Resolution of Generic Safety Issues: Issue 90: Technical Specifications for Anticipatory Trips (Rev. 2) ( NUREG-0933, Main Report with Supplements 1–34 )
Reactor protection systems (RPS) or "scram" systems are tripped by many diverse signals. The purposes of these various signals can be broadly divided into three classes: protection of the reactor core (e.g., overpower signals); protection of major components (e.g., vessel overpressurization signals); and anticipatory trips. The purpose of an anticipatory trip signal is to scram the reactor at the very beginning of a transient and thus minimize the degree of upset of the plant and avoid actuation of engineered safety features.
By definition, credit is not taken for an anticipatory trip in a plant's safety analyses, even to satisfy a single failure criterion. (Conversely, if a transient analysis calculation upon which TS are based takes credit for early reactor scram due to an "anticipatory" trip, the trip can no longer be considered "anticipatory.") Originally, it was AEC (and then NRC) regulatory policy to consider such trips to be installed for the licensees' convenience rather than for licensing purposes.630 Thus, no TS requirements were placed on these trips. At the time the STS was introduced, this policy was changed and all trips in the RPS were included in the STS. However, plants licensed prior to the advent of STS were not backfitted with TS on anticipatory trips. In addition, post-STS reviews of custom TS done after the STS were introduced did not require the inclusion of specifications on anticipatory trips.
This issue originated when Region II noted the anomaly and found that licensees were often performing little or no maintenance on these trips in the absence of TS requirements.630 With no mention of these trips in the TS, Region II could take little action even in cases where inspectors had reason to doubt the operability of these trips. DSI/NRR recommended that the issue be included in the broader study involving the overall adequacy of the TS. However, at the time this issue was evaluated, RSB/DSI had not defined a scope or schedule for a systematic study of all TS.
The safety significance identified630 was that, "because anticipatory trips are a part of the protection system, a failure or maintenance action in the anticipatory trip could cause other trips relied on in the accident analysis to be degraded below an acceptable level." The design of the RPS, however, leans strongly towards a fail-safe direction, i.e., failure of any channel may cause an inadvertent reactor scram, but should never prevent a trip of another channel from scramming the reactor. Questions of this nature are valid (e.g., can the W scram-on-turbine-trip interact with other channels via the P-7 interlock?), but the DL/NRR memorandum630 did not provide any specific concerns. Moreover, ICSB/DSI/NRR mentioned none in its examination of this issue,631 and the fault tree analysis of WASH-140016 (Appendix II, §5.2) found no such interaction. In the absence of specific concerns, it was concluded that this part of the issue was better treated under the auspices of the ATWS program and was not taken into consideration in this evaluation.
The second area of potential safety significance was a plant's response to a transient. Since anticipatory trips reduce the degree of plant upset, they also (in principle) reduce the frequency of challenges to engineered safety features, and thus (again, in principle) reduce the frequency of transient-initiated accidents.
The proposed solution was implicit in the definition of the issue: impose TS requirements on anticipatory trips. Such specifications would include: (1) limiting safety system settings (LSSS), which provide setpoints; (2) limiting conditions for operation (LCO), which require equipment to be operable during appropriate operational modes; and (3) surveillance requirements (SR).
A search through some older TS uncovered the following anticipatory trips:
|(1)||PWR - High neutron flux, source range|
|PWR - High neutron flux, intermediate range|
|(2)||PWR - Turbine trip|
|(3)||PWR - Low steam generator level coincident with steam-feed mismatch|
|(4)||BWR - High neutron flux, source range|
Because these trips are notable via their absence in the TS, there is no guarantee that this list is complete. Nevertheless, these anticipatory trips should be representative. There are other trips (e.g., PWR RCP breaker open) which have LCOs and SRs, but do not appear in the LSSS. Historically, this is because it is difficult to define "setpoints" for such trips. This is not a safety but a licensing improvement (enforceability) aspect and will not be considered here. Normally, calculations of scram failure probabilities are done for each channel of the RPS, explicitly accounting for the various 1/2, 2/3, and 2/4 logic matrices with estimates of common mode failure rates included. Using this data, the following simplified assumptions were made.
