.

3 N ah S mt I + jI4 il!;, + 1234 i

j,. i

{t_ GW nw NO i u L D 0 I d O.F ti. t_ D se .p.+ I ;p,;. ,j I 7 hhhh ,5 . p4 Ncccc D 2 wt E ~ nnnn AllL oc Gaaaa NUA li 1.,j,f 7 - Errrr RBN + fd LBBBB O R re T .E i r

1

==

0 YT ap 3 i._ !4 . 1 - Mt pi ,0 A t} 4! oa+= 2 hWN p I d s H ea ._~ t,o ti Il mTD . pi9. k{- t AE a) e 0PT ro s miyl. 0 A ed n 7 it ,7 0 1 . \\ 3, l..y ER na 5 )s A

i _

1 EVE en . /. r/. p;.- ( N gr EAEE o m. ,I / ]l - l M 1LG ot i.f s S d 'f 0T SRO ah /. P CTA I E D np 1+ f l t. - (,j 9i 0 AO rm p _ ,//. o-AN T0 LR _0 i 7_ i. [.p }, 4 l li UO 2 /_ PT .( t[ 5 N

1. 1 I

s 0 4 4 7.w pt i _ *. 7 p_ I . \\ A 5B M 1 J

Oy ge f!p y,

] 4 4 i s _ h+\\ 0 F a e L .0 c oS om o.. ,o ,7 y g' g7 d n j,,. 3 s:. a

o-i:

l onwowc. ei enn t ? - *j -'] bo ' g g i bo O e Q- '.. ';I -_ _. ~._L - - +.. " "] j ..{ ..L. (b)- 1. p. i o f + o l { O" 4 4._ { '} { [ g. ._.,__.._.} 9 -_.2. 5 -_.a.._ 4 ._4_ .j_

/,

a' + j... _ _j.. ._.L. _ L.._,.41 ...i. .__ i .I !/ ' y i J j i l /' t o 1 ..g. 4 ..~..j.. ...? l /. ,L 3

g.._g_._p_

j a. ~ o I . _. L a. _. _.. _. 4 . } _.. .4 g...._.._1...y._.- .r. .. f .. _g. t l +t -a. t i c. s 1 l i o j-an ~ q.Og.. o_ j

/

/, l,, i.j l L 3

k..

-4 _- .._Q /. . T.:-kQy. y 1 - - - ~ ,[ I ____.. _. 7c .{. ..j iy . _ ___.. _._._. r _ t. 4 1 .../l. f .. g 1 .__..n_..__7..___. . J.. _; zi p-g s - r a a _..__.7_. .._ _ p -...a .k. %.g. 4 g>.. __J_ _ j__... A _ a. i. j j Li._. d_ ___ ____j(/l ...1__...__ ql h 4 i +l-q / - I-- i - 7, -- b- -- M -

i w,, _

Y' %_ g _ t _.I i I / i e A. _ g, _.. 4. 4 -. 4....+ o_ q-- y 3 ._3 l l i 3 5 r- - + - - - - -- ) 9 !--{} i p w+'I l ~* -' ' ~ - * ,._. l 7" { j r .,a-. , _.... J ..\\, _.. T _ /. _(___ o i i i o s ._ _\\ l g ___.tg. m, _. 7

3... __ 7..

4 I 1___ _.._ o = Branch 5 LEGEND 1 1 I l = e LEGEND i T (1- --T -l _- ^ = Branch 6 k-ig - ![ - o = Branch 1 \\ y- ._ y a = Branch 2 + = Branch 7 i + = Branch 3

=Bsanch 28 g

= = Branch 4 g + = Branch 29 25 0 30 0 0.0 50 10.0 15.0 20.0 25.0 30.0 0.0 50 10 0 15 0 20.0 TIME (s) TIME (s) CASE IC - 300 m h TORNADO CASE IC - 300 m h TORNADO MANIPULATOR SLEEVE PATifWAY. BUILD 1NG INTACT MANIPULATOR SLEEVE PATifWAY, BUILDING 1NTACT TORNADO GENERATED INTERNAL FLOWS TORNADO GENERATED INTERNAL FLOWS I Fig. 5.2. (a) Tornado-generated airflows in manipulator sleeve pathway for case IC1 (300-mph tornado). (b) Tornado-generated airflows in branches leading into chemical process cell for case IC1 (300-mph tornado).

5-10 indicating that some potentially contaminated air f rom the GPC could be included in the release f rom the building to the environment. Flow through branch 3, the manipulator port path, is of primary concern because this flow would be unfiltered. Subs;antial flow does occur from the PMC in a reverse direction through the venttiation supply path represented by branch 2 to the atmosphere; however, this air aupply pathway contains HEPA filters at the air intakes, and any flow out of + this pathway would be filtered. As discus sed in Sect. 3. 2, all HEPA filters within the facility would remain intact under tornado accident conditions, provided the struct 2re housing the filter remained intact. Also shown in Fig. 5.2 is a plot of airflows in branches adjoining the CPC for the 300 mph tornado. Substantial flow occurs from the cell through branch 6 to the surrounding corridor and ambient; however, this pathway represents the supply pathway to the cell, and any re tease would be filtered. It had also been determined previously (see Sect. 5.1 and Part II) that very little resuspendable radioactive material was present in the CPC. Thus it was determined that this pathway did not have the potential for release of radioactive material during ternado accident conditions. Pred'.cted flows for the manipulator pathway with outer bfilding walls and doors destroyed, cases I A2, IB2, and IC2, are given in Figs. 5.3 and 5.4 for the three tornadoes. Flows for branch I are not included because this branch is removed to simulate loss of exterior walls. Flows through the manipulator path with outer wails destroyed are 25% highe r than with walls intact for the 100-mph tornado, 94% higher for the 200-mph tornado, and 111% higher for the 300 mph tornado; thus, loss of exterior building walls substantially af fects predicted consequences. 5.2.2 Case 2. CPC-EDR pathway Figures 5.5 and 5.6 are plots of flow rate transients in branches 28 and 29 representing the CPC-hDR release pathway for 100, 200, and 300 mph tornadoes. Branch 7 represents the cpen hatch between the CPC and the GPC, with the normal flow direction (indicated by positive flow rate values) from the CPC to the CPU. Branches 5 and 6, representing the normal ventilation supply pathway, are also shown in the figures; hcwever, the direction of airflow never reverses in these paths, and thus the pathway is not of concern as an unfiltered release potential. Notice tha, only in case 2C (300 mph tornado) does the flow in branch 7 become negative, indicating reversal of airflow direction with flow occurring from the GPC into the CPC. Reversal of airflow occurs in branch 7 for case 2C and not for cases 2A and 2B because the forcing function applied at the facility boundary points for case 2C, the 300 mph tornado, is much greater (i.e., the pressure drop within the tornado is much lower) than the forcing function applied for the 100-

1 ,i p 0 T S 4e; 4 - ._ j 2 S 5 F d ,4 O 0 9 ~ L s G ~ pri4 G o _a I 8 ~ 234 w ~ [ _d W rl NS of 8t O L _.. & 'o.. D IW ir hhh D N Nccc 0 L O ri E nnn 0 O L eA 0 2 I Gaaa DUF t ~ ~ x E rrr AB e. Y LBBB N.L ~ ~ R A )

=

OYN f o ~ oa+ TAR od WE a 3- \\ ~ 4 0 hHT sn n \\ (g,\\. \\. ' l 5 pTN sr 1 AI oo mP / g l t w 'r.. ,7_f. D / 0EE hh /. _il 2VT ) .k s0 t p i m g/. j /

0. E 0

( EA . /. f[- ER w-fy LE j 0M y0 ?;+ l ji 2J,!.e u,,y '_jg BSN 1 a2 lq TIR w( I E /. G h EO _l. l4, j] .l.lh. t . / / t2 4 STO aB A._/( AAD p1 / CL A UN ee /_ _[ [. 7 _ j' 0 PR vs

4i1j1_

.lf 5 INO ea ~ T ec A l ~ ra M sd ~ n ~ _ g 4 o -cgm _ 0 t ) ~ 0 ao ag 9g o c~ o o-eo ,7 gI {i

Y l d ua b7

_3y *gA pn i r no 0 T 5 at 1 ;* [,- ,.l

4' q f;. * }j

, y ,4. f. S 2 m O h 234 a L np i m ._ i

L G

i' J_ l I_ NS s0 -. i' {;.*, i,; D IW w0 hhh DO o1 y p,,. ,2' ,+.j i.i Nccc 0 LL l( . _.. j L. Ennn [.,* >t ,t ;l it ,0 OI 2 E. a a a DUF f G r2 ~ rrr AB iA ,s LBBB N.L . h 6_ ! ,,.' p ( ' O [.. h. - ti g,. R A al t

=

YN s. s AR d e \\,\\ oA+ TWE es y.' \\ I t a e\\ '.j 3 ,g j. 0 HIT j: [ dii f; d 1l:0 pTN ac >, At 8 r e ~ mP AI er . _ ~ ~~ D no t,t/.

t.,.jf[_:t,p.

