APPENDIX F ACCIDENTS CONSIDERED DURING INTERIM ACTIVITIES
TABLE OF CONTENTS
ACCIDENTS CONSIDERED DURING INTERIM ACTIVITIES F-1
F.1 RETRIEVAL AND STORAGE F-4
F.1.1 MITIGATION MIXER PUMPS F-5
F.1.2 INITIAL TANK RETRIEVAL SYSTEM F-5
F.1.2.1 Unfiltered Riser Release F-6
F.1.2.2 Waste Spill From Contaminated Pump F-8
F.1.2.3 Transfer Pipe Break F-8
F.1.2.4 Pressurized Spray Release F-10
F.1.2.5 Toxic Gas Release F-11
F.1.3 NEW TANK FACILITIES F-12
F.1.3.1 Pressurized Spray Releases F-13
F.1.3.2 Transfer Pipe Leaks F-15
F.1.3.3 Leaks From Failures of the Waste Tank F-16
F.1.3.4 Leaks From Waste Misrouting F-18
F.1.3.5 Pressurization of a Contaminated Process Pit F-18
F.1.3.6 Nuclear Criticality F-19
F.1.3.7 Flammable Gas Burn F-19
F.1.3.8 Tank Bumps F-20
F.1.3.9 Overheating of a Waste Tank F-20
F.1.3.10 Gaseous Release of Toxic Material F-21
F.1.3.11 Release of Materials From a Pressurized Tank
Dome F-21
F.1.3.12 Chemical Reactions Due to Waste Misrouting F-22
F.1.3.13 Aircraft Crash F-23
F.1.4 PAST PRACTICES SLUICING SYSTEM F-23
F.1.4.1 Transfer Pipe Breaks F-24
F.1.4.2 Pressurized Spray Leaks F-25
F.2 INTERIM STABILIZATION OF SINGLE SHELL TANKS F-26
F.2.1 SALT WELL TRANSFER PIPING LEAKS F-27
F.2.2 SALT WELL SYSTEM SPRAY LEAKS F-30
F.2.2.1 Spray Leak During SST to DCRT Transfer F-30
F.2.2.2 Spray Leak During DCRT To DST Transfer F-31
F.3 UNDERGROUND CROSS-SITE TRANSFER F-32
F.3.1 EXISTING CROSS-SITE TRANSFER SYSTEM F-33
F.3.1.1 Waste Transfer Line Leak F-35
F.3.1.2 Overflow of The 241-EW-151 Vent Station
Catch Tank F-36
F.3.1.3 Rupture of the Encasement and Pipeline F-36
F.3.1.4 Spray Release From a Diversion Box With
Cover Blocks Installed F-37
F.3.1.5 Spray Release From a Diversion Box With
Cover Blocks Not Installed F-38
F.3.2 REPLACEMENT CROSS-SITE TRANSFER SYSTEM F-38
F.3.2.1 Transfer Line Breaks F-40
F.3.2.2 Spray Releases F-42
F.4 ABOVE-GROUND CROSS-SITE TRANSFER F-45
F.4.1 TANKER TRAILER TRUCKS F-45
F.4.1.1 In-Transit Punctures F-46
F.4.1.2 Fire Induced Breaching F-47
F.4.1.3 Collisions and Rollovers F-47
F.4.1.4 Criticality F-49
F.4.2 RAIL TANK CARS F-51
F.4.2.1 In-Transit Punctures F-52
F.4.2.2 Fire-Induced Breaches F-52
F.4.2.3 Collisions and Derailments F-52
F.4.2.4 Criticality F-53
F.4.3 LOAD AND UNLOAD FACILITIES F-54
F.4.3.1 Spills F-56
F.4.3.2 Spray Releases F-57
F.4.3.3 Fires F-58
APPENDIX F REFERENCES F-59
LIST OF TABLES
F-1 Accident Frequency Categories F-1
F-2 Summary of ITRS Accident Releases F-7
F-3 Summary of NTF Accident Releases F-14
F-4 Summary of PPSS Accident Releases F-25
F-5 Estimated Frequencies and Maximum Volumes for Salt Well Transfer
Line Leaks F-28
F-6 Summary of Accident Releases During Pumping and Transfer of SWL F-29
F-7 Summary of ECSTS Accident Releases F-34
F-8 Summary of RCSTS Accident Releases F-40
F-9 Accident Frequencies for Trucks at the Hanford Site F-48
F-10 Fractional Release Frequencies for Rail Accidents F-49
F-11 Summary of Maximum Accident Releases from Transport Vehicles
During Cross-Site Transfers F-50
F-12 Summary of Accident Releases for the HLW Load and Unload
Facilities F-56
APPENDIX F ACCIDENTS CONSIDERED DURING INTERIM ACTIVITIES
This appendix discusses potential accidents which could occur during implementation of the proposed alternatives. The discussion includes estimates of the frequency of occurrence of the accident scenarios and the quantity of hazardous materials released during each accident. Design features and institutional and organizational controls which can prevent or mitigate potential accidents are also discussed. This appendix is organized by the systems which would be utilized by one or more alternatives identified in Section 3. Accidents can be initiated by operational events, natural phenomena, and external events and may be categorized according to their frequency of occurrence, as shown in Table F-1.
Accident Frequency Categories
Annual Frequency
Accident Description (yr-1) Category
Anticipated - May occur more 1
than once during the lifetime
of the facility
Reasonably
Foreseeable
10-1
10-2
Unlikely - May occur at some 10-3
time during the lifetime of
the facility
10-4
Extremely Unlikely - Probably 10-5
will not occur during the
lifetime of the facility
10-6
Incredible - Not credible 10-7
during the lifetime of the
facility
< 10-7 Not Reasonably
Foreseeable
It is important to distinguish between the frequency of the event that
initiates the sequence of events and the frequency of the accidental release
of hazardous materials expected to result from the sequence of events.
Equipment or component failures or human errors are common initiators of
accidents and often have frequencies in the anticipated and unlikely ranges.
Natural phenomena can also initiate accident sequences. The frequencies of
occurrence of natural events such as earthquakes exceeding Uniform Building
Code levels, 100-year floods, and maximum wind gusts are usually within the
unlikely range. Natural phenomena such as severe earthquakes, tornados, and
lightning strikes are usually in the extremely unlikely range. The frequency
of the accidental release of hazardous materials is the product of the
frequency of each event in the sequence leading to the event.
DOE orders establish a design process for nuclear facilities that ensures that
there is an inverse relationship between the frequency of occurrence of an
accident and its consequences. DOE Order 5480.23, Nuclear Safety Analysis
Reports, defines hazard categories and Section 1300-3 of DOE Order 6430.1A,
General Design Criteria, establishes a safety classification system for
structures, systems, and components. Risk acceptance guidelines used in the
design process at Hanford are defined in WHC-CM-4-46, Nonreactor Facility
Safety Analysis Manual (WHC 1988). The purpose of the risk acceptance
guidelines is to determine whether additional mitigative features need to be
incorporated into the design of a structure, system, or component and not to
define any particular level of risk as acceptable.
The principal actions to be accomplished during the interim period, 1995
through 2000, assessed by this EIS are:
. Continued removal of SWL from SSTs
. Provision of a capability for cross-site transfer of waste from the 200
West Area to the 200 East Area via a system that complies with
applicable rules and regulations
. Provision of adequate tank waste storage capacity for wastes associated
with tank farm operations and other Hanford facility operations during
the interim period
. Mitigation of flammable gas buildup in Watchlist Tank 101-SY.
Actions considered to continue mitigation of flammable gas buildup in Tank
101-SY are active mitigation using mixer pumps and passive mitigation by
retrieval and dilution of tank contents. Accidents associated with the mixer
pumps and with facilities to retrieve, dilute, and store the contents of Tank
101-SY are discussed in Section F.1.
SST interim stabilization is accomplished by removing drainable liquids from
SST salt wells using submersible pumps or jet pumps. The liquids removed are
collected in DCRTs and transferred to designated DSTs. The shielding piping
used during these operations is old and some is above ground. Accidents
associated with salt well pumping are discussed in Section F.2.
Eventually, the liquids removed from the SST salt wells and Tank 101-SY could
be processed through Evaporator 242-A in the 200 East Area to reduce the waste
volume. The SY Tank Farm and several million gallons of SWL are in the 200
West Area. Both below-ground and above-ground alternatives are considered for
the cross-site transfer of these wastes. Accidents associated with
underground transfers are discussed in Section F.3 and those associated with
above-ground transfers are discussed in Section F.4.
The primary pathway for exposure of workers and the general public as the
result of accidents associated with the actions considered here is inhalation
of tank waste released as aerosols and vapors during potential accidents. To
support the evaluation of health effects using the methods discussed in
Appendix E, discussion of each accident includes an estimate of the
"respirable volume" at the point of release. Hanford safety assessment
documents generally assume that material released as vapors and as droplets
with diameters of 50 -m (0.002 in) or less are respirable. Droplets with
diameters of greater than 50 -m (0.002 in) are assumed to condense and settle
and those smaller than 50 -m (0.002 in) are assumed to remain airborne and
evaporate to a respirable size. Calculation of airborne concentrations of
hazardous materials at downwind locations occupied by workers and members of
the general public is discussed in Appendix E.
F.1 RETRIEVAL AND STORAGE
Prior to the initiation of full-scale retrieval of tank wastes under the TWRS program, it may be necessary to retrieve and store wastes now stored in Tanks 101-SY and 102-SY. Tank 101-SY is currently designated as a hydrogen Watchlist tank as the result of GREs that have occurred in the past. In the absence of mitigative measures, concentrations of flammable gases such as hydrogen and NOx increase in the tank vapor space during GREs. If concentrations approach the LFL for the gas mixture, combustion could occur and lead to the release of hazardous materials. Flammable gas buildup is considered to be mitigated by actions which limit the concentration of flammable gases in the tank vapor space to no more the 25 percent of the LFL. The LFL is defined as 3 percent hydrogen by volume in a NOx atmosphere (PNL 1994a). Section F.1.1 discusses an active approach to mitigating the buildup of flammable gases. The active approach relies on a mixer pump to keep the waste mixed so that stratification cannot occur. Although flammable gases are still generated in the waste, they are released gradually thereby preventing the short-duration increases in gas concentrations in the tank vapor space associated with GREs. A passive approach to mitigation is to dilute the waste to dissolve the components of the sludge layer that trap gases. This can be accomplished by retrieving, diluting, and storing the waste. Accidents associated with the tank retrieval using the ITRS are discussed in Sections F.1.2 and those associated with storage are discussed in Section F.1.3. Accidents associated with PPSS, an ITRS alternative, are discussed in Section F.1.4. If storage tanks for the diluted waste are located in the 200 East Area, cross-site transfer will be necessary. Accidents associated with such transfer are discussed separately in Section F.3.1. The waste currently stored in Tank 102-SY presents a different problem. This tank is the only non-Watchlist tank in the SY Tank Farm and is used for storage of SWL and facility wastes generated in the 200 West Area. It is also used as the staging tank for cross-site transfers. Tank 102-SY currently contains 269,000 L (71,000 gal) of sludge that is classified as TRU waste. Some of the SWL contains complexing agents that could dissolve this sludge if added to Tank 102-SY. Dissolution of the sludge could result in an increase in the volume of TRU waste. This could be avoided by retrieving the sludge prior to the introduction of complexed SWL. Retrieval options considered are the ITRS and PPSS. Accidents associated with retrieval using the ITRS are discussed in Section F.1.2 and those associated with retrieval using the PPSS are discussed in Section F.1.4.
F.1.1 MITIGATION MIXER PUMPS
Accidents involving operation of the mitigation mixer pump installed in Tank 101-SY were evaluated in Environmental Assessment for Proposed Mixing Operations to Mitigate Episodic Gas Releases in Tank 241-SY-101, (DOE 1992a) and found not to result in any significant impact. The evaluation included removal of the existing slurry distributor, and installation, operation, and removal of the mitigation jet mixer pump. The pump was assumed to be operated for four hours per day, seven days per week. A spare 150-hp mixer pump from the Hanford Grout Program was modified and installed in Tank 101-SY on July 3, 1993. The ability of this pump to mitigate flammable gas buildup has been demonstrated on Tank 101-SY during approximately 1 year of testing (PNL 1994b, KEHC 1993). Based on these tests, it was concluded (PNL 1994a) that the jet mixer pump is capable of maintaining mitigation indefinitely and a long-term pump operations plan was developed for Tank 101-SY. The plan calls for operation of the pump from 5 minutes to 180 minutes per day 3 days per week. Since this is well within the operating conditions evaluated in the EA, no additional accident analysis has been performed.
F.1.2 INITIAL TANK RETRIEVAL SYSTEM
The ITRS is designed to retrieve both solids and liquids from selected DSTs. The selected DSTs include hydrogen Watchlist tanks, proposed process feed tanks, and tanks selected for retrieval to provide additional liquid waste storage space. The selected tanks include Tank 101-SY but not Tank 102-SY. The Safety Assessment, Initial Tank Retrieval Systems. Project W-211 (WHC 1995a) considers a range of accidents that could occur during ITRS retrieval of these tanks. The hazards identified in this SA are considered to be generally applicable to retrieval of DSTs, including Tank 102-SY. The SA included quantitative estimates of the following five accident scenarios involving both radioactive materials and toxic chemicals: . Unfiltered riser release . Waste spill from contaminated pump . Transfer pipe break . Pressurized spray release . Toxic gas release. The SA also evaluated a release of toxic gases from the tank ventilation system during tank drawdown. Accident frequencies and quantities of respirable radioactive materials released and release rates of "flash" gases derived from the safety assessment (WHC 1995a) are summarized in Table F-2 and discussed in the following sections.
F.1.2.1 Unfiltered Riser Release
- This accident scenario is identical to that described in A Safety Assessment for Proposed Pump Mixing Operations to Mitigate Episodic Gas Release in Tank 241-SY-101, Rev. 8 (LANL 1994). It is assumed that the primary tank ventilation system fails while a riser is open for installation of a pump. In this scenario, natural convection flow patterns are established due to the heating of the dome gas from the hotter waste surface. Based on the prior safety assessment, 327 g (0.72 lb) of suspended waste is convected to the environment in 1 hour at ground level (DOE 1992b). Based on a density of 1.4 g/ml (11.7 lb/gal) (WHC 1993a), this is equivalent to a release of 0.023 L (0.006 gal) of respirable material. The release is categorized as extremely unlikely (WHC 1995a). There are several mitigating factors that would be expected to reduce the duration and consequences of an unfiltered release from a riser. These factors include use of temporary enclosures, use of protective equipment by workers, monitoring of direct radiation and airborne radioactivity by HP Technicians, and use of work procedures specific to the task.