First, imposing TS on anticipatory trips does not affect the common mode failure rate of the RPS. Thus, when the changes in failure probabilities are calculated, the common mode failures are subtracted out. Second, lack of maintenance of an anticipatory trip is a common mode of both trip channels associated with the anticipatory trip parameter, i.e., if one channel fails, the other is also quite likely to fail. In the calculations, it was assumed that the probability that a trip signal will fail to cause a scram (i.e., both channels failing) is 0.01/demand on an anticipatory trip which is not in the TS. This assumption was based on judgment augmented by conversations with the senior inspector who originated the issue. All other trip parameters were assumed to have failure rates of 10-4/demand, based on Appendix II, §5.2 of WASH-1400.16
(1) PWR - High Neutron Flux, Source and Intermediate Range
These trips are functional only during reactor startup. They are backed up by the low setpoint of the power range neutron flux channels. Usually, the source range monitor setpoint is set at 105 counts/second (which is equivalent to roughly 0.01% of rated reactor power) and the intermediate range setpoint and power range low setpoint are both set at 25% of rated power. Thus, the source range is truly anticipatory in the sense of attempting to stop a transient early while the intermediate range backs up, but does not anticipate, the low power setpoint of the power range channels.
Rod Ejection: The first "transient" of interest is not an anticipated operational occurrence, but is considered an accident. A rod ejection is a reactivity excursion which is, if significant at all, too rapid for the anticipatory nature of the source range trip to make much difference. Thus, the safety contribution comes from increased scram reliability, not early scram. The following assumptions were made: 10 reactor startups/RY178; two days of vulnerability/startup (based on judgment); and 10-5 rod ejections/RY (based on WASH-1400,16 Appendix I, §4.3). This translates to 5.5 x 10-7 rod ejections/RY at low power. The change in scram failure rate is (10-2 - 10-4) for the source and intermediate range trips multiplied by 10-4 for the low setpoint power range trip. Thus, the change in core-melt frequency was given by:
F = (5.5 x 10-7/RY)[(10-2)2 - (10-4)2](10-4)
= 5.5 x 10-15/RY
The consequences are those of a partial core-melt (if all these trips fail) which was bounded by a PWR-5 release (core-melt with the containment not isolated). Using the assumption of a uniform population density of 340 persons/square-mile, a 50-mile radius, a central midwest plain meteorology, and no ingestion pathways, a PWR-5 release results in 106 man-rem. Thus, the estimated risk reduction was given by:
FR <(5.5 x 10-15/RY)(106 man-rem)
<5.5 x 10-9 man-rem/RY.
Short Periods: Under certain conditions of core burnup and high xenon inventory, differential rod worth tends to concentrate in a relatively narrow vertical range in the core during startup. This effect is much more pronounced in a BWR core (see Issue 6), but can also occur in PWRs. Should this occur, the reactor core will go suddenly from subcritical to supercritical with a rapid positive period (on the order of 10 seconds). Older plants do not have flux rate trips.
If the SRM trip fails and the operator does not terminate the transient manually (assume 10% chance), the reactor core will not be shut down until power reaches the 25% intermediate and power range setpoints. Fuel failure could also occur due to pellet/cladding interaction (PCI) during the rapid power ascension, or due to DNB because of the highly axially peaked power shape. At the time this issue was first evaluated in August 1984, no such events had occurred at a PWR. IE Bulletin No. 79-126 and IE Circular No. 77-075 listed 5 such events in BWRs as of May 31, 1979. This corresponded to about 155 BWR-years of experience.
It was assumed that the PWR frequency was at most one-tenth of the BWR frequency, or 3 x 10-3 event/PWR-year. The consequences were also bounded by those of PWR-8 and PWR-9 releases which corresponded to mitigated large break LOCAs without and with containment isolation (i.e., widespread cladding failure but no fuel melting). The WASH-140016 assumption of containment isolation failure probability is 0.1. The risk associated with the short period scenario was given by:
|FR||<(3 x 10-3 event/RY)[(10-2 - 10-4) SRM trip failure/event] x (0.1 operator failure/event) x [(120 man-rem/accident, containment isolated) + (0.1 isolation failure/accident)(7.5 x 104 man-rem/accident, containment not isolated)]|
|2.3 x 10-2 man-rem/RY.|
Rod Bank Withdrawal Error: This transient is characterized by a slower reactivity insertion rate than those of the transients discussed above. Thus, fuel failure is not likely to occur because of a high rate of power ascension at the beginning of the transient, but instead may occur due to DNB as the core comes into the power range, possibly with an adverse power distribution due to some rod banks remaining in the core.