,! f._ 4 M]__j' 0EE ef ) 3 g s0VT EA - d 1 ( oe ~ - ER ~

0. E d t f

U, 0M LE /.-l )<. ' rl t

p;.

,; 4 i.j.j _.I !z. i_ aa. . h., ASN nl T 1 I T1 E ruN R G omE EO Lkt 4 TiV STO sT ,e-a_ ' 1 .i 1 AAD ~ CL ,7 _ ~ sh A UN 3.l i j gx,g)/._.. 0 PR l t ~ 3* .I + l! .;I , t,. { ;A ! q1,did 1 + 5 NO 5 aw I e _ T w A .s. p 4_ !_ d N N-4, [L

.7

,tlt . T t_ t M gge

1. ;'

J. i nt Fi c 0 ld di x - ~ ~ 0 i e ur oe g qe .o 3 o.* 2 ,E' $7 b p b7 _33 *3A n

5-12 ORNL-DWG 81-9876 'O Q 1 j I l i i i, I i. i e i t I i Y \\ i. t i l, i l I l !. M 'O..I i o es /. \\\\ i 4 // e \\e,3 i I l a/ \\ i s\\.g:\\s'! i i

i.

i .j .! j, ! 3 a f o/ / h. i E k] So "- i i i is l b; 1 o++ l l I i j t 9 I. C i n l cic i ) J llI I f-S.,_ 1 , 'p/ l, t i \\; i / !! i I 4\\l N ? f/ o, s \\\\ ! \\ l' l /. >\\ \\/e[C/ ' i ! I l i { O 1 A i I / LEGEND /l i \\ _3 .,!..__ _ _. _ _..*!^ ~ - ~ ~ ' ~ ~ a = Branch 3 c = Branch 2 ,i \\;/ + = Branch ~ 4 q + l 1 I 20.0 25 0 0.0 50 10.0 15 0 TIME (s) ' CASE IC - 300 mph TORNADO MANIPULATOR SLEEVE PATHWAY, BUILDING LOST TORNADO GENERATED -INTERNAL FLOWS Fig. 5.4. Tornado-generated airflows in manipulator sleeve pathway as predicted with TVENT for case IC2 (300-mph tornado). Loss of exterior building walls simulated.

ORNL-DeG51 9877 { _..~~~7 -~ T j f i l h $- f sa[#]A g.

I ___. \\

I d l o Al .\\ - ',f ,. W-6 '/ I .-5/w o ' A' i e g. i ./ gi a. .k--- y ~' / _1_z__._. g 1 /jg";1 1 n; a ~ i 1 i // __4___ c_ _r 3 1 \\\\ l 1

N

,,' /p d_. Y _3 o i s 4 i + f_. l i-w I I l [ l l 84 l 0 8 l l ..j i 1 3 i f, i i s s 4 o i f' t f i l y "9-i 3 l 1 i y U 3 /- I I y-r / -: i t / 9 /. 'T -i. I i / LEGEND 2- -j -I--- LEGEND ._/ o = Branch 5 I o = Branch 5 3..

r. = Branch G

T 7* ~~ ^ = Branch G og_ _ _ _ + = Branch 7 + = Branch 7 x = Branch 28 = = Hranch 28 g { l + = Branch 29 l o = Branch 29 ' O0 50 10.0 15 0 20.0 25 0 30.0 0.0 50 10.0 15 0 20.0 25 0 30.0 TIME (s) TlME (s) CASE 2A - 100 mph TORNADO CASE 2D - 200 mph TORNADO CPC TO EDR PATHWAY. DUILDING INTACT CPC TO EDR PATHWAY. BUILDING INTACT TORNADO GENERATED INTERNAL FLOWS TORNADO GENERATED INTERNAL FLOWS Fig. 5.5. Tornado-generated airflows in pathway from chemical process cell through equipment decontamination room to atmosphere for case 2A (100-mph tornado) and case 2B (200-mph tornada). Exterior building walls intact. Airflows predicted with TVENT.

5-14 ORNL-DWG 81 9878 "o o 7b" i l l I i l o 9 f i l l i i j i j oo I .i_ s ^ /V \\\\ i i o ' I,/ { / --- p \\K I I h- ) l -L g/ 1 C j n 'N i E o i /. So N

2

- + - - -. }/. o S i C f /./ i l Il. {' .9 4o. -_a_.___ k i i /, i 3 o,. j.y! o s -n--- i /f 3 ~ // i o /- l . / j', -.4__.,_ \\ l / \\\\ / / o a \\\\ / LEGEND \\. / -k i / o = Branch 5 ^ = Branch 6 8- \\ +' " - o + = Branch 7 \\ x = Branch 28 9 = Branch 29 8, I 0.0 50 10.0 15 0 20.0 25 0 30.0 TIME (s) CASE 20 - 300 mph TORNADO CPL TO EDR PATHWAY, BUILDING INTACT TORNADO GENERATED INTERNAL FLOWS Fig 5.6. Tornado-generated airflows in pathway from chemical process cell through equipment decontamination room to atmosphere for case 2C (300-mph tornado) with exterior building walls intact. Airflows predicted with TVENT, t

5-15 and 200- mth tornadoes; thus, more of a pull is exerted on the flow in branch 7 from the boundary point (node 5), causing the flow to reverse. The volume of air discharged to the CPC from the GPC for case 2C (45 ft 3. in Table 5.2) is less than the previously established 7700-ft 3 minimum and would not cause a discharge of radioactive material from the GPC. As previously noted, a substantial release of air is predicted from the CPC to the atmosphere, but the lack of dispersible material in the CPC makes this release inconsequential. L

5. 2. 3 Case 3.

GPC to operating aisle pathway Figures 5.7 and 5.8a present flow rate transients in the branches leading into the GPC from the GOA, branches 10, 11, and 12, with the exterior building walls intact. Figures 5.8b and 5.9 present the same case with exterior building walls removed. As discussed in Sect. 5.1, the predicted flows from the GPC through manipulator ports and a fire protection port (branches 10, 11, and 12) are so low that no significant amounts cf radioactive material could be released froc the GPC through these pathways. Again, sharp peaks in the flow rate curves represent the point at which cardboard covers are dislodged from ports leading into the CPC. This case also represents the condition of potential release pathways from the head-end cells if the manipulator ports from the PMC were termanently sealed. l 5.2.4 Case 4. GPC-MSM building pathway Figures 5.10 and 5.11 present flow through the GPC-MSM building pathway, branches 13 and 27, for 100, 200, and 300-mph tornadoes. Fxterior walls and the roof of the MSM building are assumed destroyed by tornado loadings, as simulated by deletion of branches 23, 24, 25, and 26 and application of the tornado at node 24 (see Fig. 4.4). The direction of flow in branch 27 determines whether a release of air will occur through this pathway to the atmosphere. As indicated in Fig. 4.4, the normal flow direction in this branch is into the Suilding toward node 12. Flow in branch 27 does not reverse for case 4A but does reverse for a short period of time in cases 4B and 4C (as shown in Figs. 5.10 and 5.11). Flow in branch 13, representing the exhaust duct from the GPC, is always positive and will contribute to a discharge through branch 27 in all three cases. As summarized previously, releases of air from the CF are insignificant for all cases.

l' I ,li]illi ,; ;a f' y"e e 0 h g 0 T t n 7 I,M }f l34 .l.l !{ 3 C i 9 A od 8 59 T tl 4 i,,i,t , j-89012 N i G 111 IS su s W O t s4i tb 0 GO r D hhhhh NL or Nccccc ,5 t j* 's },lt 2 IF Ennnnn OD i R po e O Gaaaaa DLL

A._

Errrrr AIA hr

j;;t;i.,

i' LBHBBB NUN ge ut

===

0 RBR ox y;ljj, r, *j '{_ o^+=+

20 O E

+ 1 ip ;:!ue g rE ,j T.T H h N hT t 7* s' .j.. } gi ;i t'.*i pA I

ie'f

{ , u

y. -

lT mPD / E lN 4im.j.: '~ . I 0 E eE 5 )s 0 S0LT A cV 1 y 1 2IR T ( / e fy,f. .y-il E AE sh c' IL[_ l N t,i+. }.

+[

M GE ot l D pi .,,i j,.. ! I T NG rw f?' 0 3 w I jI,* 0 u 1i},lt .+ TO pd 1 EAD - e /./ ~ 7 S RA / A '. ' * {iV],'t l t EN ac CPR ri r - - '.. { r g OO ed 0 T ne h fi o 5 O A_,,-]_."h'l P ms l, j i' er N T vf 4' gp k_ C .M. \\';1 y+ jj+ ,t 4J G oa r +k r, 0 \\ f s 0 o' e o oh, n; g-oi oI-o .g. oo8u. o g., g,' f li yo ad -ESgS6 wa hn 0 0 T t r j 9j 4,

.l,i ij 3

C ao ~j1 i,q.i,;1 A pt T ,t;. i 4 g i. ,.4Ij. 89012 N nh 111 IS i p W m 0 .GO s-D hhhhh 5 y. , i .',1g 94' !<;I. i't Nccccc ,2 NL w0

,i F

o0 Ennnnn O3 l2 I Gaaaaa DLL f p4' i Errrrr .IJ, j; I ,,i1.I L BBBBB NUN i n AIA rd _7

= = =

0 RBR aa o t- ., + 2 0 O E .L, i! -..:.. 12. T.T d - H ., s N e0 hT t0 I t pA a1 y ,;i l I. P !.i 4 . s mD r E er ~ = 0) E ( ?. 2 - ,5 s0LT no 0SA ef a ~ 1 ( . y,o 1 IR g i 1 E AE N o3 j M GE i d I A T NG ae . /q 0 3 ns . / . h. !