Summary of ITRS Accident Releases
Accident Frequency Exposure Duration Respirable Volume
Scenario (yr-1) (hr) (L)
Unfiltered Riser Extremely On-site 1 0.023
Release Unlikely
Off-site 1 0.023
Spill from On-site 1 0.00262
Contaminated Pump Unlikely 7 0.00165
0.00427
Off-site 1 0.00262
2 0.00542
3 0.00804
Unmitigated Pipe On-site 1 1.0203
Break (Seismic) Incredible 7 0.6428
1.66
Off-site 1 1.0203
7 0.6428
1.66
Mitigated Pipe Break On-site 1 0.2559
(Seismic) Unlikely 7 0.1612
0.417
Off-site 1 0.2559
7 0.1612
0.417
Unmitigated Pipe On-site 1 0.384
Break (Excavation) Incredible 7 0.242
0.625
Off-site 1 0.384
7 0.242
0.625
Mitigated Pipe Break On-site 1 0.0122
(Excavation) Unlikely 7 0.0077
0.0198
Off-site 1 0.0122
7 0.0077
0.0198
Unmitigated Spray Extremely On-site 8 4,550
Release Unlikely to
Incredible
Off-site 8 4,550
Mitigated Spray Anticipated to On-site 8 0.00033
Release Unlikely
Off-site 8 0.00033
Toxic Gas Release Anticipated Not Applicable 0.236 m3/s or
84.7 mg/s
F.1.2.2 Waste Spill From Contaminated Pump
- This accident scenario is similar to that described in A Safety Assessment for Proposed Pump Mixing Operations to Mitigate Episodic Gas Release in Tank 241-SY-101, Rev. 8 (LANL 1994). The accident scenario assumes that the pump installation is unsuccessful and that the pump is removed from the tank after it becomes contaminated with waste material. Following removal of the pump, waste trapped in the pump inlet is spilled on the ground surface. The accident scenario does not take credit for the fact that the pump assembly will be drained and bagged in plastic as it is removed from the riser. An SA for a similar event estimated that 654 L (173 gal), assuming a density of 1.4 g/ml (11.6 lbs/gal) of waste would be spilled and that 0.18 kg (0.4 lbs) would be expected to become airborne (LANL 1994). For consistency with treatment of spills in more recent SAs, the quantity of airborne respirable material is calculated using ARRs from Mishima (DOE 1993) and exposure times adjusted. The spill of 654 L (173 gal) of waste is assumed to form a pool on the surface for one hour and then soak into the ground causing the soil to remain saturated with waste for 23 hours. While waste is pooled on the surface, respirable material is released at a rate of 4.0 x 10-6 of the pool volume per hour (DOE 1993). Respirable material is released from the saturated ground at a rate of 3.6 x 10-7 of the liquid volume per hour (DOE 1993). Workers are assumed to be exposed for 8 hours and the general public for 24 hours. The resultant volumes of respirable material are shown in Table F-2. Mitigating factors that would reduce the consequences of a spill from a contaminated pump include high pressure flush rings for rinsing and bagging the pump as it is removed from the riser, use of protective equipment such as respirators, and quickly covering the spill area to prevent subsequent airborne release.
F.1.2.3 Transfer Pipe Break
- Transfer pipe breaks could be caused by
excavation accidents or BDBE. The ITRS safety assessment developed different
accident scenarios for each initiating event (WHC 1995a). Each of these two
accidents are described in the following paragraphs.
. Excavation Accident - This accident scenario assumes that excavation
equipment causes a breach of the primary transfer pipe and its
secondary containment in the interval between leak testing and
initiation of pumping. Pumping then begins at the rate of 530 L/min
(140 gpm) and causes a pool of waste to form on the ground surface.
Pumping would continue until the leak is detected. Once pumping is
stopped, waste in the ruptured line is assumed drain back into the pool
390 L (103 gal). The ITRS safety assessment assumed that pumping would
be terminated after 3 hours and that 390 L (103 gal) would drain back.
The unmitigated case assumes (WHC 1995a) that the leak is detected by
the first material balance performed after pumping begins. Considering
the time to fill the transfer line, this corresponds to a 4-hour leak.
The mitigated case assumes that the leak detection system shuts down
the pump within 5 minutes. Both cases assume a 0.5:1 dilution
(diluent:waste) of the waste spilled and that the spilled waste is
available for resuspension for 8 hours. For consistency with pipe
break scenarios for similar systems, this EIS assumes an 8-hour leak
for the unmitigated case and a 2-hour leak for the mitigated case.
Since dilution is performed in-line, the transfer line would initially
contain undiluted waste. With these assumptions, the unmitigated leak
volume would be 255,000 L (67,300 gal) and the mitigated leak volume
would be 64,000 L (16,900 gal).
The quantity of respirable material released in 8 hours is estimated by
assuming (WHC 1995a) that the pool remains on the surface for 1 hour
and then saturates the ground for the remaining 7 hours. An airborne
release rate of 4.0 x 10-6/hr is assumed for the pool and of 3.6 x
10-7/hr for the saturated soil. The quantities of respirable material
released during each phase are shown in Table F-2.
The ITRS SA (WHC 1995a) classifies both the unmitigated and mitigated
excavation accidents as incredible based on the analysis used to
estimate the frequency of RCSTS excavation accidents (WHC 1995b) but
assumes somewhat different event trees. For this EIS, the unmitigated
accident is assumed to be incredible and the mitigated accident to be
unlikely.
. Seismic Accident - The seismic transfer pipe break accident is
initiated by a beyond design basis accident and causes rupture of the
primary transfer pipe and its encasement. The ITRS SA (WHC 1995a)
assumes that the unmitigated leak continues for 8 hours at 530 L/s (140
gal/s). This is considered to be an incredible event. The quantity of
respirable material released is estimated by assuming that the pool
remains on the surface for 1 hour and then saturates the ground for the
remaining 7 hours. An airborne release rate of 4.0 x 10-6/hr is
assumed for the pool and of 3.6 x 10-7/hr for the saturated soil. The
quantities of respirable material released during each phase, assuming
undiluted waste is spilled, are shown in Table F-2.
The ITRS SA treats the mitigated case as an impossible event on the
basis that the system is designed to withstand a design basis
earthquake. For this EIS, the mitigated case is treated as a 2-hour
leak that remains available for resuspension for 8 hours and is
considered an unlikely event. The quantity of respirable material
released is estimated by assuming that the pool remains on the surface
for 1 hour and then saturates the ground for the remaining 7 hours. An
airborne release rate of 4.0 x 10-6/hr is assumed for the pool and of
3.6 x 10-7/hr for the saturated soil. The quantities of respirable
material released during each phase, assuming undiluted waste is
spilled, are shown in Table F-2.
F.1.2.4 Pressurized Spray Release
- Spray releases may be defined as a
pressurized release of a liquid from a hole/opening such that droplets and
mists are formed. Some of the droplets and mists remain airborne and are
released to the environment. Unmitigated and mitigated accident scenarios
have been evaluated for pressurized spray releases.
. Unmitigated Spray Release - This accident scenario involves a spray
release in a pump pit, which is advertently left uncovered during the
transfer operations (WHC 1995a). Equipment in a pump pit includes
piping, motor-operated three-way valves, and pumps. Valves are welded
and are provided with double stem packing. Even though the system will
have been hydrostatically tested, it is postulated that a leak develops
through the stem packing. The length of the opening is assumed to be
5 cm (2 in), the typical circumference of a 8 cm (3 in) valve stem and
the width is assumed to be 0.1 mm (0.004 in) (WHC 1995a). These
dimensions maximize the fraction of respirable material in the spray.
The pressure of the transfer system is assumed to be 28 kg/cm2 (400
psig). The resulting total flow of 0.190 L/s (0.05 gal/s) through the
opening. Based on a respirable fraction of 0.8, respirable material
would be released at the rate of 0.158 L/s (0.04 gal/s). The spray
release continues for 8 hours releasing in the release of 4,500 L
(1,200 gal) of respirable waste.
The frequency of this unmitigated spray release is considered to be
extremely unlikely to incredible. The most significant factor
controlling the consequences of this accident is that it assumes that
the pit cover blocks are not in place. The frequency of the
unmitigated seismic pipe break is assumed to incredible. The frequency
of the mitigated case is assumed to be extremely unlikely.
. Mitigated Spray Release - This accident scenario is similar to the
unmitigated spray scenario, with the exception that the pump pit covers
are assumed to be in place during system operation. The leak paths for
dispersing aerosols from the pump pit into the environment would be
through openings around the cover blocks. The volume of waste released
is calculated using the assumptions that the pump pit drain is plugged
and that the leak detection system fails to detect accumulating liquid.
It is further assumed that escaping spray will saturate the pit volume.
Saturated air will hold 100 mg/m3 [1.0 x 10-4 ounces (oz)/ft3] of
respirable liquid aerosol particles. This saturated air is assumed to
be forced from the pit at the volumetric flow rate of the leak, 1.58 x
10-4 m3/s (5.58 x 10-3 ft3/s) (WHC 1995a). The leak is assumed to
continue for 8 hours resulting in the release of 3.3 x 10-4 L (9.7 x
10-5 gal) of respirable waste.
F.1.2.5 Toxic Gas Release
- The ITRS safety assessment (WHC 1995a) evaluated a release of toxic gases from the tank ventilation system as the level of waste in the tank was reduced during a waste transfer. Gases are released as the hydrostatic pressure on the liquid decreases. The release of ammonia (NH3) and N20 were modeled using a process simulator. A drop of 92.9 cm/d (36.6 in/d) was assumed in the waste level. The simulation included release of gases already in the waste and additional gases generated by radiolysis during the drawdown period. The initial NH3 concentration was 11.3 g/L (1.5 oz/gal) and the initial average hydrostatic pressure was 1.80 atmosphere. The initial concentration of N20 was not specified. Ventilation rates of 14 and 28 m3/min (500 and 1,000 ft3/min) were considered. Concentrations of the gases were found to be greatest at the lower ventilation rate. The peak concentrations were 103 mg/m3 (1.0 x 10-4 oz/ft3) for NH3 and 256 mg/m3 (2.6 x 10-4 oz/ft3) for N20. The total gas release rate was 0.236 m3/s (8.3 ft3/s) or 84.7 mg/s (2.99 x 10-3 oz/s).
F.1.3 NEW TANK FACILITIES
New HLW tank facilities would be required to support mitigation of flammable gas potential by dilution. These facilities would be designed and constructed to current DOE high safety standards. These standards apply to the high- integrity DSTs, and require monitoring and control systems, ventilation system, double containment, and rigorous OSRs. Two new tanks and support facilities would be constructed in either the 200 West Area or at one of two potential sites in the 200 East Area. To allow maximum flexibility in the use of the new tanks, they would be connected to the RCSTS. The RCSTS is described in Section F.3.2 The operation of these new tanks would be similar to past and continuing tank farm operations on the Hanford Site. No new or unique accident scenarios would be anticipated. With new construction using current technology and lessons learned from past operations, it is anticipated that the probability of abnormal events and accidents would be lower than for similar existing facilities. It is also anticipated that the consequences, because of better instrumentation and control, would be less severe. The section discusses accidents that could occur if the NTF were constructed. The information presented is based primarily on Multi-Function Waste Tank Facility - Preliminary Safety Analysis Report, Rev. A, Vol. II, (WHC 1994a) and additional analysis related to the multi-function waste tank facility (WHC 1994b). These documents evaluated the frequency of occurrence and consequences of a large number of accident categories. These categories include: . Pressurized Spray Releases . Transfer Pipe Leaks . Leaks from Failures of the Waste Tank . Leaks from Waste Misrouting . Pressurization of a Contaminated Process Pit . Nuclear Criticality . Flammable Gas Burn . Tank Bump . Overheating of a Waste Tank . Gaseous Release of Toxic Material . Release of Materials from a Pressurized Tank Dome . Chemical Reactions due to Misrouting . Aircraft Crash. These accidents are discussed in the following sections. The frequencies, durations, and source terms (respirable volume released) of selected accidents are summarized in Table F-3. These accidents have been selected to show the full range of frequencies and consequences for the NTF.
F.1.3.1 Pressurized Spray Releases
- Spray releases may be defined as a pressurized release of a liquid from a hole/opening such that droplets and mists are formed. Some of the droplets and mists remain airborne and are released to the environment. The NTF would use pumps to move radioactive liquid waste to and from the waste tanks in a system of underground transfer pipes and pump and valve pits. Although constructed, tested, and controlled to high standards some leaks would be anticipated. Spray release accidents could be initiated by events such as valve failures and cracking of pipes. These initiating events are in the anticipated to unlikely range. Seismic events could also initiate spray releases. The resulting spray release can have severe consequences but can be easily mitigated by requiring the pit cover blocks to be in place. For this reason,
Summary of NTF Accident Releases
Exposure Respirable
Accident Single Event Frequency Duration Volume
Scenario (yr-1) (hr) (L)
Pressurized Spray 5 x 10-4 to 1 x 10-5 On-site 31.4 0.00304
Release Unlikely to Extremely
(Mitigated) Unlikely
Off-site 94 0.0091
Pressurized Spray 2 x 10-7 to 8 x 10-8 On-site 8 625
Release Incredible to Not
(Unmitigated) Reasonably Foreseeable
Off-site 8 625
Tank Leak (BDBE) 7 x 10-6 to 8 x 10-8 On-site 1 1.76
Extremely Unlikely to 7 1.11
Not Reasonably 2.87
Foreseeable
Off-site 1 1.76
23 3.63
5.39
Misrouting Leak 2 x 10-4 to 4 x 10-6 On-site 1 1.45
Unlikely to Extremely 7 0.915
Unlikely 2.36
Off-site 1 1.45
23 3.01
4.46
Pressurization of 3 x 10-3 to 1 x 10-4 On-site 2 4.5 x 10-4
Contaminated Pit Unlikely
(Unmitigated)
Off-site 2 4.5 x 10-4
Tank Overheating 5 x 10-5 to 5 x 10-7 On-site 8 0.80
(Unmitigated) Extremely Unlikely to
Incredible
Off-site 24 2.39
pit cover blocks are designed to Safety Class 1S and are required to be in
place (WHC 1994a). The confinement provided by the pit structure also
prevents the direct release of leakage to the soil column.
A large range of pressurized spray leaks has been evaluated by Muhlestein
using a quantified event tree approach (WHC 1994b). Spray leaks initiated by
human error or equipment failure were found to be more likely than those
initiated by seismic events (WHC 1994b). Regardless of the initiating event,
spray leaks were found to be in the unlikely to extremely unlikely category
and result in small releases when cover blocks are in place. This section
considers a spray leak in a process pit with cover blocks in place when most
other safety systems have failed. Section F.1.3.2 considers a similar spray
leak in transfer piping in a pit when the cover blocks are not in place.