For this to happen, the source, intermediate, and power range trips must fail. In addition, the rod stop must fail. A failure rate of 0.1 (rather than 0.01) for the rod stop was assumed since it was associated with the intermediate range detectors. No credit was taken for operator action since the cause of the event was probably operator error. Again, as discussed in the short-period transient, the consequences were bounded by those of PWR-8 and PWR-9 releases, assuming a containment isolation failure probability of 0.1.
In WASH-1400,16 the frequency of PWR uncontrolled rod withdrawal transients was estimated to be 0.01/RY. It was assumed that half of these occur during startup maneuvers. The risk estimate was given by:
|FR||< (0.01 rod withdrawal events/RY) x (0.5 percentage in startup)|
|x (10-2 - 10-4 source range failure rate change)|
|x (0.1 rod stop failure rate)|
|x (10-2 - 10-4 intermediate range failure rate change)|
|x (10-4 low setpoint power range failure rate)|
|x [(120 man-rem/PWR-9 release) + (0.1 containment failure rate) x (7.5 x 104 man-rem/PWR-8 release)]|
|< 4 x 10-8 man-rem/RY.|
For this transient to progress to core-melt, the high pressurizer pressure, high pressurizer water level, overtemperature and overpower T, and several other trips must fail. However, the DNB event frequency above is already down to about 5 x 10-12. Thus, even if no credit was taken for these additional trips and the core-melt resulted in the worst case consequences (5.4 x 106 man-rem from a PWR-1 release, where the core-melt causes a steam explosion which ruptures both the reactor vessel and the containment), the resulting public risk would be only 3 x 10-5 man-rem/RY.
Boron Dilution: This transient is caused by a CVCS malfunction which dilutes soluble boron in the reactor moderator. The reactivity insertion rate is very slow, on the order of 10-5 /second at the start and diminishing asymptotically to zero as the moderator becomes more dilute. Thus, the transient is also very slow, giving the operator as much as an hour or more to take action for a dilution event during startup. Because the power increase is slow, fuel failures due to PCI are not expected.
For a PWR core at BOC conditions, there is insufficient reactivity worth in the control rods to maintain the core subcritical with no soluble boron in the moderator. Thus, a reactor scram does not permanently terminate the transient; operator action is necessary to stop the dilution and re-borate the moderator. However, a reactor scram does give the operator more time to respond. In EPRI NP-801,178 the frequency of boron dilution events was estimated to be 0.03 event/RY. Again, it was assumed that almost half of these events occur during startup.
As the event progresses, the operator (for whom credit was given for this slow transient) must fail to observe the transient and take no action (assume probability 0.10). The source range trip must fail (0.01 - 0.0001), the intermediate range trip must fail (0.01 -0.0001) and the low setpoint power range trip must fail (0.0001). At this point, the estimated change in frequency is reduced to 1.5 x 10-11/RY. Reactor power increases past 25% and thermal energy is dumped via the main condenser steam dump, the ADVs, and/or the steam generator safety valves, depending on plant conditions. Eventually, as reactor power increases to a value too great to be dissipated by these means, trip signals on pressurizer high pressure, pressurizer high water level, steam generator low-low water level, turbine trip (which functions as a 50% power trip as the P-9 permissive is reached), overtemperature T, and several others may scram the reactor. (Fuel still has not been damaged.) However, even with no credit for these non-neutron-flux trips, a frequency of 1.5 x 10-11/RY was estimated. Thus, this transient was not considered further.