,0 I

i ay c HJu. TO ra 1 E

j;..

u 5 AD oc _I,<. e-S RA T A ~ . ; { ;i

!g!i.'

f /', CEN PR e l ,/ / OO . l e f. l s. O. T 7.i t .S O i_ !jqj .e, ;r .{ 1.i .::.j: 9{:4 j,i... T 5 ac . h _ ^. .g-C gt a P gnn

,5. -

j.- , '! tt . i Li_ \\ G iii yi. .i#t.k 0 as 0 Ft 4< ~ rl op .g-e3 og' og7 .g' , g.,o l. el e3g3W pa ow t {' ,i, j t

l' i1 I1l~ Tw" g 0 n 0 i 1 l , ij t II- {-ll. I " 4t I 'l4+ 3 d o t l s a j.;,J! - e + T oi = yi

'h

.1 7i f+ 9012 S t u 1 ; (' g 111 G O b e - r L s D D 6 S a 5 th hhhh F1;i! Ncccc ='o GW rt L .} t!4 ;' !4

!. ;i i,j' N

Ennnn NO oi L O Gaaaa pw Errrr DF I 71,r4 4i + ti 7' r' ,l + il, LBBBB OL ho DIL gd

==

0 AUA . jL l*-lp ;: 6 oa+= ,0 NBN ua pli r;: i.I}!f on 2 R.R rr - + E OHT ho T f!4 i ttI4' ll-l[J T1 T t t t hAN pP lh I a' 4 O l p i4,! I 7-7i ,S )s mED 4l '. h J~ -l j> 7! em 7 IJ l I F L c - ( ~ 0ST - I', ; p i^ 0 1 ~ E 0IA fL N,. e0 M 1AR .t!; +, s1 E I o = - ~

0. T GN pr f r

,e f ;g,4 + ~ NE ro -l4! I l[I t4 !]>p]4 7 0 A IG uf j 1 3T p b EAO - ) / li' tl

. t Ily/ r/-

T RD lb S AEA a( CPN r F 0 OR e; }j . i , ; y t' lfj / - k ' 5 O nt OT ec ~ T ga A - ,+ . e \\4 C j t l, j,. 4' ) / . iq'. q{ ,t + l it P mn b / 7 - 0 G oi iq1 ( \\ j /~ i r -o"* f g 0 q og o}. ol ol o[- o}7

f/

og' o {;ol. h yt g, S. goau. ad wl 0 0 T hi s 3 C t u A ab T p . i Cn_' _:t ;t &!+11I 89o!2 N h t!1 S nt I W ii D 0 GO w hhhhh ,5 y, Nccccc 2 NL s Ennnnn IF wo OD od .=%l, J". Gaaaaa } ;4 i Errrrr DLL l a J' t LBBBBB AIA f n NUN rr

=

RBR i o TE O at ,0 tl ,,^ * = + 2 T H N dh s bl y, il AI hT ep I A g i '.. i 4 p t m t P mD a-0 EE r0 _l,5 )s 0 LT e0 .\\ - - + ' L -+. 4ii 'i 0SA n3 1 ( 'f ~ 3IR e .' [ /' t ',.h E AE gr [ / t1L if t M N - o ,. t,;I! GE oF l y;

/

I CNG d / 1f,/

0. T 3 a)

J,4 hf t I /+jj il,0 Lt]; TO n EAD r (a 1 / / S RA o + A l }' EN T } i. . ' 3' l,! .1 ~ CPR OO e ~ 0 T . l 4.' .,... u-. _..,. i ,l jjtj l P ...i ~ ; 5 O

8. is T

-.a' sS 5 a g C a )f, {... I i

s' I -

. g - ) \\ G gn ~ ( Li 0 ii r, . ' ;i 0 Ft oe 2 2 2' o,. ' 3' g7 a. rt or 7,3.,oau. es po ol {l j' l

F 3-18 L'. 't 4 i - O". og - -. _ _ _. ~.. - _........ _. I s g w . 4.~.._-... On c_ ej h E .a-... - !. 4. b,. -..... 4 o a m s'..,4

===== ; J e.=o o d$$$g '30 4 W AO 4,. ..-, 4 ~4'_ .R 1B 4 m = A"* ; .....,1 u w aywgy J .C k w3 M i cc g n i i ll ll 18 11 o pg,4 at D gu .......+..._.. ;.. ..J g n4.. i i O

W

. 1_4 _ f. _..j - - N ..y. .../-. i .4 .p 7 ..m... .s. n._ ~ e- .

EWn"o l

j w -...___,.C._._.....+,~~__.+.i..- ., ~.. 1.. - c o "d y W y *;. a 6 C y;t J' 5 i j l y g 4'* r.e . _. p.,. p, _ 7..a 2 7...q g gy g .,. %'... 4 Ls .g. + r.,_.,,. W p 807 o.u 4

i i

l j 1 ( ' 4. ( _a g .. j.s Lp_{. '..y g g3

c.,4

')_..._....,_& s ; int ; +' (7 m. l ..H j y 5 8 g,.g g ,_._.... ; _ r. j.. m. . m..+. a _. _._.. 1.3Y! J__ 4 ( < W,4 c.ue 9... _d f _ b 1 i 4 4 a-D .. h -.._ b, --f., - % 9 d ,.6 J. f y] I Ig} O= Q w 'i N; l i t i h 00 Q. 4

r.,_. Ats_.__.-.......

i .,6_.4... +. !.~. _% ' }I - l W 5m I k L-4 . )M",,-g .J -+M ' ' A d., j 1 O (* e o O w" j 4 ~ m O tt 4 25 00s 09 09 Ot et 00 07-O P-09-09-0 01- 0 21-O f f-0- N ,01. (mp).gou e.Oo 4 an 9 .C '* s +__._..,y,...__,.. g a g ,,r 1 t nO I Au . d... p - ~4 .h 9.--.h.' O c g 0] u_. a _._ L. q_, i_ p; g.=.c o.= g J mf - o. ~ a .o s H.euuu ,a 00 t j ._ +. M_ t, 0:e=ce i - W n. s e. s. Qw sm s s L _, A...

4... A. - _,,. g

.A _.._4, _ 3, 1 t an a a ~ 4- %.g =JAmmm d,,,3 i j j i, 9 6 / <D<

- - C il 11 il li o

gag' 4 m 1 L,,1 J_.. .p _ . _4 -- h._ f g o4,, i

l i

caw t I i l H H Q^ u. y - r.,.-+ a' .y af. .s ... - +. .-.._--n 1 .g,,.- g m

,..g i

hef_ .,~,.___4 Eg EWg g g-H. p<. ~ 4, i w m oc. 2 t 1 4. d. e A. 1 i I I .q: 2 4 s.. .,,f' W om { l

p I

I 07. "U ~ I p J. ._. A " = w - zw g* g e i,- i- . w! ..A s a r i I i i i t tx Wyg g o . (' m._ I .. -.. _. - e. t g j I 'N.,s [ A 7. W *O n - _.t . _._(. 'p.

w,4_,_.-

o og

~ n j

o mmx g.. v-N. I i 1 / Op

*at O

j f ~ /- F. c ec k 3 is 5 u.. n.- .t ->.. -.H A u ) gg g I I oc c e j-i 't

~ _ nlx-o n=

00500 0 0000 0 000t ' O0002 00 0 OdO2-0 OdCb 00009- 0 0009- 'h j tm m.uou g.g O c 4 ,,,,--m .,w--.- w.

q - + - - qg3

g. (, j,,

m, __'7 t ,f f.g -n , fj g .t s 3,,g caset-ow.*; si oso ( g bo b i 1 78' r8'~~--. i i i. I t + c. .s s p " " ',f _... _ _... L ' + j .g { t t -t' i o I. 's g 4 3 ..p.. o 4 ~, { ~. \\ _ I _.... d. _ _ }. _ -. _ _ / f " ~%, g. \\l \\.\\, j i \\ i I , { +-+(. j g. ,f. ~

s.nQ:M c; h u p urd

!\\,s'd ::q n r.:2_: (,- I /\\ - > x%...._ s / N l l _ o.. y. +. \\. i .,,..[...v. ... 4 w._. u o x 3 p 'l c. + e. e e e n- .._ 7 . __. i l] y j E i l t N - / f 3 g. i. /: -2 i i i bc j s: o I o g ..__e.___y_.... 4 g [ U 4 t. t _.. _? _. _. _.. ._4_. > b. ,a i l l t o o. o_. 9, .j. } 7 i l l i i i l l _4, p i i I LEGEND 3. . f. j _ __. I.EGEND I 3. _..._f 3 = Dranch 7 7 o = Ilranch 7 a . +. 7 i I I ^ = Hranch 13 ^ = Branch 13 ~~~-I-~~~~i + : Uranch 14 -t . = Branch 14 +- l l x = Branch 27 l l-l

= Branch 27 8__

2-0.0 50 10.0 15.0 20.0 25 0 30.0 00 50 10.0 15 0 20 0 25 0 30.0 TIME (s) TIME (s) CASE 4 A - 100 mph TORNADO. CASE 4D - 200 mph TORNADO GPC TO MSM BUILDING PATH. MAIN BLDG. INTACT GPC TO MSM BUILDING PATif. MAIN BLDG. INTACT TORNADO GENERATED INTERNAL FLOWS TORNADO GENERATED INTERNAL FLOWS Fig. 5.10. Tornado-generated airflows from general-purpose cell through main exhaust duct to master-slave manipulator building and atmosphere, case 4, for 100- and 200-mph tornadoes. Airflows predicted with TVEhT.