The accident frequency and release quantity correspond to a spray leak in a
NTF valve pit during a waste transfer. The cover blocks are in place but the
ventilation system is inoperative, the pit drain is blocked, and the pit leak
detection is inoperative. This sequence of events has an estimated frequency
of 5 x 10-4 to 1 x 10-5/yr (WHC 1994b). Under these conditions, the air inside
the pit is assumed to become saturated with vapor from the spray and to be
forced out through small spaces around the cover blocks. The release is
terminated when liquid begins to overflow to the ground surface around the
pit. This is estimated (WHC 1994a) to require 94 hours and to result in the
release of a total of 0.0091 L (0.0024 gal) of respirable liquid. If the
release begins at the start of a worker's shift, the worker could be exposed
for 31.4 of the 94-hour release. It is conservatively assumed that the off-
site individual is exposed for the entire 94 hours.
F.1.3.2 Transfer Pipe Leaks
- The waste to be stored in the NTF would be transferred into, out of, and among the tanks via encased underground transfer pipes. If the NTF is constructed in the 200 West Area, encased transfer piping would connect the NTF transfer pit to RCSTS Diversion Box 1. If the NTF is constructed at either of the two sites in the 200 East Area, encased transfer piping would connect the NTF transfer pit to an RCSTS Diversion Box 2 that would be added for this purpose. It is anticipated that the distance between transfer pit and diversion box would be greater for an NTF in the 200 East Area. As discussed in Section F.1.3.1, transfer pipe leaks can be considered as a type of pressurized spray leak. Some of these events result in no release, some in release to the soil column, some in release of aerosols, and some in both soil column and aerosol releases. To bound the entire category, this section considers an unmitigated spray release from a leak detection riser in a transfer pit. The event is assumed to be initiated by a seismic event during a waste transfer when the pit cover blocks are not installed. The earthquake causes rupture of the primary transfer pipe but leaves the secondary encasement pipe intact. This creates the potential for pressurization of the transfer pipe leak detection system and a large spray release directly to the atmosphere. The release is maximized by assuming that the following events also occur: . Leak detection fails in the transfer pipe . Transfer pit drain is blocked . Leak detection in the diversion box fails . Transfer pipe leak detection riser flange is not tight . Leak detection riser is pressurized. The estimated frequency for the sequence of events is 2 x 10-7 to 8 x 10-8/yr (WHC 1994b). Based on adjustment of the size of the opening in the riser flange to optimize the production of respirable aerosol, a release rate of 0.0217 L/s (0.006 gal/s) was calculated (WHC 1994a). Leach (WHC 1994a) assumed workers were exposed to this aerosol for 8 hours and the public for 24 hours. Since a spray of this magnitude would be visible at some distance, it is assumed for this evaluation that the spray release is detected and terminated in 8 hours. Workers and the public would then be exposed to 625 L (165 gal) of respirable aerosol. The accident scenario developed by Muhlestein (WHC 1994b) postulates a BDBE as the initiating event but reduces the frequency of occurrence by assuming independent failure of design features such as pipe and pit leak detection systems. Because a BDBE is assumed, no design feature can be assumed to function. The leak detection riser would not be pressurized unless the primary pipe ruptured but the secondary pipe remained intact. Although no quantitative estimate is available, it is anticipated that this reasoning would yield an accident frequency in the range of extremely unlikely to incredible.
F.1.3.3 Leaks From Failures of the Waste Tank
- The NTF would be constructed utilizing the best current technology and lessons learned for DSTs and instrumentation. However, it is assumed that leaks from failure of the primary waste tank could occur either through corrosion or a seismic event. The following features may mitigate an HLW tank leak: . Level indication . Secondary (annulus) tank containment . Liquid detection system in the annulus . Ventilation flow in the annulus . Annulus ventilation continuous monitor . Annulus HEPA filter system. If all of these systems should fail, the entire contents of the tank could be released to the soil column. Some of the waste could reach the surface and be released to the atmosphere, particularly if there were large cracks in the primary and secondary tank shells. A BDBE is considered to be the only credible initiating event. Muhlestein (WHC 1994b) estimated the frequency of a BDBE initiated simultaneous failure of all of these systems at 7 x 10-6 to 8 x 10-8/yr. To estimate the consequences of the loss of the entire contents 4.39 million L (1.16 million gal) of a new DST, Leach assumed (WHC 1994a) that 10 percent of the volume formed a pool on the ground surface and was released to the atmosphere at a rate of 10-10 per second. The release was assumed to occur for 24 hours. For consistency with more recent SAs (WHC 1995c), modeling of release to the atmosphere has been changed for this evaluation. The very conservative assumption that 10 percent of the tank volume reaches the surface is retained. The pool is assumed to be present for 1 hour and the ground is assumed to remain saturated with waste liquid for an additional 23 hours. A release rate of 4 x 10-4/hr is assumed for the pool and 3.6 x 10-7/hr for the saturated ground. Using these assumptions, a respirable volume of 2.87 L (0.76 gal) would be released during 8 hours and of 5.39 L (1.42 gal) during 24 hours, the assumed exposure time of worker and the public, respectively. The method used to estimate the accident frequency assumes a BDBE but also assumes that the failure of each design feature remains independent of the others. It could be argued that the occurrence of a DBE would cause simultaneous failure of all design safety systems and, in this case, the frequency of the accident would be equal to the frequency of the earthquake, 1.44 x 10-5/yr.
F.1.3.4 Leaks From Waste Misrouting
- Misrouting during waste transfers could result in transfer of liquid waste to an open pipe in a process pit. This misrouting could result in either an overflow of the process pit or structural damage to or overflowing of the tank to which the process pit drains. The following features serve to mitigate the consequences of such a misrouting: . Level indication differences between withdrawal and receiving tanks . Pit drain . Liquid detection system in the process pit. The frequency of a waste misrouting event in conjunction with failure of the potentially mitigating features is estimated between 2 x 10-4/yr and 4 x 10-6 per year (WHC 1994b). If the process pit drain is blocked and the pit liquid detection system does not operate, liquid would fill the pit and leak out, forming a pool at grade. The transfer pump pit could be filled in less than 1 hour at a transfer rate of 6.31 L per second (100 gpm). Since material balances are performed at approximately 2-hour intervals the transfer pit could overflow. Leach assumed (WHC 1994a) that the overflow was detected after 16 hours and that an additional 8 hours was required to cover the spill. Using a combination of pool and saturated soil release rates (WHC 1995c), 2.36 L (0.62 gal) of respirable material would be released in 8 hours and 4.46 L (1.18 gal) in 24 hours. If the pit drain is open and the pit liquid detection system does not operate, the tank might be damaged from overfilling. However, the soil release consequences would be less than from a process pit, because the tank free space at the maximum liquid level is greater than the volume of the process pits. If a tank failure was caused by overfilling, the consequences would be bounded by the postulated full tank leak accident discussed in Section F.1.3.3.
F.1.3.5 Pressurization of a Contaminated Process Pit
- During long-term operation, process pits would become contaminated with dried waste solutions resulting from spills or leaks. Contamination is normally confined by the induced flow HEPA filter system. An accident involving pressurization of a process pit might be caused by the inadvertent transfer of incompatible materials from a process pit resulting in failure of a test port during a pneumatic testing of the pipe encasement. Air leaking from the test port could resuspend contamination and transport it from the process pit, either via the pit ventilation system or around the cover blocks. Features that may mitigate the consequences of such an event include the process pit cover block(s), the ventilation flow path from the process pit, and the ventilation system filters. Although the frequency of this accident has been estimated (WHC 1994b) at 3 x 10-3 to 1 x 10-4/yr, its consequences are very small, even without filtration of the release (WHC 1994a). Leach estimated a release of 4.5 x 10-4 L (1.2 x 10-4/gal) of respirable material.
F.1.3.6 Nuclear Criticality
- A criticality SA of Hanford Site HLW tanks has concluded that the waste is highly subcritical (WHC 1994c). The subcritical nature of the waste in the waste tanks was demonstrated by evaluating the two independent criticality parameters. The first criticality parameter is the concentration of Pu in the waste. If the Pu concentration in the waste is less than critical Pu concentration, the waste is subcritical. The safe critical Pu concentration is 2.6 g/L (0.02 lbs/gal). The highest measured Pu concentration in the HLW solids was 0.35 g/L (0.003 lbs/gal). The second criticality parameter is the relative amounts of neutron absorbers and Pu in the waste. If the ratio of the mass of neutron absorbers to Pu mass is less than the minimum subcritical ratio for any neutron absorber present, the waste is subcritical. Analysis has shown that each solid waste sample contained sufficient neutron absorbers to be subcritical. Because the waste in the NTF waste tanks would be highly subcritical, no facility-specific criticality accident scenarios were developed for the NTF.
F.1.3.7 Flammable Gas Burn
- A major function of the NTF would be to provide mitigation for an accident involving a flammable gas burn event. The NTF would be constructed and operated to provide state-of-the-art protection from this event. The NTF's key prevention measures are: Reliable ventilation to prevent high gas concentrations in the tank vapor space and good mixing to prevent gas accumulation in the waste. In addition to the accumulation of a flammable gas mixture, there must be an initiating source to cause a burn. Factors such as electrical sources, sparks, and lighting have been considered in assessing this risk. It has been concluded that the frequency of flammable gas burn is in the range of 1 x 10-6 to 1 x 10-7/yr (WHC 1993b). A more recent review, using a quantified event tree approach, estimated the frequency of a flammable gas burn in NTF tank without primary tank or riser failure at extremely unlikely to incredible (WHC 1994b). Evaluation of the results of mitigation mixer pump test data from Tank 101-SY led to the conclusion that the mixer pump can mitigate flammable gas buildup indefinitely in this tank (PNL 1994a).
F.1.3.8 Tank Bumps
- A tank bump is a sudden release of vapor resulting from the loss of circulation in the waste and resultant buildup of temperature gradients in a waste tank. Tank bumps occurred at the Hanford Site between 1954 and 1968 in HLW tanks containing waste with a high radioactive heat load. Existing high-heat load waste tanks are equipped with airlift circulators to prevent buildup of temperature gradients. The NTF storage tanks would be equipped with mixing pumps that can serve the same purpose. Tank bumps are considered to be bounded by the tank overheating scenario described in Section F.1.3.9.
F.1.3.9 Overheating of a Waste Tank
- The temperature of HLW in the NTF would be controlled by two parallel ventilation systems. These ventilation systems would be designed to prevent thermal damage to the tank and to prevent boiling of the waste. If the waste in the tanks boils, excessive quantities of vapor could enter the ventilation train and result in the loss of filtration. Overheating could be caused by addition of incompatible wastes or process chemicals resulting in an exothermic reaction. The heat generated would be in addition to the radioactive decay heat of the waste and the heat produced by the operation of the mixing pump. If one of the tank ventilation systems fails and the mixing pump continues to run, the heat production rate could exceed the remaining cooling capacity. A longer term failure could result in partial crystallization of the waste. This is considered to be an operational problem. For a relatively large airborne release to occur, the filtration on the once-through ventilation system must also be ineffective. Because the existing tank waste and transfer criteria are not completely defined, it is not possible to calculate the possible heat production due to adverse reactions. Leach evaluated (WHC 1994a) a scenario that assumes that the tank contains waste with the design-basis radioactive-decay heat-load, that no ventilation is operating, that there is no filtration on the once- through ventilation system, and that the mixer pump continues to operate for 24 hours under these conditions. It was assumed that all the heat from mixer pump operation was used to boil off waste. The frequency of this sequence of events was estimated (WHC 1994b) at 5 x 10-5 to 5 x 10-7/yr. Leach estimated that 0.80 L (0.211 gal) of respirable liquid would be released in 8 hours and 2.39 L (0.63 gal) in 24 hours.
F.1.3.10 Gaseous Release of Toxic Material
- Leach (WHC 1994a) chose two scenarios to evaluate gaseous release of toxic materials from the NTF. The first scenario assumes a sudden increase in ammonia release resulting from a rollover (burp) or sudden mixing of the contents of the waste tank. The second scenario assumes an increase in ammonia release rate due to elevated temperature in the tank. It was assumed that the off-gas condenser was inoperable and that no other device in the ventilation system was effective in reducing ammonia release. The ammonia release rate for the burp scenario is 4,200 mg/s and 229 mg/s for the high-temperature scenario. Muhlestein estimated (WHC 1994b) the frequency of both the these scenarios at 1 x 10-2 to 3 x 10-4/yr.
F.1.3.11 Release of Materials From a Pressurized Tank Dome
- A loss of the heat removal capability of both ventilation systems combined with continued full-power operation of both mixing pumps and the maximum radioactive heat load is not sufficient to pressurize the tank dome space. The adiabatic boil- off rate is less than the design volumetric flow for the NTF once-through ventilation system (Cloud 1993). Chemical reactions and unregulated high- volumetric steam flow into the tank dome space may lead to pressurization of the tank dome space. These postulated events and their consequences are considered to be bounded by other accident scenarios.
F.1.3.12 Chemical Reactions Due to Waste Misrouting
- Section F.1.3.4
considered spills resulting from misrouting of waste transfers to open pipes
and overfull tanks. This section discusses accidents due to chemical
reactions where waste is transferred to a tank containing incompatible waste.
There are chemical reactions, other that those that generate hydrogen gas,
that are of potential concern to the NTF. These reactions do not necessarily
require incompatible waste mixing to occur. Chemical reactions have been
studied for existing tank waste (WHC 1993b) and are briefly summarized as
follows.
. Nitrated Organics - Organic compounds used at the Hanford Site are
present in the waste as salts and are mixed with nitrates. There are
two types of reactions to consider; first an exothermic reaction between
nitrates or nitrites and organic compounds, and second exothermic
reactions involving nitrate and organic compounds in the presence of
uranium, sometimes called a "red oil" reaction. A "red oil" formation
requires a high temperature, exceeding 135yC (275y F) and cannot be
formed in the alkaline wastes stored in the Hanford Site waste tanks.
To produce a nitrate or nitrite salt reaction with organic compounds,
the following four conditions are necessary:
- High concentration of organic material (ARH 1976)
- An optimum near-stoichiometric mixture
- High temperature, greater than 200yC (392y F) (ARH 1977)
- Water content less than 20 percent by weight (ARH 1976).
. Organic Materials - Organic materials identified in this section could
also produce flammable gases by radiolytic decomposition. This is
essentially the same mechanism as the hydrogen burn described in Section
F.1.3.7 and is limited and controlled by the same mitigation effects.
. Ammonium Nitrate - Ammonium nitrate is stable at standard temperature
and pressure (i.e., normal tank conditions). However, the aqueous
alkaline environment of the waste tanks could drive the equilibrium
toward ammonia which could be released to the tank vapor space. If
ammonia gas and nitrogen oxides exist together, then ammonia could
combine with nitrogen dioxide to form ammonium nitrate.
. Ferrocyanide Reactions - Early studies (DOE 1987) suggested that a
ferrocyanide reaction could cause an explosion within the tank. This
scenario was subsequently deemed not Reasonably Foreseeable (WHC 1994d).