(2) PWR - Turbine Trip
A turbine trip, normally sensed as either two out of three low autostop oil pressure signals or four out of four turbine stop valves closed, causes a reactor scram when the plant is operating above a preset power level (e.g., 10%). If a turbine trip occurs and this scram fails, steam pressure will rise in the secondary system. The atmospheric dump valves and steam generator safety valves will be available to limit the pressure rise, and the steam dump (which is usually in Tavg mode during power operation) will open and dump steam directly to the condenser. However, before these alternate energy sinks become available, the primary system will experience a rapid heat up. Expansion of the primary coolant will force coolant into the pressurizer, compressing the steam bubble. Reactor scram signals will be generated by the high pressurizer pressure, over-temperature T and high pressurizer level signals, and the transient will "turn around." Pressurizer spray will also turn on to limit the primary pressure transient, but it is likely that the PORVs will open.
This evaluation was restricted to W and CE plants. Upgraded anticipatory reactor trips on turbine trips were required of B&W plants by TMI Action Plan Item II.K.2(10) which was implemented under MPA F-28. The safety significance of this trip is two-fold: (1) the anticipatory trip increases the reliability of the entire RPS, thus decreasing the frequency of ATWS events; and (2) by preventing opening of a pressurizer PORV, the frequency of a small LOCA is also decreased. In EPRI NP-801,178 the following transient frequencies involving a turbine trip are listed:
|Loss of Condenser Vacuum||0.12/RY|
|Loss of Circulating Water||0.07/RY|
The last two initiators will also disable the steam dump and, on plants with turbine-driven main feedwater pumps, cause a loss of feedwater. However, tripping of the main turbine is the first event to cause a reactor transient.
It was assumed that a PORV opens about half the time in these transients. Also, the probability that a PORV will fail to re-close was estimated to be 1% per actuation (WASH-1400,16 Appendix V, §4.3.1). The S2 LOCA sequence is then a turbine trip transient (2.12/RY), failure of the anticipatory trip channels (10-2 - 10-4), opening of a PORV (0.5), failure of the PORV to re-close (0.01), and failure of the operator to correctly diagnose the problem and close the PORV block valves (assumed to be 0.1). The change in S2 frequency was then 10-5/RY.
The scram reliability is increased by the diversity of an anticipatory trip. However, placing TS on the anticipatory trip does not affect the common mode failure rate (CM) of the RPS. If the high pressurizer pressure, overtemperature T, and high pressurizer level signals have failure rates of 0.0001/demand, and the imposition of TS reduces the turbine trip signal failure rate from 0.01 to 0.0001, the RPS failure rate is:
(0.0001)3(0.01) + CM without anticipatory trip TS
(0.0001)3(0.0001) + CM with anticipatory trip TS.
The change in RPS failure rate is 9.9 x 10-15; the common mode contribution cancels out. This is negligible compared with the S2 sequences, therefore, it was not considered further.
The public risk associated with the S2 sequences was obtained by normalizing the WASH-140016 results (where an S2 frequency of 10-3/RY was assumed) to a frequency of 10-5/RY (see WASH-140016 Table V, 3-14. The results are shown in Table 3.90-1.
|Release Category||Normalized S2 Frequency/RY||Consequence (man-rem)||FR
|PWR-1||1.0 x 10-9||5.4 x 106||5.4 x 10-1|
|PWR-2||3.0 x 10-9||4.8 x 106||1.4 x 10-2|
|PWR-3||3.0 x 10-8||5.4 x 106||1.6 x 10-1|
|PWR-4||3.0 x 10-9||2.7 x 106||8.1 x 10-3|
|PWR-5||3.0 x 10-9||1.0 x 106||3.0 x 10-3|
|PWR-6||2.0 x 10-8||1.5 x 105||3.0 x 10-3|
|PWR-7||2.0 x 10-7||2.3 x 103||4.6 x 10-4|
|Total:||2.6 x 10-7||2.0 x 10-1|
(3) PWR - Low Steam Generator Level Coincident with Steam-Feed Mismatch
This anticipatory trip will scram the reactor on low steam generator water level coincident with steam flow greater than feedwater flow by a preset amount (usually 40% of rated). It is backed up by the low-low steam generator water level trip. The initiating event here is a total loss of feedwater event in any steam generator. Partial loss of feedwater events or total loss of feedwater flow at reduced power levels may not produce sufficient mismatch between steam and feedwater flow to actuate the anticipatory trip; such events are also slower and early scram is not as important.