5-20 ORNL-DWG81 9883 be_ l i l l l m jf '\\'! i i ,i i 9 / l I g__..._9,__ __y._-.. i I ___ i emy\\ l o e 3 ,__q. _z ga _ _ _.i. /< 1 'g i i i/ I _i 1 s \\\\ l l l l 9 -/--- nl--t // i ee g g &. { l L \\ .Li 3__ _ l t a_4 1 _._ _.' _.. _.__. 4-, je# 0; g c / I i l l i \\' i i i i \\; i fi \\L h /_. f I s 1 i / i

\\

.a__.+-_. i _l 3-__ \\ i / 2 N _ i /! _L__.. _. _ _ p _ ..__.4 j i y!;- j l--- i i e si i T i l J i i. 1 i i i LEGEND y'_ _ _1_ _l__ _i _ _ M _ _____L. o = Branch 7 a = Branch 13 i i ! _- - = Branch 14 { j i x = Branch 27 ? 0.0 50 10.0 15 0 20.0 25 0 30.0 TIME (s) CASE 4C - 300 mph TORNADO GPC TO MSM BUILDING PATH, MAIN BLDG. INTACT TORNADO GENERATED INTERNAL FLOWS Fig. 5.11. Tornado-generated airflows in pathway from general-purpose cell through main exhaust duct to master-slave manipulator building and atmosphere, case 4C, for 300-mph tornado as predicted with TVENT. J E

1-5-21 1 The following r 'etion deceribes the. determination of two-dimensional air relocities within the head-end cells as presented _by the SOLA-ICE code. This is the data that is used with information concerning material in the cells as described in Part II to determine whether tornado-induced airflows witha. the cells will b3 sufficient to resuspend radioactive particles and carry them.from the facility to the atmosphere. 4 o v

6. TORNADO-INDUCED AIR VELOCITIES IN. HEAD-END CELLS In Sect. 5, cases 1B1, ICl, 182, and IC2 were identified as having the potential for significant radiological releases f rom the head end cells under tornado accident conditions; additional analysis was required to determine the quantities of radioactive material in these releases. q 4 An essential step in estimating amounts of material discharged is i* the determination of tornado-induced air velocities near contaminated l surfaces of the head-end cells. As will be discussed in Part II, f_ near-floor velocities above a certain magnitude can reentrain or resuspend radioactive particles which may eventually be released at the i plant's atmospheric boundaries. i Los Alamos Scientific Laboratory had developed a computer code, SO LA-ICE,6 that would address the problem of evaluating in cell velocities. The NRC requested that LASL analyze cases IAl, IB1, ICl, and 4C based lon the TVENT predictions of flows at the cell. boundaries utilizing SOLA ICE to determine near-floor air velocities in the PMC and the GPC during the tornado transient. Results for. casec 1B2 and IC2 l with. the exterior building lost would not dif fer substantially in the SOLA-ICE analysis, so these cases were not analyzed. 'The following sections describe analytical techniques utilized and analysis results for the four-specified cases. 6.1: Modeling Techniques l The procedure for determining tornado-induced air velocities within - the head-end' cells involves the following steps: (1) geometrical l modeling of the rooms ' to be a,nalyzed; (2) converston of TV3NT output predicting the magnitude and direction of system airflows into velocities. which are the boundary conditions for SOLA4 ICE; 'and (3) calculation of air velocities within the cells,.using SOLA-ICE based on.the results of steps 1'and 2. SOLA-ICE is a two-dimensional finite dif ference algorithm for solving the equations of motion for an arbitrarily compressible fluid 6 Dimensional limitations 'of.the code require simulation of the ' effects-occurring ia a three-dimensional: room with a two-dimensional model. The model used for 'the 'GPC_ (Fig. 6.1) is, essentially,1af vertical' slice through the cell' alongrits length as indicated by the cross section shown below the isometric view.E To' simulate: the effect of the ventilation exhaust duct (branch 13. in the TVENT. schematic model).. located in ' the f ront f ace of the cell shown in.the ' isometric drawing, 6-1 k y %r-y- a w 4 , [ r ', .y ,-s g e', ..--.-,e

6-2 i ORNL-0 W 79-0820 (o) AL b N '~' > N?N 't CHEklCAL 7'"ocEss g i. g c, a.. x Q x f' 4A 1 ,,= 8 CELL 1 ~ 'N y i secTion ron oE. ERAT 3, u SOLA-ICE MODE E ^ ,,,,3 ISOMETRIC. VIEW l= 475ft HATCH HATCH - TO PMC (BRANCH (4)] TO CPC (BRANCH (7)) - VENTILATION EXHAUST 19.5 f t WITH FILTERS. [ BRANCH (t))) l[ FLOOR I I l l I I JL l i I I I l 1 SECTION FOR SOLA'-ICE MODEL GPC Fig. 6.1.(a) Isometric drawing of head-end cells; (b) SOLA-ICE model of general-purpose cell, l.

6-3 the duct was rotated 90' and translated so that all flow paths would be in the same plane (this procedure is described in ref. 4). Pathways through ports into the GPC from the GOA are not included in the SOLA-ICE model of the CPC because of the dif ficulties of simulating these pathways in the two-dimensional model. The normal flow into the GOA from the GPC is less than 5% of the total flow from all pathways into the GPC. This means that the normal flows in the GPC remain' essentially unchanged by neglecting these pathways. This is not the case, however, if the port covers are lost during the transient and these pathways then become significant in allowing flow out of the GPC. When this happens, however, there is a reduction in flow f rom the GPC into the PMC, thereby decreasing the importance of radioactive release to the environment via the PMC. The SOLA-ICE model for the PMC is shown in Fig. 6.2. This model is also a vertical slice through the cell. The seven manipulator ports which are evenly distributed along the length of the cell are divided into two boundary paths on either side of the SOLA-ICE model. The TVENT predicted flow through branch 3 (the manipulator ports), which is input as a cell boundary condition for the SOLA-ICE analysis, is equally divided between the two paths. Inleakage airflow is represented by a pathway at the upper edge of the SOLA-ICE mocel. When a compromise in modeling the cells for SOLA-ICE analysis was required in order to satisfy the two-dimensional modeling restrictions, a modification was always chosen whi,ch would have conservative ef fects on the results but would simulate the actual cell conditions as realistically as possible. For example, consider the magnitude of the boundary velocities (f rom TVENT analysis) for the ducts and hatches into the cells. If an airstream flows f rom a constricted path into a larger volune, the airstream expands, reducing the velocity. This effect cannot be represented two-dimensionally but could be modeled by reducing the boundary velocities by some factor. However, because of dif ficulties in determining appropriate f actors for this analysis, a factor of I was used, which predicts higher floor velocities than would actually occur. Flows determined from the TVENT analysis are converted to cell boundary velocities. In calculating these velocities, the velocity distribution across the duct was assumed to be uniform; that is, velocity v = g/A, _where 3 is the flow, and A is the cross sectional Average velocities at the cell boundaries under tornado area. conditions are required as input for the SOLA-ICE analysis. Branches 4, 7, 12, and 13 represent the boundaries of the GPC. Transient velocities in these branches for cases IBl and ICL are shown in Fig. 6.3. Transient velocities in branches 2, 3, and 4, the boundaries of the PMC, are shown in Fig. 6.4 for cases IBl and IC'..