F.1.3.13 Aircraft Crash
- The airspace above the Hanford Site is used by many types of aircraft, typically on a limited basis. The Hanford Site airspace is used by commercial air carriers, air taxis, general aviation, military aviation, contracted pesticide and herbicide aerial applicator aircraft, and Hanford Site bioscience surveillance aircraft. Commercial aircraft are considered to present the most threat to the NTF on the basis that they are the largest aircraft to routinely overfly the Hanford Site. Muhlestein estimated a frequency of approximately 8 x 10-8/yr for airplane crashes at the Hanford site. Based on this frequency, accidents involving airplane crashes are not reasonably foreseeable.F.1.4 PAST PRACTICES SLUICING SYSTEM
A PPSS is being designed to retrieve the contents of Tank 106-C. Tank 106-C is a SST with a capacity of 2,000,000 L (530,000 gal) and now contains approximately 746,000 L (197,000 gal) of sludge. Because of decay heat from the large quantity of 90Sr in the sludge, approximately 23,000 L/mo (6,000 gal/mo) of water are added to the tank to prevent drying of the sludge. An EA and FONSI have been issued for PPSS retrieval of Tank 106-C (DOE 1995). For this EIS, it is assumed that PPSS design can be adapted for retrieval of the 269,000 L (71,000 gal) of sludge in Tank 102-SY. To evaluate accidents that could be associated with this application of the PPSS, scenarios and source terms taken from Preliminary Safety Evaluation for 241-C-106 Waste Retrieval, Project W-320 (WHC 1994e) have been modified for better consistency with safety analyses for other systems considered in this EIS. Two types of accident scenarios have been evaluated, Transfer Pipe Breaks, Section F.1.4.1, and Pressurized Spray Leaks, Section F.1.4.2.
F.1.4.1 Transfer Pipe Breaks
- Transfer pipe breaks could occur as the result of BDBEs, excavation accidents, and operational failure such as corrosion or blockage of the pipe. If secondary containment is also breached, waste would be released to the environment. The spill would include waste pumped prior to recognition of the accident and shutdown of the pump as well as drainback of waste in the pipeline. The PPSS for Tank 106-C has a pumping rate of 350 gpm and includes 610 m (2,000 ft) of piping to allow circulation of sluicing fluid between Tanks 106-C and 102-AY. There is no design for a PPSS for Tank 102- SY; pipe runs would be expected to be much shorter and drainback is considered to be negligible. A seismic event would be the most likely event to rupture both pipes; however, when drainback is ignored, the volume spilled would be the same whether one or both pipelines were ruptured. The PSE for Tank 106-C (WHC 1994e) assumed a 2- hour unmitigated leak and a 10-second mitigated leak for the seismic event. The duration of the unmitigated leak is not based on any particular design or administrative control. The duration of the mitigated leak is based on the activation of a seismic cutoff switch. Based on pumping rate alone, a 2-hour leak would spill 159,000 L (42,000 gal) and a 10-second leak would spill 221 L (58 gal). It was assumed that 10 percent of the leak reaches the surface where it becomes resuspended at a rate of 1 x 10-5/hr. These assumptions result in smaller releases of respirable material than obtained using assumptions that are more consistent with recent safety assessments for other systems considered in this EIS. More recent accident analyses assume an 8-hour unmitigated pipe leak and a 2- hour mitigated pipe leak. These times assume detection of the leak during a work shift change and by periodic material balance, respectively. The entire volume of the spill is assumed to remain on the surface for 1 hour and to maintain saturation of surface soil for the remainder of the accident. These assumptions yield leak volumes of 636,000 L (168,000 gal) for the unmitigated leak and 159,000 L (42,000 gal) for the mitigated leak. The resuspension rate of the liquid pool is 4.0 x 10-6/hr and 3.6 x 10-7/hr for the saturated soil. The quantities of respirable materials that would be released under both sets of assumptions are shown in Table F-4.
Summary of PPSS Accident Releases
Accident Frequency Exposure Duration Respirable Volume
Scenario (yr-1) (hr) (L)
Unmitigated Pipe Incredible On-site 1 2.54
Break 7 1.60
4.14
Off-site 1 2.54
23 5.27
7.81
Mitigated Pipe Unlikely On-site 1 0.636
Break 7 0.401
1.04
Off-site 1 0.636
23 1.32
1.96
Unmitigated Spray Extremely On-site 2 831
Release Unlikely to
Incredible
Off-site 2 831
Unmitigated Spray Extremely On-site 8 3,312
Release Unlikely to
Incredible
Off-site 8 3,312
Mitigated Spray Anticipated to On-site 8 0.0000438
Release Unlikely
Off-site 8 0.0000438
The PSE for the Tank 106-C PPSS estimated accident frequencies were based on
the frequency of the initiating event and did not consider whether the system
was in use at the time of the accident. Using this approach, the unmitigated
seismic transfer pipe break has a frequency of unlikely (7.0 x 10-4/yr) and
the operational accident has a frequency of anticipated (1.7 x 10-2/yr). No
frequency was given for the excavation accident. This is a more conservative
approach than used to estimate accident frequencies for other systems
considered in this EIS. The unmitigated 8-hour leak is considered to
incredible and the 2-hour mitigated leak is considered to be unlikely.
F.1.4.2 Pressurized Spray Leaks
- Pressurized spray leaks could occur in valve and transfer pits due to misaligned or failed jumpers. For the unmitigated case (i.e., pit cover blocks off), the PSE (WHC 1994e) assumed a 5-cm long (2-in) 0.1-mm (0.004-in) wide crack in a valve stem operating at a pressure of 180 psig and estimated a respirable material release rate of 0.114 L/s (0.03 gal/s). As indicated in Table F-4, a 2-hour leak would result in release to the atmosphere of 831 L (220 gal) of respirable material and an 8- hour leak in the release of 3,312 L (875 gal). The PSE (WHC 1994e) estimated the quantity of respirable material released to the atmosphere for a mitigated spray leak based on the displacement of vapor- laden air from the pit as it filled with liquid emerging from a broken or disconnected jumper. An aerosol loading of 10 mg/m3 (6 x 10-7 lbs/ft3) was assumed for the vapor. The displacement rate of 0.062 m3/s (2.2 ft3/s) was based on 180 psig through a 5-cm (2-in) diameter, squared-edged orifice. Under these conditions, 4.38 x 10-5 L for respirable waste would be released to the atmosphere in 8 hours. Based on analogy to the ITRS system and consideration of the limited operating time of the PPSS, the unmitigated spray release frequency is considered to be extremely unlikely to incredible and the mitigated spray release frequency to be anticipated to unlikely.
F.2 INTERIM STABILIZATION OF SINGLE SHELL TANKS
This section describes the activities involved in interim stabilization and characterizes the potential accidents which could result. Interim stabilization of SSTs is accomplished by saltwell jet pumping. Jet pumps are used to remove interstitial liquids from saltcake and sludge in the SSTs. Pumping rates vary from approximately 0.19 to 19 L/min (0.05 to 5.0 gpm). The interstitial liquids are routed to DCRTs through transfer piping and valve pits. Depending on the location of the SST and the DCRT, transfer piping may be double-encased in concrete or another steel pipe or single- encased (direct buried). Existing salt well transfer piping is beyond its design life and is pressure tested every 6 months to minimize leaks. Transfer valve pits are equipped with cover leak detection interlocked to the appropriate pumps and covered by heavy shield blocks. The valve pits are designed to allow several SSTs to be pumped simultaneously to a common DCRT. DCRTs are housed in underground reinforced concrete vaults. The steel tanks have capacities of either 76,000 L (20,000 gal) for Tank 244-S or 95,000 L (25,000 gal) for Tanks 244-TX and 244-U. The DCRT vaults include HEPA ventilation systems, interlocked leak detection systems, and tank sluicing systems. Permanent neutron detectors are installed in DCRT 244-TX and portable detectors can be used in steel Tank 244-S. There are no provisions for neutron monitoring in DCRT 244-U. SWL accumulated in DCRTs in the 200 West Area is presently transferred to Tank 102-SY. SWL can be transferred through the ECSTS. A number of reports have evaluated accidents associated with salt well pumping activities. Safety Study of Interim Stabilization of Nonwatchlist Single- Shell Tanks (WHC 1992) evaluated spray leaks, equipment fires, hydrogen fires, waste stability, and transfer pipe leaks. Environmental Assessment, Waste Tank Safety Program (DOE 1994) focused on issues associated with interim stabilization of Watchlist SSTs. A FONSI was issued for these activities on February 25, 1994. A more recent report, Safety Analysis; Tank Farms Waste Transfer System Leaks, Breaks, and Spray Releases (WHC 1994f), is generally applicable to interim stabilization of SSTs. This more recent report is used as the basis for evaluating accidents during interim stabilization of SSTs. The accidents evaluated are: . Salt Well Transfer Piping Leaks . Salt Well System Spray Leaks.
F.2.1 SALT WELL TRANSFER PIPING LEAKS
Salt well transfer piping includes both double-encased and single-encased steel piping. More than 80 percent of the salt well transfer piping is single-encased pipe which is 2.5 cm to 8 cm (1 to 3 in) in diameter and shielded by 0.9 m (3 ft) of soil. The pipe may be either buried or bermed on the ground surface. All of this piping is beyond its 10-year design life. Stahl (WHC 1994f) estimated a failure frequency of 3.0 x 10-7/hr-ft based on the Hanford Site soil conditions and the age of the pipe. This failure frequency was then used with pipe lengths and assumptions regarding pumping rates to estimate transfer piping leak frequencies and maximum leak volumes for salt well jet pumping of individual tank farms. The maximum leak volumes are based on maximum rather than nominal transfer pump rates and assume 16- hour releases. The results are shown in Table F-5.
Estimated Frequencies and Maximum Volumes
for Salt Well Transfer Line Leaksa
Frequency Volume (L)
Farm Tanks
200 East Area
A & AX A-101, AX-101 2.6/yr 36,300
Anticipated
BX BX-106 4.2 x 10-2/yr 18,200
Anticipated
BY BY-102, BY-103, BY-105, 3.1/durationb 72,700
BY-106, BY-109 Anticipated
C C-102, C-103, C-105, 0.34/duration 72,700
C-106, C-107, C-110 Anticipated
200 West Area
S S-101, S-102, S-103, 0.85/yr 90,800
S-106, S-107, S-108, Anticipated
S-109, S-110, S-111,
S-112
SX SX-101, SX-102, SX-103, 2.0/duration 54,500
SX-104, SX-106 Anticipated
T T-104, T-107,T-110,T-111 4.7/yr 54,500
Anticipated
U U-102, U-103, U-105, 0.46/duration 72,700
U-106, U-107, U-108, Anticipated
U-109, U-111
Source: WHC 1994f
aFrequencies are based on pipe failure only and do not include operator
error or failure of other systems such as leak detection.
bSalt well pumping expected to take less than 1 year.
The maximum leak volumes are based on the assumptions that leak detection does
not function and that the leak is not discovered for 16 hours. Based on prior
experience with similar piping, it was assumed that the entire volume reaches
the ground surface where it forms a pool for 1 hour and keeps the soil
saturated thereafter. The quantity of respirable radioactive material
released was estimated using an ARR of 4 x 10-6/hr for pools for the first
hour and an ARR of 3.6 x 10-7/hr for saturated soil for the remainder of the
accident. Workers were assumed to be exposed for 8 hours and members of the
public for 24 hours. The volumes of respirable material released are shown in
Table F-6.
Summary of Accident Releases During Pumping and Transfer of SWL
Exposure Respirable
Accident Frequency Duration Volume
Scenario (yr-1) (hr) (L)
Unmitigated Transfer 2.3 x 10-3 to On-site 1 0.363
Piping Leak (90,800 L) 1.0 x 10-2 15 0.490
Unlikely 0.853
Off-site 1 0.363
23 0.752
1.11
Mitigated Transfer 0.46 to 2.0 On-site 1 0.0727
Piping Leak (18,200 L) Anticipated 7 0.0458
0.118
Off-site 1 0.0727
15 0.981
0.171
Unmitigated DCRT Spray 1.1 x 10-5 On-site 8 28.5
Leak (207 psig) Extremely Unlikely
Off-site 24 85.5
Mitigated DCRT Spray 1.1 x 10-2 On-site 8 3.03 x 10-5
Leak (207 psig) Anticipated
Off-site 24 9.09 x 10-5
Unmitigated Salt Well 1.1 x 10-4 On-site 8 7.31
Spray Leak (80 psig) Unlikely
Off-site 24 21.9
Mitigated Salt Well 0.11 On-site 8 1.92 x 10-5
Spray Leak (80 psig) Anticipated
Off-site 24 5.76 x 10-5
Source: WHC 1994f
The frequencies shown in Table F-5 are based on current interim OSRs. It was
estimated that the frequency of salt well transfer piping leaks could be
reduced by a factor of 1,000 by requiring material balance discrepancy (MBD)
surveillances (WHC 1994f). MBD surveillances combined with restrictions on
the number and types of salt well tanks (i.e., High Source Term) could reduce
the maximum leak volume to 4,500 to 18,200 L (1,200 to 4,800 gal). Table F-6
shows respirable volumes released based on a leak volume of 18,200 L (4,800
gal) as a mitigated case. Including frequencies of 0.1 failure per year for
failure of leak detection and of 0.05 failure per year for failure to respond
to a leak detection alarm (WHC 1994b) would shift the frequency of pipe leaks
in the 200 West Area Salt Well Transfer System into the Unlikely category.
F.2.2 SALT WELL SYSTEM SPRAY LEAKS
Spray leaks could occur in SST pump pits, DCRT pump pits, and valve pits during pumping and transfer of salt well liquids. Stahl estimated (WHC 1994e) frequencies of occurrence and release rates for spray leaks during transfer of SWLs from a SST to a DCRT and from a DCRT to a DST. Stahl noted that these accidents are also applicable to Diversion Boxes 241-UX-154 and 241-ER-151 on the ECSTS and to other diversion boxes in the SST/DST and Aging Waste Facility waste transfer system. Section F.2.2.1 describes the potential accidents which could occur in SST to DCRT transfers and Section F.2.2.2 describes DCRT to DST transfer accident potential.
F.2.2.1 Spray Leak During SST to DCRT Transfer
- This accident scenario
assumes (WHC 1994f) that a jumper was improperly installed during routine
maintenance in a jumper pit providing liquid transfer from the SST to a DCRT.
The jumper misalignment is assumed not to be discovered during visual
inspections or leak testing before the pit is returned to service. The jumper
is assumed to leak at both ends. A maximum flow of 19 L/min (5.0 gpm) and
pressure of 80 psig are assumed. It is also assumed that the spray leak is
not detected for 16 hours and that an additional 8 hours are required to
terminate the leak. Workers are exposed to spray aerosol for 8 hours (one
shift) while members of the public are exposed for 24 hours.
Two accident scenarios were analyzed: an unmitigated spray release with cover
blocks not in place and a mitigated spray release with cover blocks in place.