6-4 ORNL DwG 79-1829 (0) AD ,' s !. Q,;<< ,t CHEMICAL I 1 'o gESS Q m 's\\ Q .'l s . l 's l A s f s i MNiCAL CELL /h SEC~ ION FOR o \\ "Mg ts. ap' SOLA-!CE MODEL p CEtL scale (f t) ISOMETRIC VIEW 52 f t

t l

_ b) INLEAKAGE ( PATHS [BR ANCH (2)] ~ M ANIPUL ATOR MANIPULATOR ~ PORTS [8 RANCH (3)] PORTS 7 [ BRANCH (3)] % 2S f t ~ HATCH ~ TO GPC _ [l LOOR (BR ANCH (4) 16 1 I i 1 1 1 I1 I I I I I I I t_ o SECTION FOR SOLA-ICE MODEL PMC Fig. 6.2.(a) Isometric drawing of' head-end cells; (b) SOLA-ICE model of process mechanical cell. l I m

6-5 s o U C o c g p<g .,.I.__~- g xm -o O i . __ _ 9 _. -q { a- . _ _. _. _..... _ - -.. -. ~ _.. '"b cu g ] g , i <. o c . q _ _ f _. - _. 4 .C 7 k O i Wcccc 4m .\\___.,__._.+. .'._-.4'- . _\\ l C W' e e e e Z R &a3 .c y ___4 ,6666 gg-a UCCAC o H E s \\ H7 8 il il 11 11 o - ~. . --.i_ .p i + 4 _,_,... _.., Q - p-__ o< w _g g s. e< + a ~ j EmW . __ ; v. _ _.._._. _ __2.-_A _ u 8 c.cg> $ N li i ! i i s ( w/.. _M t i _-__. y -_ g, c'3 g,;I'3 o r. _.4

-4

-7 y .._.4_ , g t I on 4 g

i_p__,

N I t:3 U vm 4 ,_a.,. l o 5 $ce r ./_ _, r.. - _ _ 3 _ % _q _. 3- _ _.. _ _ __4__ - -- o, go u _4.__ 7 7.___. A' fi .,. *i i l UH d 8 _..,.___. l -w... w 4.+.1 '. -.,___,.., _. ' _ i * ,,34 ... +..,. j g g 3 I gZO.a i 4 3 . > -, g: t I I 4) _ _ __ _ j- _ _.- d__+ p3 ._j. .._4_.__

/H U

l / i i e i g1 =4, ) E i i g) 7_.-.-__.w._. _ __ _b - 1

/ j.

o 1 (x. ~ i 51. i / u i x-i o o o 52. 0 00 0 09 0 0t 0 02 0'O . O'02-00t-0'09-0'09-t (Sd)) A113013A b e o u 0 g 6 i -o gg 7_ j "" b $ Q oy Z ~l {_._ ._ d._ _.. I.. -_h - f. f, _. 1___ _ 1U %gggg.h 1 ,3 4 o 1 i i W==cc =.c ~ 6 4 Q $ ~l O' c. c c e .#4 D.m l ta3 W 6 666 u *4 I. i I (- M c;H + i t J Q::3 A C3 O d 3 k .__.4 h } D )e t .--.gi,___h._s__--__q__..g-Il li ll 11 o 4- _. 4 a<,w -g g.4 c 03 i s' i i i I ou g** i \\ i I E 0 4 H g c.:>c u.=c4 -- +_ -w l 4, N i p$4 I o I'_____'.1(_. Q _.... }. e q3 t i I l i o W 4 c] N.% y Q- . y-- -- +- -- f -t l l l 1 i Wewm i .s i w N_# ^ ,i T; L. p} d I i i WW u ss3 c:a.,o l 2 U3 a3 ..,. _h. ~ + { d.. k^l i :M._ _.h- - { ---- +.{ - -g \\l 1 I } G O

s_'.

+/ i o H dmo ,0.,.@ 7 g.._ y -p-go L.s /- -h [ h "y 6 n n { _s. / O 'o [ i a r l i_.._ -. - - _ -. - - + -. - --~~ c -i

"". p *.C o

6._4.. 4 ..__p...g / e g3, i. Li ._r.t_ 4(i4lg!_ ag ...y .j Le . W <, i 4 o. l 0 Oh9 Ob> 0 02 00 0 02-O& 0 09-OW Hg (sd;) AJ.13013A H ] .. p 1 4 e-

1 0%L-O*G el e865 o e 8-y-- -~!- g, _ _..r i- --r--- y-1 _ I _, j,,, 4 _ _f. _ _ p _,l e - q , 4 t. 7-j . p _ ;1 g. .__...r_-... /.. . _ A, 4.~ .,I _.

____i, i

i o 1 i = o. ...,.3 ., _. p.. r.. },._,4,..,, i p g j i g i . f.. - 4_' ...A. ..f. _...

4. '

,,...- f-i____?, _ 7_ _~ 's [ .t. _ pl _ __ o l g. _4

s. --... _.. }. _.M __.. L _.-.a_,

.J 8-q- ~ - -tfv - - + - l . _. _l. _. _ __. _ 4. _. t._. _./.t.r t...._.; _ N l i j ~ .a_ + s_f - _ -i.-___/!/,\\l _ gs o. _} ' ___,/._.I(..\\ ' _ _. j j ~ r -.x- .q - g- _ _ _ _... / t i i

N},h _[I. )q

_..'%nq- -- f s ---4

+----+---l-

_+_g__.-_.-w ..4 __.T -. % _._ - g_ N-t- } i 4_. _!_._. ;.. _ _. _ L.__.. _.. ' ._ h%; 4 __., o.- _.,.4..__. -. 7 l 1 N~* g,+4.- I .o,,,- .T~-{! y f --. j- --r d;p,, J i i .J _4 o. 4 9 _. _.4 p._._L. l t c. >o. f t 7[ 1 - - t- - ! i i h ) t D h l 1 t l -t.----- -- j .r i ocs4 t o i .c. t l"., .j._ d _ _3 . _. 4 _._- _._ _j i~ l ____j 8._ t i L i i _._L.__._i8, A i i i r. m I i d e-r- - + - ---n-g-- - 7 -- g-i U 1 y ...._1_._,__..._.._a_ q i b 3 -1 a o l 7 G _ t_4 _,i.___-_ __ _ 4 t i j ~f.-- 2 o s i g,._ .._? g -- __A ' ._ _ I. _. i - y: _.-.. _ _._,i u _ia \\ 'Q I i, v - f... .._.s._- o \\ \\y 1 _ ~ _... _ b.v. + g ._.-_t, e _..u __ ,g y I x I f LEGEND 8 M LEGEND o = Branch 2 I o = Branch 2 o E g f A / I A = Branch 3 a = Branch 3 + = Branch 4 V i + = Branch 4 g } ] g 0.0 50 10.0 15.0 20 0 25 0 30.0 0.0 50 10.0 15 0 20.0 25 0 30.0 i TIME (s) TIME (s) CASE 1B - 200 mph TORNADO CASE IC - 300 mph TORNADO - MANIPULATOR SLEEVE PATHWAY BUILDING INTACT MANIPULATOR SLEEVE PATHWAY, BUILDING INTACT TORNADO GENERATED VELOCITIES FROM PMC TORNADO GENERATED VELOCITIES FROM PMC Fig. 6.4. Velocities at boundaries of process mechanical cell for cases 1B1, 200-mph tornado, and 1C1, 300-mph tornado, as predicted with TVENT.

6-7 Velocitics in both cells for cases IB2 and IC2 with exterior walls 'ost did not dif fer substantially from those in the cases with exterior walls intact. Results of the SOLA-ICE analysis of the four cases are discussed in the following section.

6. 2 Airflow Velocity -- Time Histories for Cases Analyzed SOLA-ICE results for the four cases analyzed are displayed in Figs. 6.3 and 6.4.

Although cases lAl and 4C were deleted from further analysis, as discussed in Sect. 5, SOLA-ICE analysis of these cases was performed, and the results are shown for illustrative purposes. Streamlines through the GPC at 5-s intervals are shown in Figs. 6.5 and 6.6 for cases lAl and 4C. .The first 10 s represent steady state conditions in the cell with the tornado applied af ter 10 s and lasting for approximately 15 s. As indicated in frames 4, 5, and 6 of both figures, af ter 10 s, when the tornado is applied to the system, af rilow occurs from the floor of the CPC through a hatch in the ceiling to the adjoining PMC. This air could potentially resuspend deposited radioactive material from the GPC floor and carry the material into the PMC. The patterns of streamlines within the cells were used in Part -II I of this analysis to approximate the shapes of air volumes discharged from the cells. Velocity plots along the floor of the GPC for cases IC and 4C are shown in Figs. 6. 7 and 6.8. The first plot is a composite of t:11 the velocity vs time curves that describes an envelope for the floor l velocities; the other three plots are velocity time histories at a particular location along the floor. These velocities are used to determine particulate reentrainment in the CPC under tornado conditions, as discusted in Part II. Minimum air velocities required to reentrain particulate in the GPC (calculated in Part II) were compared with SOLA-ICE predicted air velocities along the floor of the cell. If the minimum required air velocity.were exceeded, some fraction of material on the cell floor would be resuspended and possibly carried from the cell. Velocity vector plots through the PMC for Cases lAl and IBl are shown in Figs. 6.9 and 6.10. As -indicated in f rames 4, 5,' and 6, substantial airflow occurs from the PMC floor _through the manipulator ports af ter the tornado transient is imposed on the system. The airflow patterns indicated in these figures were used to substantiate the assumption in Part II that air discharged from the PMC was originally in the shape of a quarter-sphere centered at the ports. The 20.0-s plots of Figs. 6.9 and 6.10 also indicate-that air discharged from the PMC porte might. originally be in the shape of a truncated cone centered at i

6-8 ORNL-DWG 81-14196 '14 ' 00 ':4 s S 90 W CSC pse CPC t a gt g es F thtges O GEM.A **0 CESS CEa GE g ea6 peoCgs3 Ctg e 71=E sil 60 tesE s ee G3 C8C

est C#C Put S IL rt.1 t is'tet

.L Maf es p.yf13 Ctw Made% ##Xf15 Cia t :st 325 Da truet all 00 M C8C 8'E CPL . non. ,a,,,, i =<. xisi a w.~ mis C.. 'lut 113 90 '14 sv6 00 P'at geC Pest CDC ,,u.. , m...s ...~ m s, u. .is u. Fig. 6.5. Streamlines through general-purpose cell for case lAl, 100-mph tornado, as determined by SOLA-ICE code. (Normal conditions from i 0 to 10 s; tornado impact at 10 s.) I i E

6-9 ORNL-DWG 81-14197 ...,0, 734 s 6 60

MC Cet out C#C F IL7E*1 F tift #5 4

GEwaatP40CE55 Cfu qq engpegCgst Cta flag 314 OG fist stl 60 - put Cet

M CPC 8 tL 481 FILTERS

.48A Pn30Cill Cta gqsai, pe0 CESS Cta fiaE *FO 09 t imE 25 00 84C C#C put Cpt F 8L485 f tiTiel s GE498LP90CE11 CER GEqket pe0CE55 Clu 714 230 09 T:mE s34 90 put C8C og ~ C#C rm m , m.m ss au ...~,oEss Cc. Fig. 6.6. Streamlines through general-purpose cell for case 4C, 300-mph tornado, as determined with SOLA-ICE code. I -r n,- n----a ,,r-

l 1 6-10 ORNL-0WG 81-14198 .4 l .j i j f I E j ] j ~~ 3 S -I y I l i n -E (a) E a-t N, v eEa g, ELE ENY 4Mt a * ,...x.. <.2 .,i 1 4 l 9 l E j% x l h 1 3 a I I I I 4 -i n -E e,

-E s,

ELE E NT Nu-6Es 13 ELE-EN' %*9E A 9 Fig. 6.7. Transient velocities along floor of general-purpose cell for case 1Cl, 300-mph tornado. (a) Composite of all velocity vs time (b-d) velocity-time histories at specific locations along the curves; floor.

6-11 ORNL-DWG 81-14199 .2._ ..<r .{ i.j ^ ? E k j a I 8 y.. i i

j l

-[ 1 (d) EuE*EN' %=8Ep *i .s-E.E*EN* L *?E4 ..a ,x ..w-. l i .{ i I, .I 1 I C .l r l E N / l E !N 8 l 8 4 4-t I .j i i l i i i r --.r ';wE <s, ';=E is-1 IC) (O ELE *EN* %* sea 9 E E"EN' %"9E8 '3 Fig. 6.8. Transient velocities along floor of general-purpose cell ~ for case 4C, 300-mph tornado. (a) Composite of all velocity vs time curves; (b-d) velocity-time histories at specific locations along the floor.

6-12 ORNL-DWG 81-9872 ~ INLEAKAGE INLEAKAGE ._.,-._7 PORTS PORTS PORTS -...a..= GPC GPC HAlCH HATCH INLEAKAGE INLEAKAGE PORTS PORTS PORTS . =q -. =. =..u. ..'. A L1. u :- GPC GPC HATCH HATCll INLEAKAGE INLEAKAGE s 4 PORTS ~' k ' PORTS PORTS --^~ < GPC GPC HATCH HATCH f INLEAKAGE INLEAKAGE . = -.. PORTS PORTS PORTS I GPC GPC HATCH HATCH Fig. 6.9. Plots of velocity vectors through process meet.anical cell' for case lAl, 100-mph tornado, as determined with SOLA-ICE. (Normal conditions from 0 to 10 s; tornado impact at 10 s.)

6-13 ORNL-DWG 81-9871 TIME = 0 sec INLEAKAGE INLEAKAGE PORTS PORTS PORTS '

L a_

GPC GPC HATCH HATCH INLEAKAGE INLEAKAGE . =, _. PORTS PORTS. - - - . n~. PORTS 20'

1.

GPC GPC HATCH HATCH INLEAKAGE INLEAKAGE

C..'. ~.

JY- ~ T2 : '.'} ' 2 2 ', L~11:!- ~

. E i u. u...

PORTS -

-- - PORTS

!! 20Qk'ttZr PORTS r M.~t tr _ E;; ' - _ ~ t.. s ~ . :.:b: 5.5 bi5hbb5b --~----:=. _ ~ u. = - :- 1. .2.:~.=----- GPC GPC HATCH' HATCH 1 30 sec 35 sec INLEAKAGE INLEAKAGE '05...' ~ .s _ L.. s s PORTS PORTS

- :tt.td : ;.

PORTS . -...y;m ~ _

~

'c >.7 d r; L. ' 2." '

5 :'.:.: :.

~ GPC GPC HATCH . HATCH Fig. 6.10. Velocity vectors through process mechanical cell 1.n-case 1B, 200-mph tornado, as determined with SOLA-ICE. i l 1 j

I 6-14 the ports. This shape was also used in Part TT .a 'rmining radioactive material released f rom the PMC. Velocities along the PMC floor for cases IB and IC are shown in Figs. 6.11 sad 6.12. Again, as in the CPC, these velocities were used to determine whe,ther material in the cells would be resuspended and carried from the cells by tornado generated airflows. m W

m.__ i 6-15 ORNI-DWG 81-14200 -xm,, .c. u 4.u... xm..,...,,,,,. i i ,i l i .j ,l t mm ' 5_ MM E. i-f l \\NN j"'N- \\ f m 4l t. t I - f, I i i 1 ^ -(a) t_t-tv v es= ~ '5' (b)- Eut*Ev %.walR ' S i -xm oc-u n_ :1m;.. ..sm y s_ a em. t ' -.,t - .t. 1 .i I E_.! O E_ i 7 l: s V s } d f { f 1 4, i: I-t. 1: t,s, 1:-e < s (c) (d) ca -tv s.-ere ta-tv esta r e Fig. 6.11. Transient velocities along floor of process mechanical 4 l~ cell for case 1B1, 200-mph tornado. (a)' Composite of'all. velocity vs. ~ _ b-d) velocity--time histories at specific locations along.the l time curves; ( l floor. ? 4 + 4 1 t d p.6..s- % f r- .w.-,.w s.w-w -. -. - -, ,, - my ,y. 5....,,.,--,~ -,m .. +, +,.,.. -..., y e +e-o

6-16 ORNL-DWG 81-14201 w... m.m. ..m m; .....=c.m.... vw y.. f. 1 ~ (a), ""E '5' t;c w v et" (b) ELEwEN' Nu"6CR 5 _e.oress ac ms tsa icastici wes we n% cem.t v a i [ 'l .i t E S . - - -..\\ O 2 \\ w g 5 -j f5-f \\ t V E"' '"E ' 5 ' ' (c) (d) c_t -s v sse : t_tw vece a Fig. 6.12. Transient velocities along floor of process mechanical' cell fot case 101,~300-mph tornado. (a) Composite 'of all velocity vs time curves;-(b-d) velocity-time histories at specific l locations along the ~ floor.

7. CONCLUSIONS Analysis to determine the potential for airflow from the contaminated head-end cells in the NFS facility has concluded that inadequately sealed manipulator ports in the PMC have the potential to release unfiltered air to the atmosphere in the event of a tornado strike. Winds or the depressurization accompanying a tornado could remove plywood or cardboard covers over these ports and draw unfiltered air to the surrounding corridor or to the atmosphere. Subscantial releases of air through this pathway would occur whether exterior building walls remained intact or not, and this air could contain radioactive particulate matter. Covers would also La removed from inadequately sealed ports in the GPC cell walls, and some air from this cell would be drawn into the surrounding corridor. However, this release would not be sufficient to reentrain radioactive material and would not present any significant hazard to the public. A release to the building corridors could present a cleanup problem for facility operators and should not be neglected in considerations of possible facility improvements. The other identified pathways from the cells were found not to have the potential to release unfiltered air and thus were eliminated from further consideration. For the cases in which an air release potential was identified, further analysis was required to determine whether the air released would contain radioactive materials. This analysis is described in Part II of this report, in which the amounts of material released from cell airflow pathways during a tornado are predicted. + 7-1

REFERENCES 1. W. Davis, Jr., Potential Radiological Impact of Tornadoes on the Safety of Nuclear Fuel Services' West Valley Fuel Reprocessing Plant. II. Reentrainnent and Discharge of Radioactive Materials, ORA /NUREG-80/ 2, NUREG/CR-1530, Vol. 2 (September 1981).* 2. " Western New York Nuclear Service Center Study," U.S. Department of Energy, December 1978 (TlD-28905-3). 3. " Interim Safety Evaluation I.. Nu: lear Fuel Services, Inc., and ~ New York State Energy Research end Development Authority. Western New York Nuclear Service. Center," U.S. Nuclear Regulatory Commission, Docket No. 50-201, August 1977. 4. R. W.'Andrae, R. A. Martin, and W. S. Gregory, Analysis of Nuclear Facilities for Tornado-Induced Flow and Reentrainment, Los Alamos - Scientific Laboratory, NUREG/CR-0521 (LA-7571-MS, informal report) - (January 1979).