As discussed in the following sections, these scenarios have different
frequencies of occurrence, release durations, and release rates.
. Unmitigated Spray Release - The frequency of a spray leak in pump and
valve pits is estimated to be 0.112/yr (WHC 1994f). This frequency
estimate assumes continuous operation and considers the length of pipe,
the number of gaskets and valves in a pit, and the frequency of jumper
misalignment. Due to the relatively low flow rate, it is assumed that
leak detection would not be activated. Inclusion of the frequency for
failure to replace cover blocks (0.001/yr) reduces the frequency for an
unmitigated spray release during SST to DCRT transfer to 1.1 x 10-4/yr.
The quantities of respirable liquids for 8-hour worker exposures and
24-hour public exposures are shown in Table F-5 and are based on
optimization of the aerosolization rate (WHC 1994f).
. Mitigated Spray Release - The accident scenario for a mitigated spray
release during SST to DCRT transfers from a misaligned jumper assumes
that cover blocks are in place. Leak detection functions are assumed
not to be active due to the low flow rate. These assumptions do not
alter the duration of the release relative to the unmitigated case but
do alter the frequency of occurrence and release rate. The frequency of
the event is 0.112/yr (WHC 1994f). The spray saturates the atmosphere
inside the pit and the aerosol is forced out as liquid accumulates in
the pit. Assuming a vapor loading of 10 g/m3 and a pipe pressure of 80
psig, workers would be exposed to 1.92 x 10-5 L of respirable liquids
and the public to 5.76 x 10-5 L.
F.2.2.2 Spray Leak During DCRT To DST Transfer
- This accident scenario
assumes that a jumper was improperly installed resulting in misalignment
during routine maintenance in a jumper pit providing liquid transfer from the
DCRT to another tank (e.g., Tank 102-SY or a load and unload facility) (WHC
1994f). The jumper misalignment is assumed not to be discovered during visual
inspections or leak testing before the pit is returned to service. The jumper
is assumed to leak at both ends. The DCRT pumps in the T and U Tank Farms
operate at a pressure of 207 psig and that in the S Tank Farm at 80 psig, all
with flow rates ranging from 189 L to 379 L/min (50 to 100 gpm). It is
assumed that the spray leak is not detected for 16 hours and that an
additional 8 hours are required to terminate the leak. Workers are exposed to
spray aerosol for 8 hours (one shift) while members of the public are exposed
for 24 hours.
Two accident scenarios were analyzed: an unmitigated spray release with cover
blocks not in place and a mitigated spray release with cover blocks in place.
As discussed in the following paragraphs, these scenarios have different
frequencies of occurrence, release durations, and release rates.
. Unmitigated Spray Release - The frequency of a spray leak in pump and
value pits is estimated to be 0.112/yr (WHC 1994f). This frequency
estimate assumes continuous operation and considers the length of pipe,
the number of gaskets and valves in a pit, and the frequency of jumper
misalignment. Inclusion of frequencies for leak detection failure
(0.1/yr) and failure to replace cover blocks (0.001/yr) reduces the
frequency for an unmitigated spray leak in a DCRT pump pit to 1.1 x
10-5/yr. The quantities of respirable liquids for 8-hour worker
exposures and 24-hour public exposures are shown in Table F-6 and are
based on optimization of the aerosolization rate (WHC 1994f).
. Mitigated Spray Release - The accident scenario for a mitigated spray
release from a DCRT pump pit from a misaligned jumper assumes that cover
blocks are in place and the leak detection functions. These assumptions
do not alter the duration of the release relative to the unmitigated
case but do alter the frequency of occurrence and release rate. The
frequency of the event is 0.112/yr (WHC 1994f). The spray saturates the
atmosphere inside the pit and the aerosol is forced out as liquid
accumulates in the pit. Assuming a vapor loading of 6.2 x 10-4 lbs/ft3
(10 g/m3) and a pipe pressure of 207 psig, workers would be exposed to
3.03 x 10-5 L (8 x 10-6 gal) of respirable liquids and the public to 9.09
x 10-5 L (2.4 x 10-5 gal).
F.3 UNDERGROUND CROSS-SITE TRANSFER
Underground cross-site transfer could be accomplished through the use of the ECSTS or, if constructed, through the RCSTS. The ECSTS was constructed in the 1950s, and is expected to have a higher failure rate than the RCSTS which would be designed to meet current safety standards. Four of the six lines in the ECSTS are believed to be plugged. One of the remaining lines passed a pressure test in June 1995. Two major types of accidents can be hypothesized for the cross-site transfer systems. The first is a pipe break, or its equivalent. Depending on the leakage rate, location, and duration, the liquid waste released could remain totally below grade or could reach the surface and be released to the atmosphere. The second is a spray leak. Spray leaks could occur in a diversion box or in pump and valve pits. Depending upon the spray characteristics (flow, geometry, particle size) and pit status (cover in place or not), the consequences of spray leaks could range from very low to severe. Although the types of accidents that could occur during operation of the ECSTS and RCSTS are similar, their frequencies of occurrence and release rates are different. Accidents involving the ECSTS are discussed in Section F.3.1 and those involving the RCSTS in Section F.3.2.
F.3.1 EXISTING CROSS-SITE TRANSFER SYSTEM
The ECSTS is comprised of the 241-ER-151 Diversion Box located in the 200 East
Area, 241-UX-154 Diversion Box located in the 200 West Area, 241-EW-151 Vent
Station located at the high point between the 200 East and West Areas, and
approximately 5.6 km (3.5 mi) of concrete-encased stainless steel pipe that is
buried at depths of 1.5 to 5 m (5 to 15 ft). There are six ECSTS pipelines,
of which four are believed to be plugged. Each has a nominal pipe size of 8
cm (3 in). There are 58 encasement test risers spaced regularly along the
concrete encasement between the diversion boxes. These test risers provide
access to the encasement void space for leak detection. The concrete
encasement slopes in both directions from the vent station and drains into a
catch tank at each diversion box. Catch tank contents can be transferred to a
designated DST and held for later processing. Additional descriptive
information regarding the ECSTS is in Chapter 3 of this EIS.
Five events associated with the ECSTS are considered to adequately encompass
the range of plausible accident scenarios. They are:
. Waste transfer line leak
. Overflow of the 241-EW-151 Vent Station Catch Tank
. Rupture of the encasement and pipeline, either by an earthquake during a
transfer of waste, or by excavation activities with a subsequent
transfer of waste
. Spray release from a diversion box with the cover blocks installed
. Spray release from a diversion box with the cover blocks not installed.
Accident scenarios for the first three events were developed in Operational
Safety Analysis Report, Cross-Country Waste Transfer System, (WHC 1989) and in
Safety Analysis; Tank Farms Waste Transfer System Leak, Breaks, and Spray
Releases, (WHC 1994f). Frequencies of occurrence and release rates for these
five accidents are summarized in Table F-7.
Summary of ECSTS Accident Releases
Respirable
Accident Frequency Exposure Duration Volume
Scenario (yr-1) (hr) (L)
Waste Transfer Line On-site N/Aa N/A
Leak Anticipated
Off-site N/A N/A
Overflow of the On-site N/A N/A
Vent Station Catch Unlikely
Tank
Off-site N/A N/A
Rupture of the On-site 1 0.318
Pipeline and Unlikely 7 0.200
Encasement 0.518
Off-site 1 0.318
23 0.658
0.976
Unmitigated Spray 1.1 x 10-5 On-site 8 28.5
Release w/cover Extremely
Blocks not Unlikely
Installed
Off-site 24 85.5
Mitigated Spray 1.1 x 10-2 On-site 8 3.03 x 10-5
Release w/Cover Anticipated
Blocks Installed
Off-site 24 9.09 x 10-5
Source: WHC 1989, WHC 1994f
aN/A = Not Applicable. Waste liquids are not expected to reach the
ground surface.
F.3.1.1 Waste Transfer Line Leak
- For this accident scenario, leaks could occur in an ECSTS pipeline during waste transfer. The majority of the leak volume would be retained within the concrete encasement. Plugged drain lines leading to the catch tanks from the encasement would result in waste liquid being backed up into the pipe chase of Diversion Box 241-ER-151 and/or 241-UX-154, which would overflow to encasements leading to various other sumps and catch tanks. Some of the waste liquid in these encasements could escape to the surrounding soil through joints or small cracks. Entrop (WHC 1989) estimated that up to 79,500 L (21,000 gal) of waste could be released into the encasement and that 10 percent of this volume would be released to the soil through small cracks in pipeline or diversion box encasements. Since the hydrostatic head for this ground release would be very small, the flow rate at any single leak point would likely be very low, approximately 2 L/min (0.5 gpm). Based on this very low flow rate, Entrop concluded that no waste liquid would reach to ground surface. Accordingly, Table F-7 shows no release of respirable material for this accident. An ECSTS waste transfer line leak is an anticipated event based on an analysis of transfer pipe failure rates by Stahl (WHC 1994f). Stahl estimated the failure rate of 3.0 x 10-7 per hr-ft for pipes similar to that in the ECSTS. The ECSTS is 5.6-km (3.5-mi) long and has a nominal pumping rate of 190 L/min (50 gpm). If the ECSTS were used to transfer all the SWL from the 200 West Area and the free liquid from Tank 102-SY, several leaks of this type would be expected to occur in the ECSTS. If drain lines within the diversion box were plugged or if there were large cracks in the ECSTS encasement, a significant volume of waste liquid could reach the ground surface. The integrity of the concrete encasement cannot be directly determined at this time. The consequences of waste transfer pipe leaks under these conditions are bounded by the encasement rupture accident considered in Section F.3.1.3. Several factors would tend to reduce the frequency of waste transfer pipe leaks. Current practice at the Hanford Site is to perform a hydrostatic test prior to each use of an ECSTS pipeline. This test entails pressurizing the pipeline to a predetermined pressure and subsequent monitoring of the pipeline pressure for a period of time to ensure that no leak path (as would be revealed by a pressure drop in the pipeline) exists. Leakage could be detected through inspection of encasement test risers, recognition of a discrepant material balance, or liquid level rise in a catch tank. Leaks could be detected by a conductivity probe located inside the ECSTS Diversion Boxes.
F.3.1.2 Overflow of The 241-EW-151 Vent Station Catch Tank
- An open vent valve on an ECSTS pipeline could cause the 241-EW-151 Vent Station Catch Tank to overflow into the pipeline encasement. To pump liquid into the 241-EW-151 Vent Station Catch Tank, it would be necessary to have an obstruction in the pipeline somewhere past the vent station and have the vent station vent valve left open. This scenario is unlikely. Liquid entering the 241-EW-151 Vent Station Catch Tank would eventually actuate the high-level alarm in the catch tank, which is transmitted to the Computer Automated Surveillance System (CASS) in the 2750E Building. Also, any increase in ambient radiation levels in the vicinity of the vent station would be transmitted to the CASS. The CASS operator would then notify the appropriate personnel so that corrective actions could be taken. Failure to identify the event or failure to take the proper corrective actions would result in solution overflowing to the ECSTS encasement with subsequent drainage into the Catch Tanks 241-UX-302-B and/or 241-ER-311. Catch Tank 241- UX-302-B is associated with Diversion Box 241-UX-154. Catch Tank 241-ER-311 is associated with Diversion Box 241-ER-151. Should the increase in level in these catch tanks go unnoticed, the leak would be detected by the recognition of a MBD. Material balances are performed every 2 hours. Entrop (WHC 1989) estimated that these factors would limit the overflow to the ECSTS encasement to 20,900 L (5,520 gal). As discussed in Section F.3.1.1, only 10 percent of this volume would be expected to escape the encasement and none would be expected to reach the ground surface.
F.3.1.3 Rupture of the Encasement and Pipeline
- Both the ECSTS pipelines and their concrete encasement could be ruptured by either an excavation accident or a seismic event. The leak would be maximized if the encasement ruptured immediately adjacent to an ECSTS diversion box. A recent study of similar accidents for the RCSTS estimated frequencies for leaks detected within 2 hours of 7.5 x 10-7/yr for the excavation-initiated events and 5.4 x 10-8/yr for events initiated by a BDBE (WHC 1995b). The frequency of the excavation-initiated event for the RCSTS is considered applicable to the ECSTS but that of the BDBE-initiated events is not. Entrop concluded that seismic events with horizontal ground motion exceeding 0.05 g would be sufficient to damage both the ECSTS pipelines and encasement. The design basis earthquake for the RCSTS is 0.2 g with a frequency of 1.44 x 10-4/yr. An earthquake with a horizontal ground motion of 0.09 g has an estimated frequency of 2 x 10-3/yr (WHC 1994a, WHC 1994g). On this basis, the frequency of a release from an earthquake-induced rupture of the ECSTS pipeline and encasement is estimated as Unlikely. Entrop estimated that a maximum of 79,500 L (21,000 gal) of waste could leak from the ruptured pipeline and encasement under these conditions. It is assumed that this entire volume reaches the ground surface and forms a pool for 1 hour. Respirable radioactive material is assumed to be released from this pool at the rate of 4.0 x 10-6/hr (DOE 1993). Thereafter, soil at the site of the release remains saturated, releasing respirable material at the rate of 3.6 x 10-7/hr (DOE 1993) until covered 24 hours after the initial spill began. The volume of respirable material released during an 8-hour work shift is 0.518 L (0.137 gal) and the total volume of respirable material released in 24 hours is 0.976 L (0.26 gal).
F.3.1.4 Spray Release From a Diversion Box With Cover Blocks Installed
- A recent safety analysis of hazards involving leaks, breaks, or spray releases from waste transfer systems associated with the tank farms at the Hanford Site included evaluation of spray releases applicable to the ECSTS Diversion Boxes (WHC 1994f). Four initiating events were identified for a spray release within a pit or diversion box. These events are failures of gaskets, valve packing, and piping, and a faulty jumper connection due to a maintenance error. The total frequency of occurrence of these initiating events was calculated to be 0.112 per year, with the dominant contributor being leaks from valve packing. This frequency assumes two gaskets, two valves, and 6 m (20 ft) of piping in the pit, and continuous operation of the diversion boxes. The ECSTS diversion boxes are equipped with conductivity probes that would detect the leak. Inclusion of frequency for leak detection failure (0.1/yr) (WHC 1994b) reduces the frequency of a mitigated spray release in an ECSTS diversion box to 1.1 x 10-2/yr which corresponds to an anticipated event. Stahl assumed that the spray leak occurred at both ends of the jumper and remained undetected for 24 hours (WHC 1994f). It is assumed that the spray saturates air inside the diversion box to a vapor loading of 10 mg/m3 (6 x 10-7 lbs/ft3) and that this air is forced between the cover blocks as the leak continues. The corresponding volume of respirable material released is 9.09 x 10-5 L. The volume released as a respirable aerosol during an 8-hour work shift is 3.03 x 10-5 L (8 x 10-6/gal).
F.3.1.5 Spray Release From a Diversion Box With Cover Blocks Not Installed
- The initiating events for a spray release accident with cover blocks not installed are the same as discussed in Section F.3.1.4; however, based on a frequency of 0.001/yr for failure to install the cover blocks (WHC 1994b), the frequency of occurrence drops to 1.1 x 10-5/yr which corresponds to Extremely Unlikely. Stahl assumed (WHC 1994f) a pressure of 207 psig and maximized the amount of respirable aerosol but adjusting the size of the slit through which the spray escapes. Stahl also assumed a 24-hour release which corresponds to a respirable volume of 85.5 L (22.6 gal) and 28.5 L (7.5 gal) during an 8-hour work shift. This release rate is approximately one to two orders of magnitude less that release rates for similar accidents involving NTF, RCSTS, and ITRS. The resulting spray may not be readily visible from a distance and would therefore be more difficult to detect.
F.3.2 REPLACEMENT CROSS-SITE TRANSFER SYSTEM
The RCSTS has many characteristics in common with the ECSTS, including similar
accident scenarios. Since the RCSTS is longer 10.5 km (6.2 mi), operates at a
higher pressure, 1,200 psig, and has a higher transfer rate, 530 L/min (140
gpm), than the ECSTS, the accident scenarios could result in larger releases.
However, as a newly designed and constructed system, it would be more risk-
free than the ECSTS. The RCSTS would not have the aging problems of the
existing system (corrosion, secondary concrete containment cracking), and its
design reflects the current state-of-the-art and incorporates lessons learned
over the past 40 years of operation.
New features which add to the system reliability and integrity include the
following:
. Hard-connected piping used in place of jumpers which historically have
had a number of significant failures.
. Diversion box(es) designed to eliminate the need to remove cover blocks
for routine maintenance and inspection activities. A stairwell with
Safety Class 1 doors provides access for these activities. The geometry
of the doors, stairwell, and support building entry does not provide a
direct path to the atmosphere for a spray leak. Cover blocks are
installed directly over heavy equipment (e.g., pumps) and are no larger
than required to allow replacement of the equipment to further reduce
the likelihood of a spray release directly to the atmosphere.
. Double containment provided for each transfer line by a concentric outer
pipe rather than the single concrete encasement for the ECSTS pipelines.
This provides greater leak detection sensitivity and separate leak
detection for each transfer line.
. The control and alarm systems are more sensitive and reliable.
Potential RCSTS accident frequencies and consequences were evaluated in Cross-
Site Transfer System Preliminary Safety Analysis Report (WHC 1995c).
Quantitative estimates were provided for four accidents:
. Transfer line breaks
- Beyond design basis (unmitigated)
- Design basis (mitigated)
. Spray releases
- Cover blocks not in place (unmitigated)
- Cover blocks in place (mitigated).
The frequencies of occurrence, exposure durations, and quantity of respirable
materials released are summarized in Table F-8. Descriptions of these
accidents are provided in Sections F.3.2.1 and F.3.2.2.
Summary of RCSTS Accident Releases
Exposure Respirable
Accident Frequency Duration Volume
Scenario (yr-1) (hr) (L)
Transfer Pipe 2.2 x 10-6 to 5.4 x 10-8 On-site 1 1.17
Break Extremely Unlikely to Not 7 0.736
(Unmitigated) Reasonably Foreseeable 1.91
Off-site 1 1.17
23 2.42
3.59
Transfer Pipe 1.5 x 10-5 to 1.1 x 10-6 On-site 1 0.404
Break Extremely Unlikely 7 0.255
(Mitigated) 0.659
Off-site 1 0.404
23 0.836
1.24
Unmitigated 1.9 x 10-9 On-site 8 6,700
Spray Release Not Reasonably
Foreseeable
Off-site 8 6,700
Mitigated 0.11 to 0.03 On-site 8 0.0075
Spray Release Anticipated
Off-site 8 0.0075
Source: (WHC 1995c)
F.3.2.1 Transfer Line Breaks
- The RCSTS transfer line consists of an 8-cm
(3-in) diameter, Schedule 40 stainless steel pipe within a 15-cm (6-in)
diameter carbon steel pipe. The primary piping is of all-welded construction.
An electronic system in the annular space is capable of detecting and
identifying the approximate location of any leak. The secondary containment
pipe is a closed system and can be pressure tested to confirm its integrity.
Such testing would be performed at least once per year (WHC 1995c). All of
these new design features add to the integrity of the RCSTS and would result
in lower probability of accident and/or failure events.
Lindberg estimated the frequencies of transfer line breaks initiated by
excavation accidents and by a BDBE (WHC 1995b). It was determined that common
backhoes used for construction were not capable of rupturing both the primary
and secondary pipes; however, an oversized backhoe could cause a double break.
The excavation scenario considered the following events and frequencies:
. Excavation occurs (1.0)
. Waste transfer in progress (0.3/yr)
. Oversized backhoe is available and used (0.01)
. Administrative control fails - wrong area excavated (0.1)
. Location fails - "terra tape" and marker posts not seen (0.1)
. Rupture occurs (0.5)
. Warnings fail - material balance, flow, pressure, level without leak
detection (0.05).
The result of this scenario would be a leak lasting more than 2 hours with a
frequency of 7.5 x 10-7/yr which corresponds to Incredible. If warnings are
recognized and responded to, the frequency becomes 1.5 x 10-5/yr (Extremely
Unlikely) but the leak is limited to not more than 2 hours.
The earthquake scenario considered (WHC 1995c) the following events and
frequencies:
. BDBE occurs (1.44 x 10-4/yr)
. Waste transfer in progress (0.3/yr)
. Administrative control fails - manual shutdown not performed (0.05)
. Leak detection fails (0.5)
. No response to leak detection occurs (0.001)
. Warnings fail - material balance, flow, pressure, level (does not
include leak detection) (0.05)
Two event tree branches lead to spills of more than 2 hours. The more likely
of the two assumes that leak detection and material balance and other warnings
fail and has a frequency of occurrence of 5.4 x 10-8/yr. This sequence is
considered not reasonably foreseeable. The frequency of a leak of not more
than 2 hours is 1.1 x 10-6/yr which corresponds to Extremely Unlikely.
It could be argued that failures of engineered systems are not independent for
a BDBE so that the frequency of leak detection failure would be 1.0. It could
also be argued that it is inappropriate to consider both failure to execute a
manual shutdown and failure to perform material balance under these
conditions. This would increase the frequency of an 8-hour release to 2.2 x
10-6/yr.
The total volume of waste spilled during an unmitigated transfer pipe break is
292,000 L (77,100 gal). This includes 254,000 L (67,200 gal) pumped in the 8
hours before the leak is detected and 37,500 L (9,900 gal) that drains from
the filled line through the break. A 2-hour leak would spill 101,000 L
(26,700 gal). It is conservatively assumed that all of the waste spilled
reaches the ground surface.
The following paragraphs describe unmitigated and mitigated RCSTS pipe break
accidents.
. Unmitigated Transfer Pipeline Break - An unmitigated transfer pipeline
break is assigned a frequency of occurrence ranging from Extremely
Unlikely to Not Reasonably Foreseeable. This frequency range reflects a
usage factor of 0.3 and the assumptions that a BDBE would cause
simultaneous failure of all engineered systems and that material balance
would not be performed. It is assumed that 292,000 L (77,000 gal) of
waste are spilled in the 8 hours before the leak is detected. The spill
forms a pool on the ground surface for 1 hour and then keeps the soil
saturated until the spill area is covered 16 hours after detection.
Respirable material is released from the pool at a rate of 4.0 x 10-6/hr
and from the saturated ground at a rate of 3.6 x 10-7/hr. The total
amount of respirable material released in 24 hours is 3.59 L (0.9 gal).
A total of 1.91 L (0.5 gal) would be released during an 8-hour work
shift.
. Mitigated Transfer Pipeline Break - The mitigated transfer pipe break
assumes that the initiating event is within the design basis and is
considered to be Extremely Unlikely. The scenario assumes that the leak
is detected within 2 hour and would spill 101,000 L (27,000 gal) of
waste to the ground surface. Respirable material is released from the
pool at a rate of 4.0 x 10-6/hr and from the saturated ground at a rate
of 3.6 x 10-7/hr. The total amount of respirable material released in
24 hours is 1.24 L (0.33 gal). A total of 0.659 L (0.174 gal) would be
released during an 8-hour work shift.
F.3.2.2 Spray Releases
- Spray releases and accidents are less likely in the
RCSTS diversion box(es) than in existing process pits since the diversion
boxes use welded stainless steel piping instead of jumpered steel piping.
Also, the system design makes it less likely that aerosol will be sprayed
directly to the atmosphere. However, since the RCSTS would operate at much
higher pressures than existing transfer piping, potential releases are
correspondingly larger.
Information on unmitigated and mitigated RCSTS spray releases and accidents
from Cross-Site Transfer System Preliminary Safety Analysis Report (WHC 1995c)
is discussed in the following paragraphs.
. Unmitigated Spray Release - The unmitigated spray release accident
assumes that a valve in a RCSTS diversion box fails creating a 3 cm
(1.2 in) by 0.15 mm (0.006 in) crack. The transfer system pressure is
assumed to be 2,000 psig. It is also assumed that there are no
additional barriers to prevent release, that is, the diversion box is
open to the sky. Respirable material is assumed to be released at the
rate of 0.231 L/s (0.06 gal/s) which, in an 8-hour release produces
6,700 L (1,800 gal) of respirable material (WHC 1995c). A release of
this magnitude would be expected to have catastrophic consequences. No
frequency was provided for the scenario.
To assign an accident frequency estimate to the RCSTS unmitigated spray
release, an event sequence was developed which reflects the unique
design features of the RCSTS diversion boxes. These features provide
access for all routine maintenance and inspection functions through an
access corridor with a 90-degree turn and doors at each end. The lower
door opens into a shield baffle within the diversion box and is entirely
below grade. This geometry eliminates the possibility of the release of
a spray directly to the atmosphere via this route. A route would exist
if it were necessary to replace heavy equipment such as a pump. A
relatively small removable shield block located in the roof of the
diversion box directly over the pump would need to be removed to allow
use of a crane to remove and replace the defective component. If the
block were not replaced before operations resumed, a direct path to the
atmosphere would exist and an unmitigated spray release could occur
should a valve or weld subsequently fail.
The frequency that the shield block would need to be removed can be
estimated from failure data for large pumps in a chemical environment
(WHC 1995d). The overall failure rate is estimated to be 6.0 x 10-5/hr
with 8.5 percent of these failures being catastrophic. Thus the
frequency of a failure requiring replacement is 5.1 x 10-6/hr. It is
assumed that there is a 0.1 percent chance that the blocks would not be
replaced before resuming operations. The failure rate of a valve in an
RCSTS diversion box (WHC 1995c) is 3.5 x 10-6/hr and there are
approximately 12 valves in a diversion box. Assuming 24-hour operation,
the frequency of an RCSTS unmitigated spray release is 1.9 x 10-9/yr and
not reasonably foreseeable.
. Mitigated Spray Release - The mitigated spray release scenario differs
from the unmitigated case in that it assumes that the diversion box is
closed. This requires that doors in the entry stairwell be closed, that
the small cover block(s) be in place, and that a small penetration to
allow connection of a portable exhauster be closed.
It is assumed that the confined spray saturates the air inside the
diversion box with vapor at a loading of 100 mg/m3 (6 x 10-6 lbs/ft3).
This saturated air is assumed to be forced out of the diversion box by
the combination of the rise in humidity due to evaporation and the
accumulation of bulk liquid in the box. The estimated total release is
0.0075 L (.002 gal) in the 8 hours assumed to elapse before the leak is
detected. Approximately 40 percent of this release occurs in the first
20 seconds. The frequency of this event for a RCSTS consisting of four
diversion boxes and operating 643 hr/yr is 0.11/yr (WHC 1995c). Current
plans are to construct only one diversion box which reduce the frequency
to 0.03/yr. In either case, the mitigated spray scenario is an
anticipated event.
This scenario does not consider the presence of leak and radiation
detectors which could limit the duration of the release; however, based
on the modeling of evaporative release, they reduce the release only by
about one-half. On the non-conservative side, drawings (e.g., H-2-
822231), (WHC 1995c) show a HEPA filter that appears to draw from the
entry stairwell and exhaust into the diversion box. The flow rate of
this filter could not be found but could be considerably higher than the
0.187 m3/min (6.6 ft3/min) assumed for the vapor release rate.
F.4 ABOVE-GROUND CROSS-SITE TRANSFER
This evaluation considers two options for above-ground cross-site transfer of high-level tank wastes from the 200 West Area. The first option is the use of tanker trailer trucks. Two tanker sizes are considered: 3,800 L (1,000 gal) and 19,000 L (5,000 gal). Accidents associated with the movement of loaded tanker trailer trucks are discussed in Section F.4.1. The second option for above-ground cross-site transfer of high-level tank waste from the 200 West Area is the use of 38,000 L (10,000 gal) rail tanker cars. Accidents associated with the movement of loaded rail tanker cars are discussed in Section F.4.2. Either of these above-ground transfer options would require construction and operation of new HLW load and unload facilities in the 200 East and 200 West Areas to process the HLW in the tankers. Accidents at these facilities are discussed in Section F.4.3.
F.4.1 TANKER TRAILER TRUCKS
This section discusses accidents that could occur while using tanker trailer
trucks to transport of HLW between the 200 East and West Areas. Two types of
tanker trailer trucks are considered. The LR-56(H) Cask System is licensed
and operated in France as a Type B quantity package for radioactive liquids
and has a capacity of approximately 3,800 L (1,000 gal) (WHC 1994h). The
second tanker trailer truck has a capacity of 19,000 L (5,000 gal). No tanker
of this size currently exists for the transport of HLW; however, a shielded
19,000-L (5,000-gal) tank truck has been used to transport high activity LLW
at the Savannah River Site. It is assumed that neither the LR-56(H) nor the
19,000-L (5,000-gal) tanker trailer truck would be licensed as Type B packages
for use in the United States.
The movement of vehicles carrying radioactive materials is more strictly
controlled on the Hanford Site than on public roadways. This control
includes:
. Speed restrictions - Vehicles transporting HLW are limited to 40 kmph
(25 mph) under good conditions and 16 kmph (10 mph) under icy, snowy,
and reduced visibility conditions.
. Escorts required - Escorts accompany the vehicle to monitor road
conditions, control local traffic, and control proximity of other
vehicle to the transport vehicle. When appropriate, roads would be
closed.
This evaluation does not specifically address whether either of these tanker
trailer trucks meet equivalent safety requirements at the Hanford Site. Such
a determination would be made based on a SARP. A SARP has not been prepared
for either the LR-56(H) or 19,000-L (5,000-gal) package. Packaging design
criteria have been developed for the slightly modified LR-56(H) planned for
use at Hanford (WHC 1994h).