5.

K. _H. Duerre, R. W. Andrac, and W. S. Gregory, TVENT, A Computer Program for Analysis of Tornado-Induced Transients in Ventilation Systems, LA-7397-M (July 1978).

6. - L. D. Cloutman, C. W. P ~. t, - and N. J. Romero, SOLA-ICE: A Numerical Solution Algorithm for Transient Compressible Fluid Flows, Los Alamos Scientific _ Laboratory, LA-6236 (July 1976).

7. J. E. Ayer and W. Burkhardt, " Analysis of the Effects of Abnormal Natural Phenomena on Existing Plutonium Fabrication Plants," preprint of paper. from ANS Topical Meeting on the Plutonium Fuel - Cycle,.May 2-4, 1977. 8. The Effects of Natural Phenomena on the Babcock and Wilcox Co. P_l,utonium Fabrication Plant at the Parks Township Site, Leechburg, Pennsylvania, U.S. Nuclear Regulatory Commission, NUREG-0547 (April 1979). * ' 9. " Tornado Loads for Design' and Evaluation of Structures," Shorc Course' Conducted by Institute of Disar er Research, Department of - Civil Engr., Texas ~ Tech University, Lubbock, June-27-29, 1979. 10. A. G. Croff, M. A. Bjerke, G. W. Morrison, and L'. A. Petrie, Revised Uranium-Plutonium Cycle PWR and BWR Models for the ORIGEN-Computer Code, ORNL/TM-6051 (September 1978). 11. E. D. Arnold and B. F. MAskewitz, SDC, A Shielding-Design Code for' Fuel Handling Facilities, ORNL-3041~(March 1966). Also available_ from the Radiation Shielding Information Center of Oak ~ Ridge National Laboratory as SDC, Kernel Integration Shield Design Code for Radioactive Fuel. Handling Facilities, CCC-60 -(Nov. ~ 25,1973). R-1

R-2 12. T. T. Fujita, Workbook of Tornadoes and High Winds for Engineering Applications, University of Chicago, SMRP Research Paper 165 (September 1978). 13. W. S. Cregory, C. I. Ricketts, P. R. Smith, et al, An Investigation of Analytical and Experimental Behavior of Nuclear Facility Ventilation Systems, LA-UR-79-2891 (October 1979). u J

Available' for purchase from the NRC/GPO Sales ' Program, U.S.' Nuclear Regulatory Commission, Washington, DC:20555 and/or the National Technical Information Service, Springfield, VA 22161.

Appendix A DESCRIPTION OF TVENT TVENT is a FORTRAN IV digital computer program developed by R. W. Andrae, K. H. Duerre, and W. S. Gregory of Los Alamos Scientific Laboratory to predict the response of ventilation systems to tornado generated depressurization areas. TVENT provides the analyst with the capability of determi.ning steady state and transient pressure distributions and of calculating system flows in complex ventilation ne tworks. The code was written for a CDC computer but is designed to be easily transferred from one computer to another with a minimum of change. Replacing seven source statements in the CDC version with nine source statements for other versions (IBM, Xerox, etc.) makes the conversion straightforward. The code occupies less than 150 K bytes on the IBM 360 at ORNL and is relatively inexpensive to run. TVENT is documented with a user's manual 5 and is available f rom the Argonne Code Center. The code was not verified during this job because of time limitations and primarily because data for verification of TVENT predictions are not available. Several test problems documented in ref. 5 were analyzed using TVENT, and the results gave good agreement with the same problems analyzed by LASL staff. The flow-dynamic algorithms, that is, the flow, pressure, friction, and capacitance relationships, used in TVENT are all well accepted approximations, and documentation supporting these approximations can be found in the literature. Only their use in a transient situation remains unverified. Correlation tests of this nature are in progress at LASL, but it is still too early to evaluate their results. TVENT is based on the following assumptions: one-dimensional, isothermal flow; a " lumped" parameter approximation that neglects spatial distribution of variables; and incompressible flow with fluid storage at the nodes. The lumped parameter approach includes a number of system elements or branches joined together at the upstream and downstream ends by connection points called nodes. The pressure variable of the system is lumped at the nodes. Air cleaning system components such as dampers, filters, and blowers that have a resistive nature are located within the branches of the system. The ventilation system duct work represented in the system model as a branch without a ' component also has a resistive a nature, and this f rictional resistance to flow is lumped with that for the system components in the branches of the network. The compressibility of the system is accountei for at nodal points by permitting storage of fluid at these points. l'oints within a structure with relatively large volumes, such as ro ms, cells, globe boxes, or duct plenums, are represented as nodal points where l A-1 t.

A-2 accumulation of fluid mass is allowed. The general fluid dynamic equations upon which TVENT is based are evaluated on an order-of magnitude basis, which indicates that the ef fects of inertia and shock can be neglected for Mach numbers significantly less than 1. Output from TVENT is organized so that the analyst can easily determine the effects of a pressure transient on components of a system and can check the value of pressure dif ferentials between containment areas for possible loss of confinement integrity. The code also generates line printer plots of selected pressures and flows. In order to facilitate reading of the plots and to generate plots suitable for publication, TVENT has be en modified to output a file of point pairs for plotting predicted pressure and flow transients. A plotting package, TPLOT, was written which uses the ORNL plotting software, DISSPLA, to 'erate pen and ink plots on the CALCOMP plotter. a Some problems were encountered with convergence when using TVENT for analysis of cases involvirg steep pressure gradients, in particular, for the 300 mph tornado. The numerical solution techniques selected for TVENT are susceptible to instabilities that can lead to a divergent solution. This is not a serious problem, providing certain precautions are observed to avoid this situation. Eistorically, any iterative scheme such as this suffers in the same way. When TVENT encounters a solution with a numerical instability, it aborts, of course, but first prints a diagnostic message suggesting minor changes to the model that could avoid this condition. No TVENT problem has yet been found that could not be solved. In most cases the problems were alleviated by switching node numberings across the blowers or by adding a room at the node where the problem occurred. Switching node numbers did not alter the solution results but did facilitate the solution technique, When a room was added, the volumes were so small that results were not significantly affected; however, instabilities were alleviated. Several cases were analyzed using a smootbed tornado forcing function (no sharp changes in slope) in an attempt to remove sharp variations in the flow ra te curves. The altered forcing function did not appear to af fect the results, and again, peaks in the flow rate curves were attributed to instabilities introduced to the system by model characteristic alterations. o O

Appendix B DETAILFD ANALYTICAL TECHNIQUES In view of the fact that the TVENT code was developed primarily for modeling ventilation systems and components rather than the system of interconnecting volumes and pathways analyzed :in this study, it was necessary to develop special techniques to properly model the NFS cell exhaust system. Most of these techniques were suggested by or discussed wi th authors of the code. In performing the " worst case" analysis, it was necessary to simulate complete loss of the exterior walls _ and doors of the building at tornado impact.- To avoid the necessity for generating a separate-system model for this case, the original schematic model (Fig. B.1) was modified to reflect exterior wall loss. This simulation was accomplished by the following steps: 1. steadyf state analysis of the cell exhaust model (Fig. B.1) with normal operating system pressures and flows to establish normal branch resistances; 2. generation of a restart deck of system pressures, flows and resistances at completion of steady state analysis; 3. citeration in nodes considered to be boundary nodes in the model to simulate loss of exterior walls; that is, nodes 2, 6, 9, and'10 were changed to boundary nodes in addition to nodes 1, 5, 8, 20, 21, and 25, and the rooms at nodes 2, 6, 9, and 10 were removed; 4. transient analysis with the tornado pressure function applied at interior boundary nodes, 2, 6, and 10, while other boundary nodes remain at atmospheric pressure. The most significant assumption nanociated with this procedure is that the building walls and doors are destroyed at the instant _of tornado impact. This is a conservative but 'necessary assumption, since structural analysis of the building was not performed to determine whether the walls would fail. Another modeling technique required for this study _ was a method for - simulating the dislodging of plywood and cardboard covers -over manipulator ports in the PMC and the GPC during the tornado-induced depressurization event. It was : determined that a pressure dif ferential .across the covers uf. 0.4 in. H 0 '(0.014 psi, cell pressure positive 2 relative to' surroundings) was necessary to dislodge the covers during the The 0.4 in. H O value is based 'on the estimate that pressure transient. 2 the force required to remove a cover taped to the~ cell wall is approximately 2 lb.: It was assumed that all-of the manipulator covers in l I B-1 l a l l I

ORNL-DWG 81-13178 e, 29 2e 1300 f - 3400 C'" (201 C'" o. (23: y . mEPRE SEtTS Her tf aaf a0m ~'" assu g ~{ F80es Comma 00s. CseusCAL O 26 tem Dec.= Am (! I4J00 se M $ 4 S2,000 se Mi CRANE IIO0es SMeLO 000m am0 SC#aP REts0waL ROOu ~~ 4900 CFu "t291 ^ 4500 CFu__ '. j