The following initiating events are postulated for potential releases during
truck transport of HLW on the Hanford Site:
. In-transit punctures
. Fire-induced breaches
. Collisions and rollovers
. Criticality.
The frequencies of these initiating events, the likelihood that they would
result in the release of package contents, and the quantity of material
released are discussed in the following sections.
F.4.1.1 In-Transit Punctures
- Objects thrown up by passing vehicles, propelled by wind or near-by explosions, or deliberately directed projectiles could strike the tanker trailer truck during transit. Both International (IAEA Safety Series No. 6, Regulations for the Safe Transport of Radioactive Material, 1985 Edition) and DOE regulations (DOE/RL Order 5480.3, Safety Requirements for the Packaging and Transportation of Hazardous Materials, Hazardous Substances, and Hazardous Wastes) contain the same requirement regarding puncture of Type B packages. The package must withstand without failure a free drop through a distance of 1 m (40 in) onto the top end of a vertical cylindrical mild steel bar mounted on an essentially unyielding horizontal surface and striking in a position for which maximum damage is expected. Although the French LR-56(H) met this requirement, a separate demonstration would be required for Type B certification in the United States. The ability of the larger and yet undesigned 19,000 L (5,000 gal) tanker to meet this requirement is unknown. Certification is not required if the tankers are not transported by public conveyance and if equivalent safety can be demonstrated (DOE 1985). The accident scenarios discussed in Section F.4.1.3 bound any reasonably foreseeable puncture scenarios.
F.4.1.2 Fire Induced Breaching
- A fire of sufficient energy and duration to cause the breaching of an LR-56(H) Cask System or a 19,000-L (5,000-gal) HLW package in the absence of the collisions and rollovers discussed in Section F.4.1.3 is considered to be extremely unlikely. Based on statistics for trucks over 4,000 kg (8,500 lb) operated on the Hanford Site, the frequency of a random fire (one not associated with a collision or rollover) is 6.5 x 10-10/km (1.04 x 10-9/mi) (WHC 1993c). The contribution of random truck fires is considered in Section F.4.1.3.
F.4.1.3 Collisions and Rollovers
- Collisions and rollovers could lead to release of the contents of the LR-56(H) or 19,000-L (5,000-gal) tanker trailer truck, particularly if a fire also occurred. A previous analysis concluded that only accidents involving uncontrolled fires and rollovers would result in the release of radioactive materials from packages transported by truck. The frequencies of these types of accidents have been estimated from Hanford Site accident report data for trucks with gross vehicle weights over 4,000 kg (8,500 lb) (Wilson 1992, WHC 1993c) and are summarized in Table F-9. The total frequency of events that could result in release of radioactive materials is 2.7 x 10-9/km (4.3 x 10-9/mi). The frequency of truck accidents in the State of Washington is 2.62 x 10-7/km (4.2 x 10-7/mi). Based on data presented in NUREG/CR-4829, Shipping Container Response to Severe Highway and Railway Accident Conditions, a similar analysis of transportation of Hanford radioactive materials concluded that only 0.6 percent of these accidents would result in a release (WHC 1994i). This corresponds to a release frequency of 2.5 x 10-9/mi.
Accident Frequencies for Trucks at the Hanford Site
Accident Frequency Initiator Result (/mile) (/kilometer) Random Fire Uncontrolled Fire 1.04 x 10-9 6.46 x 10-10 Collision Uncontrolled Fire 6.27 x 10-11 3.90 x 10-11 Rollover, Good Road Uncontrolled Fire 2.38 x 10-11 1.48 x 10-11 Rollover, Harsh Road Breach, No Fire 3.10 x 10-9 1.93 x 10-9 Rollover, Harsh Road Breach, Fire 3.77 x 10-11 2.34 x 10-11 Total 4.26 x 10-9 2.65 x 10-9 Source: Wilson 1992 Both sets of frequency data include radioactive material packages ranging from strong tight containers to Type B fissile containers. In the period 1971 through 1982, 45 Type B packages survived transportation accidents with no package failure and no release of radioactive materials (SNL 1984). Although the trucks considered here would transport Type B quantities, the transport packages cannot be assumed to meet Type B package requirements. To estimate the frequency of accidents for the two tanker trailer truck options, a total frequency of an accident resulting in release of any radioactive material is assumed to be 4.3 x 10-9/mi, the more conservative of the two estimates. The distance between the SY and A Tank Farms, the assumed locations of the HLW load and unload facilities, is 10.7 km (6.5 mi) (WHC 1995e). Based on this distance, the frequency of an accident releasing any radioactive material would be 2.8 x 10-8 per trip. The quantity of material released is assumed to be a function of the severity of the accident. Fractional release frequencies of rail accidents (WHC 1993d) are shown in Table F-10 and are considered to be applicable to truck accidents. Of those accidents resulting in any release, 27.9 percent release less than 1 percent of the contents, 24.5 percent release 90 to 100 percent of the contents. No other release fraction interval accounts for more that 4 percent of the total release frequency. The 100 percent release accident is therefore used as the basis of evaluated tanker trailer truck accidents. The frequency of this accident is 6.8 x 10-9 per trip.
Fractional Release Frequencies for Rail Accidents
Percent Released Frequency/mile 0.0001 - 0.99 6.70 x 10-9 1.00 - 9.999 2.39 x 10-10 10 - 19.999 6.45 x 10-10 20 - 29.999 7.09 x 10-10 30 - 39.999 7.09 x 10-10 40 - 49.999 6.45 x 10-10 50 - 59.999 8.38 x 10-10 60 - 69.999 5.80 x 10-10 70 - 79.999 2.58 x 10-10 80 - 89.999 7.74 x 10-10 90 - 100 5.89 x 10-9 Total 2.4 x 10-8 Source: WHC 1993d, Appendix 6.4. Table F-11 shows the total probability and quantity of respirable material released during accidents in which 100 percent of the tanker contents is lost for each of the types of high-level tank waste considered in this EIS. The amount of respirable material released is based on the assumption that the entire contents are spilled on the ground and that the area remains uncovered for 24 hours. During the first hour, the spill forms a pool that releases material at the rate of 4.0 x 10-6/hr. The spill then sinks into the ground but keeps the area wet. During this phase, respirable radioactive material is released at the rate of 3.6 x 10-7/hr. Workers are assumed to be exposed for one shift (8 hour). The public is exposed until the spill area is covered.
F.4.1.4 Criticality
- The LR-56(H) Cask System and 19,000-L (5,000-gal) tanker trailer truck, and the contents allowed in them, would be required to meet the subcriticality requirements of DOE/RL Order 5480.3 for hypothetical accident conditions. Analyses are required to demonstrate that criticality is not a credible event. Criticality analyses routinely use extremely
Summary of Maximum Accident Releases from Transport Vehicles
During Cross-Site Transfers
Frequency of Exposure Respirable
Accident Complete Loss of Duration Volume Released
Scenario Contents (hr) (L)
LR-56(H) Tank Breach 6.8 x 10-9/trip On-site 1 0.0151
SWL 2.9 x 10-5 7 0.0095
102-SY Supernatant 4.7 x 10-6 0.0246
101-SY (1:1 Dilution) 1.5 x 10-5
4.9 x 10-5
Off-site 1 0.0151
23 0.0313
0.0464
5,000 Gal Tank Truck Breach 6.8 x 10-9/trip On-site 1 0.0756
SWL 5.7 x 10-6 7 0.0476
102-SY Supernatant 9.4 x 10-7 0.123
101-SY (1:1 Dilution) 3.0 x 10-6
9.6 x 10-6
Off-site 1 0.0756
23 0.156
0.232
Rail Tank Breach 5.7 x 10-8/trip On-site 1 0.152
SWL 2.4 x 10-5 7 0.0955
102-SY Supernatant 3.9 x 10-6 0.248
101-SY (1:1 Dilution) 1.3 x 10-5
4.1 x 10-5
Off-site 1 0.152
23 0.314
0.466
conservative bounding conditions (such as ideal moderation, perfect reflection
of neutrons at the boundaries, and optimum geometry. Mechanisms that could
segregate fissionable isotopes from the rest of the liquid waste (e.g.,
settling, chemical or thermally induced precipitation out of solution,
unintended filtering effects) would be evaluated. Administrative controls
would be established to limit the quantities of fissionable materials.
Sampling would be performed as necessary to ensure that administrative
controls are being met.
Hanford Site high-level tank wastes have been found to be highly subcritical
because of the relatively high concentrations of neutron absorbers to Pu in
the wastes (WHC 1994c). Since it is very likely that the high-level tank
wastes that would be transported by tanker trailer truck would be highly
subcritical, no criticality accident scenarios have been developed.
F.4.2 RAIL TANK CARS
This section discusses accidents that could occur while using a 38,000-L
(10,000-gal) rail tanker car to transport high-level tank waste from the 200
West to the 200 East Area. The tanker car has not been designed for HLW at
Hanford but is assumed to be a heavily-shielded, double-wall steel tank. A
shielded 19,000-L (5,000-gal) rail tank car is used to transport high activity
LLW at the Savannah River Site.
This evaluation does not address whether this type of HLW transport package
can be designed or will meet Hanford equivalent safety requirements. The
latter determination would be made based on a SARP.
The movement of rail cars carrying radioactive materials at the Hanford Site
is strictly controlled (WHC 1993d). Controls include:
. Train speed is limited to 6 kmph (4 mph) while coupling.
. Speed is limited to maximum of 40 kmph (25 mph). Speed is limited to
16 kmph (10 mph) on paved road crossings, during icy or snowy
conditions, and when visibility is limited. Speed is limited to 8 kmph
(5 mph) on rail spurs.
. Patrol blockages are established to stop traffic at paved road
crossings.
. At least one spacer car is required on either side of the HLW tank car.
. At least one person in the locomotive cab must keep the cars under
observation at all times.
These controls would reduce the likelihood and consequences of rail car
accidents.
The frequencies of these initiating events, the likelihood that they would
result in the release of container contents, and the quantity of material
released are discussed in the following sections.
The following initiating events are postulated for potential releases during
rail car transport of SWLs or 200 West Area facility wastes on the Hanford
Site:
. In-transit punctures
. Fire-induced breaches
. Collisions and Derailments
. Criticality.
F.4.2.1 In-Transit Punctures
- In-transit punctures could conceivably occur if a piece of broken rail were thrown up by the car wheels. Due to the limited train speeds, the energy of such an object would be low. Track maintenance and routine and pre-transfer track inspections would keep the tracks clear of spikes and rail debris and, in general, greatly minimize the likelihood of a puncture event. A misaligned coupler with another rail car could also conceivably cause a puncture, but with minimum coupling operations and with the use of automatic hood-and-shelf type couplers such an event would essentially be precluded.
F.4.2.2 Fire-Induced Breaches
- A fire of sufficient energy and duration to cause the breaching of a rail tank car designed to transport HLW is considered to be extremely unlikely if the fire is not related to derailment or a collision. Track crews currently spray weed killer at least once every 2 years in a 3 m (10 ft) swath on either side of the centerline of the tracks, thus decreasing the possibility of a vegetation fire next to the tracks. Loose tumbleweeds are removed as necessary. To minimize the potential impact from the diesel fuel of the locomotive catching on fire at least one railroad spacing car would be placed between the engine and the HLW rail tank car.F.4.2.3 Collisions and Derailments
- Collisions or derailments could cause loss of the integrity of the HLW rail tanker car. In view of speed limitations and requirements for blockades at road crossings only, collisions involving derailments or fires are considered to be credible initiating events. Derailment is considered to be the most likely of the potential initiating events (WHC 1993d). The SARP for the rail tank cars used to transport LLW at the Hanford Site includes frequencies and release fractions. These data are shown in Table F-11 and are based on national rail statistics with adjustments to reflect the more controlled and safer transport conditions within the Hanford Site. The total frequency of a rail accident that would result in any release of radioactive materials is 1.5 x 10-8/km (2.4 x 10-8/mi). The entire contents are lost with a frequency of 3.7 x 10-9/km (5.9 x 10-9/mi). The distance by rail between the A and SY Tank Farms, the assumed locations of the HLW load and unload facilities, is 15.5 km (9.7 mi). The frequency of any accidental release 2.3 x 10-7/trip and 5.7 x 10-8/trip for a total loss of contents. Table F-11 shows the probability of an accident involving complete loss of the rail tanker car contents for each type of high-level tank waste considered in this EIS and the amount of respirable material that would be released. The amount of respirable material released is based on the assumption that the entire contents are spilled on the ground and that the area remains uncovered for 24 hours. During the first hour, the spill forms a pool that releases material at the rate of 4.0 x 10-6/hr. The spill then sinks into the ground but keeps the area wet. During this phase, respirable radioactive material is released at the rate of 3.6 x 10-7/hr. Workers are assumed to be exposed for one shift (8 hours). The public is exposed until the spill area is covered.F.4.2.4 Criticality
- The HLW rail tanker car and its contents would be required to meet the subcriticality requirements or DOE/RL Order 5480.3 for hypothetical accident conditions. Analyses are required to demonstrate that criticality is not a credible event. Criticality analyses routinely use extremely conservative bounding conditions (such as ideal moderation, perfect reflection of neutrons at the boundaries and optimum geometry). Mechanisms that could segregate fissionable isotopes from the rest of the liquid waste (e.g., settling, chemical or thermally induced precipitation out of solution, unintended filtering effects) would be evaluated. Administrative controls would be established to limit the quantities of fissionable materials. Sampling would be performed as necessary to ensure that administrative controls are being met. Hanford Site high-level tank wastes have been found to be highly subcritical because of the relatively high concentrations of neutron absorbers to Pu in the wastes (WHC 1994c). Since it is very likely that the high-level tank waste transported in the HLW rail tank car would be highly subcritical, no criticality accident scenarios are developed.
F.4.3 LOAD AND UNLOAD FACILITIES
High-level waste load and unload facilities would be designed to protect both
workers and the general public from the hazards associated with the loading
and unloading of the tanker trailer trucks and rail tank cars. Design
criteria for a load and unload facility would comply with the following DOE
orders:
. DOE Order 6430.1A, General Design Criteria
. DOE Order 5820.2A, Radioactive Waste Management
. DOE Order 5480.28, Natural Phenomena Hazards Mitigation.
Compliance with these DOE orders assures that adequate accident related
preventive and mitigative measures would be considered during the design of
the load and unload facilities. Furthermore, such a facility would not be
allowed to begin operations until a Final SAR, prepared in accordance with the
extensive requirements of DOE Order 5480.23, Nuclear Safety Analysis Reports,
has been approved by the appropriate authorities.
The load and unload facilities would accommodate rail tanker cars, tanker
trailer trucks, or both depending on the option selected. The facility would
be similar in concept to the existing 204-AR facility for unloading of low-
level waste but would differ considerably in design and operation to
accommodate handling of HLW liquids. Some of the features that would
distinguish the HLW load and unload facilities from the 204-AR facility
include (WHC 1993c):
. Placement of pumps and valves in shielded cells equipped with remotely
operated controls
. Drive-through bays to eliminate backing of transport vehicles
. Remotely-operated fill/drain lines and vent lines
. Heavily-shielded bay access doors
. Differential pressure zones in the HVAC system to isolate areas based on
expected level of contamination with dual stage HEPA filtration in
contaminated zones.
Three categories of initiating events are often examined in the performance of
an accident safety analysis: operational events, natural phenomena, and
external events. Each category of initiating event is capable of resulting in
different types of accidents (e.g., fires and spills). Accident scenarios
induced by seismic events and high winds or tornadoes have often received the
most attention in the performance of safety analyses. External events include
the crashing of aircraft and energetic (i.e., accident) events at facilities
located near the facility being evaluated. An examination of transportation
routes and facilities around likely potential sites for an HLW load and unload
facility in the 200 Areas revealed no sources that could provide the necessary
energy to cause an accident scenario more severe than those postulated in the
following sections. Though not methodically evaluated, an accident at an HLW
load and unload facility induced by nearby transportation routes and/or
facilities is considered to be extremely unlikely.
Accident analyses specific to the HLW load and unload facilities are not yet
available. For purposes of this evaluation, scenarios for the following
accident categories are considered:
. Spills
. Spray releases
. Fires.
The frequency of occurrence and quantities of respirable material released
during these accident events are summarized in Table F-12. These accident
categories are further discussed in Sections F.4.3.1, F.4.3.2, and F.4.3.3,
respectively. Other accident scenarios could be postulated, but their
consequences would likely be relatively small and limited to local
contamination and to in-facility workers. Examples of such accidents include
Summary of Accident Releases for the HLW Load and Unload Facilities
Exposure Respirable
Accident Frequency/100 Duration Volume
Scenario Tankers (hr) (L)
Loading/Unloading Anticipated to On-site 0.5 0.00568
Spill Unlikely
Off-site 0.5 0.00568
LR-56(H) Earthquake 3.2 x 10-5 On-site 8 2.53
Breach Extremely
Unlikely
Off-site 12 3.79
5,000 Gal Tank Truck 3.2 x 10-5 On-site 8 12.6
Earthquake Breach Extremely
Unlikely
Off-site 12 18.9
Rail Car Tank 3.2 x 10-5 On-site 8 25.2
Earthquake Breach Extremely
Unlikely
Off-site 12 37.9
Source: WHC 1991
overflow of a catch tank located in a pit, containment breach during changeout
of a HEPA filter, and loss of off-site power (LOOP). Regarding LOOP, a
standby generator would be provided to power key loads in the event of a LOOP.
If the standby generator also failed, then an uninterruptible power supply
would provide power to essential items such as emergency lights and fire
monitoring and alarm components. LOOP and failure of the standby generator
would result in loss of the differential pressure zones previously discussed.
F.4.3.1 Spills
- Various hardware failures or human errors could lead to a spill within the HLW load and unload facilities. Leaks could occur in pipes, valves, pumps, and connectors due to mechanical failure or human error. Holding tanks, catch tanks, or the transport tanker could fail or be overfilled. A BDBE could also cause spills, either by damaging facility components or by rupture of a loaded transport vehicle. Small leaks up to 38 L (10 gal) are categorized (WHC 1994j) as anticipated events, and larger spills up to 1,900 L (500 gal) are categorized as unlikely. A DBE (0.25 g) has a frequency of 4 x 10-4 to 1 x 10-3/yr (WHC 1991). Frequencies and released quantities of waste are shown in Table F-12. The safety assessment for the 204-AR facility considered a spill during filling of a tank car (WHC 1991). It was assumed that a gasket failed during filling at 757 L/min (200 gpm) and 10 percent of the flow spilled to the floor over a 30-minute period. The frequency of the spill was estimated at 2.7 x 10-2/yr based on a failure rate of 3 x 10-6/hr for each of 23 gaskets in a facility requiring 4 hours to fill each of 100 rail cars. Although a quantitative estimate cannot be made without design information, it is reasonable to expect that the frequency would be less at the HLW load and unload facilities. It was assumed that 0.1 percent of the spill became airborne and that each stage of a two-stage HEPA filter removed 99.95 percent of the material. The resulting quantity of respirable material released is 0.00568 L (0.0015 gal) with a corresponding frequency of anticipated to unlikely. The monitoring and alarm systems common to facilities of this type at the Hanford Site would be expected to reduce both the frequency and consequences of this spill. An earthquake-induced spill was analyzed for the 204-AR facility (WHC 1991). The frequency of the DBE was taken as 7 x 10-4/yr and it was assumed that 100 tankers per year were processed at the rate of 4 hour/tanker. The frequency of the accident is therefore 3.2 x 10-5/yr. It was assumed that the entire contents of the transport vehicle were spilled to the ground and not covered for 12 hours. During this time, 0.001 of the spill was assumed to become airborne. The corresponding amounts of respirable liquid released from each of the different transport vehicles is shown in Table F-12. The frequencies assume 100 tankers/year, 4 hours per tanker regardless of vehicle capacity.
F.4.3.2 Spray Releases
- The same mechanisms that cause leakage can also cause spray releases (e.g., failure of pipes, valves, pumps, and connections between them). A spray leak is an anticipated event. The frequency and magnitude of any resultant release is difficult to assess without design information. The magnitude of the release would probably be less than those for mitigated spray leaks postulated for other proposed facilities such as the RCSTS and NTF. The frequencies could be somewhat higher since new connections must be made each time a vehicle is loaded or unloaded. Spray releases at the HLW load and unload facilities are considered to be bounded by the spill scenario discussed in Section F.4.3.1 due to the large fractional release assumed for respirable material.
F.4.3.3 Fires
- Mechanisms for a fire at an HLW load and unload facility are extremely limited. There would be very little combustible material in the load/unload bay, and throughout the facility in general. The likelihood of electrical fires would be minimized by the use of industry accepted electrical system design standards, which include the use of overcurrent and short circuit protection devices. Further electrical fires, with no significant fuel supply, are not significant. Lightning protection equipment would also be installed in accordance with applicable industry standards. An aircraft crash at an HLW load and unload facility could lead to a fire scenario. However, the total frequency of aircraft accidents (due to all types of aircraft) at the postulated NTF in the 200 Areas is less than 1 x 10-7/yr (WHC 1994b). Accordingly, accidents in the HLW load and unload facilities due to fires are considered to be bounded by the transportation accidents considered in Section F.3.
APPENDIX F REFERENCES
ARH, 1977, Exothermic Potential of Sodium Nitrate Saltcake, Beitel, G.A., ARH- LD-163, Atlantic Richfield Hanford Company, Richland, WA ARH, 1976, Chemical Stability of Salt Cake in the Presence of Organic Materials, Beitel, G.A., ARH-LD-119, Atlantic Richfield Hanford Company, Richland, WA Cloud, 1993, Position Paper, Project W-236A Multifunction Waste Tank Facility Waste Storage Tank Heat Removal Systems, Kaiser Engineers Hanford, Richland, WA DOE, 1995, Environmental Assessment Tank 241-C-106 Past-Practice Sluicing Waste Retrieval, Hanford Site, Richland, WA, DOE/EA-0944, February 1995, United States Department of Energy, Richland, WA DOE, 1994, Environmental Assessment, Waste Tank Safety Program, DOE/EA-0915, U.S. Department of Energy, Hanford Site, Richland, WA DOE, 1993, Recommended Values and Technical Bases for Airborne Release Fractions (ARFs), Airborne Release Rates (ARRs), and Respirable Fractions (RFs) at DOE Non-Reactor Nuclear Facilities, DOE Handbook, DOE-HDBK-0013-93, July 1993, U.S. Department of Energy, Washington D.C. DOE, 1992a, Environmental Assessment for Proposed Pump Mixing Operations to Mitigate Episodic Gas Releases in Tank 241-SY-101, Hanford Site, Richland, Washington, DOE/EA-0803, U.S. Department of Energy, Washington, D.C. DOE, 1992b, An Environmental Assessment for Proposed Pump Mixing Operations to Mitigate Episodic Gas Releases in Tank 241-101-SY. Hanford Site, Richland, Washington, DOE/EA/0803, U.S., Department of Energy, Washington, D.C., August 1992 DOE, 1987, Final Environmental Impact Statement, Disposal of Hanford Defense High-Level, Transuranic and Tank Wastes, Hanford Site, Richland Washington, DOE/EIS-0113, U.S. Department of Energy, Washington, D.C. DOE, 1985, Safety Requirements for the Packaging and Transportation of Hazardous Materials, Hazardous Substances, and Hazardous Wastes, DOE/RL Order 5480.3, July 9, 1985 KEHC, 1993, Project W-030 Process Flow Diagrams, Stack Elevation and Civil Plot Plan, Kaiser Engineers Hanford Company for the U.S. Department of Energy, Richland Field Office, Richland, WA LANL, 1994, Los Alamos National Laboratory, A Safety Assessment for Proposed Pump Mixing Operational to Mitigate Episodic Gas Release in Tank 241-SY-101, LA-UR-92, Rev. 8 (DRAFT) PNL, 1994a, Mitigation of Tank 241-SY-101 by Pump Mixing: Results of Full- Scale Testing PNL-9959/UC-510, Pacific Northwest Laboratory PNL, 1994b, Mitigation of Tank 241-SY-101 by Pump Mixing: Results of Testing Phases A and B, PNL-9423, UC-510, Pacific Northwest Laboratory, Richland, WA SNL, 1984, The Transportation of Nuclear Materials, SAND84-0062, TTC-0471, Unlimited Release, UC-71, T.A. Wolff, December, Sandia National Laboratories, Albuquerque, NM WHC, 1995a, Safety Assessment, Initial Tank Retrieval Systems, Project W-211, WHC-SD-W211-SAD-001, Rev. A, June 1, 1995 by R.J. Kidder WHC, 1995b, Human Reliability Analysis of Backhoe and Seismi Events Involving The Cross-Site Transfer Line (Internal Memo, OM640-sel-95005), Westinghouse Hanford Company, Richland, WA WHC, 1995c, Cross Site Transfer System Preliminary Safety Analysis Report, WHC-SD-W058-PSAR-001 Rev. 1, (DRAFT) Westinghouse Hanford Company, Richland, WA WHC, 1995d, Multi-Function Waste Tank Facility Path Forward Engineering Analysis Technical Task 3.6, Estimate of Operations Risk in the 200 West Area, WHC-SD-W236A-ES-014, Revision OA, G.A. Coles, Westinghouse Hanford Company, Richland, WA WHC, 1995e, Replacement of the Cross Site Transfer System Liquid Waste Transport Alternatives Evaluation, Project W-058, WHC-SD-W058-TA-001, Revision O., D.V.Vo, E.M. Epperson, Westinghouse Hanford Company, Richland, WA WHC, 1994a, Leach, D. S., 1994, Multi-Function Waste Tank Facility - Preliminary Safety Analysis Report, WHC-SD-W236A-PSAR-001, Rev. A, Vol. II, SEAC Review Issue Draft, Westinghouse Hanford Company, Richland, WA WHC, 1994b, Supplemental PSAR Analysis, WHC-SD-W236A-ANAL-002, August 1994, Additional Analysis Related to the MWTF WHC, 1994c, High-level Waste Tank Subcritically Safety Assessment, Braun, D.J., et al., WHC-SD-WMSARR-003, Rev. 0, Westinghouse Hanford Company, Richland, WA WHC, 1994d, Ferrocyanide Safety Program: Safety Criteria for Ferrocyanide Watch List Tanks, WHC-EP-0691, UC-600, Westinghouse Hanford Company, Richland, WA WHC, 1994e, Preliminary Safety Evaluation for 241-C-106 Waste Retrieval, Project W-320, J.C.Conner, WHC-SD-WM-PSE-010, Rev. 2. October 1994, Westinghouse Hanford Company, Richland, WA WHC, 1994f, Safety Analysis; Tank Farms Waste Transfer System Leaks, Breaks, and Spray Releases, WHC-SD-WM-SARR-005, Rev 0, July 28, 1994, Westinghouse Hanford Company, Richland, WA WHC, 1994g, Leach, D. S., 1994, Multi-Function Waste Tank Facility - Preliminary Safety Analysis Report, WHC-SD-W236A-PSAR-001, Rev. A, Vol. I & II, SEAC Review Issue Draft, Westinghouse Hanford Company, Richland, WA WHC, 1994h, Packaging Design Criteria for the LR-56(H) Cask System, WHC-SD-TP- PDC-021, Rev 0, EDT 142720, March 15, 1994, Westinghouse Hanford Company, Richland, WA WHC, 1994i, Transportation Impact Analysis for the Tank Waste Remediation System Environmental Impact Statement Proposed Alternatives, WHC-SD-TP-RPT- 0161, Draft, Rev A, Westinghouse Hanford Company, Richland, WA WHC, 1994j, Hazard and Accident Analysis (Interim Chapter 3), WHC-SD-WM-SAR- 065, Rev 0, July 1994, Westinghouse Hanford Company, Richland, WA WHC, 1993a, Radionuclide and Chemical Inventories for the Double Shell Tanks, Oscarson, E. E. and Tusler, L.A., WHC-SD-WM-TI-543, July 1993 WHC, 1993b, Interim Safety Basis Topical Report Flammable Gas Waste Tanks, Section 5.3 of WHC-SD-WM-ISB-001, Vol. 1, Rev. 0, Westinghouse Hanford Company, Richland, WA WHC, 1993c, Pilot Plant Hot Test Facility Siting Study, WHC-SD-WM-TA-143, Draft, Rev 0, Westinghouse Hanford Company, Richland, WA WHC, 1993d, SAR for Packaging Railroad Liquid Waste Tank Car, WHC-SD-RE-SAP- 013, Rev 5, Westinghouse Hanford Company, Richland, WA WHC, 1992, Safety Study of Interim Stabilization of Non-Watchlist Single Shell Tanks, SD-WM-RPT-048, Revision 0, Westinghouse Hanford Company, Richland, WA WHC, 1991, SAR for the 204-AR Waste Unloading Facility, WHC-SD-WM-SAR-040, Rev 0, Westinghouse Hanford Company, Richland, WA WHC, 1989, Operational Safety Analysis Report, Cross-Country Waste Transfer System, SD-WM-SAR-039, Rev 0, Westinghouse Hanford Company, Richland, WA WHC, 1988, Nonreactor Facility Safety Analysis Manual, Safety, Quality Assurance and Security, WHC-CM-4-46, Level II, Westinghouse Hanford Company, Richland, WA Wilson, G.P., 1992, Above Ground Transportation System, Memorandum from G.P. Wilson to O.S. Wang, May 12, 1992
|
NEWSLETTER
|
| Join the GlobalSecurity.org mailing list |
|
|
|