2 4) t ?00 II'I C }

~] 23 n

+,

c. f e, ! tr-C'" o 144.450sefti 141 CPC 6960 CFu. (Si 7 0 4425.500se hi Cr es 24 REF#f 5EgTS psFETeaYa04 FRoss CORRColt 10320 4990 CFes 2000 CFy f, 3ggP* 27) 13920 CFes a!920 Cres 2000 CF,u (2S 000 ce M I (IS 900 es hl es M ) 12 0 comet 00m ,,,e 0 3 L} ,2 33 2 44, ~' u s, o*i oss osi on tiei siti (221 in gg3 rw.n., n ,s e eepa se en 'U'" US I IA' (20a N sos 0 CFu WWi ,o (Si @D t 301 80 C'" '[ m C'" 4,,, LEGEND ,,g,,, ~ C' " i,*[ ouc7 F**cice C,, y jg

W n

a s. s0. t.300s="11 no us 000 se h a rarr e met =1 _7 =, e- -G ~ +- e o.e or, NFS CELL EXHAUST MODEL FOR TvtNT COestauTER ANALYST $ date 1-25-79 den.LJH dru DAT UCC-ND dwg. H3C14781-GOOS Fig. B.l. Schematic model of NFS cell exhaust system for TVENT computer analysis. s s'

B-3 the P!!C would dislodge at the same f r.stant, since the pressure dif ferential across each cover w >uld be the same. The 0.4 in. H O 2 pressure dif ferential would also be required to remove the covers f rom ports into the GPC. Before the tornado reaches the f acility, the cells are held at a negative pressure with respect to the surroundings. It is not until a depressurization area created by the tornado passes across the f acility that the ce)1 pressure can become positive relative to t'.ne surroundings, causing cover dislodgement. Thus the covers would be dislodged at some time during the transient analysis of the system. The capability for altering the model during a t ransient analysis was not available with TVENT; so a restart procedure was used to make the required alterations. Steps in the procedure are as follows: 1. assign a very low flow (10 cf m) to branches 3,10, and 11, (manipulator ports into the PMC and the GPC and fire protection duct into the GPC) to simulate the ports with covers intact; 2. run a steady state ar.a transient analysis of this system, using TVENT; 3. from the pressure dif ferential across the ports during the transient analysis (AP, between nodes 3 and 2 and between nodes 4 and 10), determine the time at which the pressure dif ferential across the covers reaches 0.4 in. H O (cell pressure positive relative to 2 surroundings); at this time the covers sheuld be dislodged; 4. run a steady state and transient analysis of the same system with 10 cfm flow in branches 3, 10, and 11, but stop the transient at the (determined from step 3) at which the covers are dislodged and time generate a restart deck; 5. change (lower) the resistances for branches 3, 10, and 11 in the restart deck to simulate loss of the covers; 6. restart and continue the transient analysis with covers now removed. In sten 3, it was determined that the manipulator covers over ports in the PMC (branch 3) would be dislodged 2.7 s af ter the tornado strike for case 1A, 0.9 s for case IB, and 0.6 s for case IC. The covers in branch 11 (manipulators into the GPC) would dislodge 1 or 2 s later for each case, and the cover in branch 10 (the fire protection duct to the GPC) would never be dislodged. It is not possible to stop and restart a ~ problem twice with TVENT, so the conserv itive assumption was made that the covers in branch 11 would be dislodged at the same instant as those in branch 3. Using this prore cases 2 and 4, it was found that the covers would not be dislodt the transient. For case 3, with manipulator ports into sealed, the cover in branch 11 would be dislodged but not the cova in branch 10 as long as the building exterior

B-4 valls were intact. With the building exterior walls destroyed, both covers would be dislodged. With the exception of cases involving the 100 mph tornado, in which the covers dislodge af ter 2 to 3 s, in most of the cases analyzed, results would not be significantly af fected by assuming that the covers would dislodge at the start of the transient. This assumption would substantially reduce the number of steps in this procedure. r In analyzing case 4, the pathway f rom the GPC exhaust duct to the atmosphere through the destroyed MSM building, it was necessary to simulate loss of the MSM building walls. The procedure discussed previously to simulate loss of the main process building walls was used for this case. Af ter the steady state analysis with generation of a restart deck, nodes 22, 23, and 24 were changed to boundary nodes, and the tornado pressure function was applied at node 24. Again, the conservative assumption was made that the MSM building walls would be destroyed at the instant of tornado impact. This assumption was not supported by a structural analysis of the building. One final modeling technique used in analyzing the NFS facility involved Case 2, the pathway f rom the CPC through the EDR to the atmosphere. Examination of the cell configuration block diagram shown previously (Fig. 3.3) reveals several potential release pathways from the CPC to the environment. Air cou d flow f rom the CPC through an adjoining crane room to the chemical operat.ng aisle and outside. Another flow pathway would be f rom the CPC through the adjoining scrap removal room to the atmosphere. Air could e'.so travel from the CPC through the EDR to the outside. Instead of generating a detailed model including all of these pathways, each pathway was ana yzed separately to determine which had the least resistance to airflow. The pathway with least resistance would be the path through which air would tend to flow out of the facility, and this pathway was determined to be through the EDR to the atmosphere. Resistances in the other pathways were at least twice as large as resistance in the EDR path. The other two paths were eliminated from consideration, and the model was simplified with the assumption that all airflow would be through the EDR to the outside. In each of the described modeling techniques, the purpose was to simplify the analysis with conservative assumptions while still modeling the system as realistically as yossible.

i NUREC/CR-1530,Vol. 1 ORNL/NUREG-80/1 INTERIM DISTRIBUTION t 1. W. K. Crowley 2. W. Davis, Jr. 3. E. J. Frederick s 4. R. W. Glass 5-7. L. J. Holloway 8. A. L. Lotts 9. J. A. Parsons 10. F. S. Patton 11. T. W. Pickel 12. C. H. Shappert 13. Laboratory Records, RC 14. Laboratory Records 15. ORNL Patent Section 16-17. Technical Information Center, Oak Ridge, TN 37830 18. Office of Assistant Manager, Energy Research and Development, Department of Energy, Oak Ridge Operations Office, Oak Ridge, TN 37830 e e 6

"m U S. NUCLE AR cEGULATO3Y COMMISSION NUREG/CR-1530, Vol. 1 BIBLIOGRAPHIC DATA SHEET ORNL/NUREG-80/1 4 TtTLE AND SUBTITLE (Aed Volume No. st aporcloroet\\ri 2 tleave btersk) Potential Radiological Impact of Tornadoes on the Safety of Nuclear Fuel Services' West Valley Fuel Reprocessin9 Plant a RECIPIENT'S ACCESSION NO.

7. AUTHOHS) 5 DATE REPORT COMPLE TED MON TH l YEAR L.J. Hol'oway, R.W. Andrae July 1981 9 PERFORVING ORGANIZATION N AME AND MAILING ADORESS Itac/ude 20 Codel DATE REPORT ISSUED l YEAR voNTH Oak Ridge National Laboratory September 1981 Oak Ridge, TN 37830 6**"S*'*'

8 ILeave N shi S 12 SPONSORING ORG ANIZATION N AME AND M AILING ADDRESS Itac/ woe 2,p coaal 10 PPJJECT<TASKMORK UNIT NO Division of Fuel Cycle and Material Safety Office of Nuclear Material Safety and Safeguards ii. CONTRACT NO U.S. Nuclear Regulatory Commission Washington, DC 20555 FIN B0102

13. TYPE Of REPORT Fi R.0 3 CCv E RE o lloscius ve cases) 15 SUPPLEMENTARY NOTES 14 (Leave o/a*/

16. ABSTR ACT (200 words or ress)

This report describes results of a parametric study of the impacts of a tornado-generated depressurization on airflow in the contaminated process cells within the presently inoperative Nuclear Fuel Services fuel re-processing facility near West Valley, N.Y. The study involved the following tasks: (1) mathematical modeling of installed ventilation and abnormal exhaust pathways from the cells and prediction of tornado-induced airflows in these pathways; (2) mathematical modeling of individual cell flow characteristics and prediction of in-cell velocities induced by flows from step 1; and (3) evaluation of the results of steps 1 and 2 to determine whether any of the pathways investigated have the potential for releasing quantities of radioactively contaminated air from the main process cells. The study has concluded that in the event of a tornado strike, certain pathways from the cells have the potential to release radioactive' materials to the atmosphere. Determination of the quantities of radioactive material released from the cells through pathways identified in step 3 is presented in Volume 2 of this report. 17b tDENTIFIERS OPEN ENDED TERVS 18 AV AIL ABILITY ST ATEMENT 19 SE CURITY CL ASS ITn,s reporti 21 NO OF P A iES Unclassified Unlimited 2oggggYypigj(mso,*/ s 22 PRICE NRC F oRV 335 17 776}}

ML20031F005

Contents

Text

SUMMARY

SUMMARY

ac he==

==

==

=

=

= = =

==

=

Navigation menu

ML20031F005

Text

SUMMARY

SUMMARY

ac he==

==

==

=

=

= = =

==

=

Navigation menu

Search