If all feedwater pumps are lost, the secondary side water temperature will rise because of the loss of the relatively cool feedwater, the heat transfer across the steam generator tubes is reduced and the primary side heats up. Simultaneously, the secondary water level decreases. The anticipatory trip signal will occur, and then the low-low steam generator level trip. The increasing primary temperature and resultant coolant swell will force more coolant into the pressurizer, compressing the steam bubble and causing an increase in primary system pressure. Reactor trip signals on overtemperature T, high pressurizer pressure, and high pressurizer level will occur if the reactor has not already been scrammed, and the PORVs may open to limit the pressurizer pressure. The AFW system will also be initiated by the loss of main feedwater pumps or by low-low steam generator water level. If the AFW system fails and the steam generator tubes uncover, primary side temperature and pressure will rise more rapidly and the pressurizer safety valves will open. However, the probability of this event is not greatly affected by reliability of the anticipatory scram. Thus, it was not considered in this evaluation.
As in the turbine trip transient evaluated previously, the anticipatory trip decreases the probability of an ATWS event and also helps prevent a pressurizer PORV from opening, thus decreasing the frequency of a small LOCA. The transients that will cause a loss of feedwater were taken from EPRI NP-801178 and are listed in Table 3.90-2.
As in the turbine trip transient, it was assumed that a PORV opens half the time if the anticipatory trip fails and that the PORV fails to re-close 1% of the time. The S2 frequency is then 2.28 transients/RY multiplied by the change in probability of anticipatory trip failure (0.01 - 0.0001), the probability of PORV opening (0.5), the probability of failure of the PORV to re-close (0.01), and the probability of operator failure to close the block valve (0.1). The result of 10-5/RY is the same as the turbine trip case and thus the risk figures are the same: 2.6 x 10-7 core-melt/RY, 0.2 man-rem/RY.
The change in the probability of RPS failure is given by the change in anticipatory trip failure (0.0099) multiplied by the non-common mode failure probabilities of the trips on low-low steam generator level, over-temperature T, high pressurizer pressure, and high pressurizer level, each of which is 0.0001. The resulting change in ATWS frequency is on the order of 2.3 x 10-18/RY which is negligible compared to the S2 LOCA considerations above.
(4) BWR - High Neutron Flux, Source Range.
As in the PWR case, BWR licensing basis transient calculations for startup events do not take credit for the source range monitor (SRM) and intermediate range monitor (IRM) trip setpoints, but instead assume that the reactor is scrammed by the startup mode setpoint of the average power range monitor (APRM) system, usually 15% of rated power. Unlike the PWRs, the IRM scram setpoints are already in the TS; the SRM scram setpoints are not required (although the monitoring function of the SRM is addressed). Moreover, it is common (if not universal) practice to disable the SRM scram inputs with shorting links after the plant's initial core loading is complete. Thus, the SRM scram generally has a failure probability of 1.
|Loss of feedwater (one loop)||0.99/RY|
|Loss of feedwater (all loops)||0.08/RY|
|Feedwater instability (operator error)||0.66/RY|
|Feedwater instability (mechanical problem)||0.50/RY|
|Loss of one condensate pump||0.05/RY|
|Loss of all condensate pumps||0.00/RY|
However, the SRM scram, at its usual setpoint of 5 x 105 counts/second, is generally not the first scram to occur during a startup transient or accident. The reason is that, in a BWR, the IRM rod block and scram setpoints are defined as percentages of full scale for each IRM range, and the IRM ranges cover five decades. The SRMs and IRMs must overlap. SRMs are interlocked such that they cannot be withdrawn unless the IRMs are on Range 3 or above. If the IRMs are on Range 1 (and if they are not, a rod block on IRM Downscale will prevent rod withdrawal), the IRM scram will occur virtually simultaneously with or (more likely) prior to the SRM scram. Therefore, the SRM scram will not reduce plant upset; it will only somewhat increase the reliability of the RPS. The events of interest are the rod drop accident (RDA) and the short period transient. (Because of the individual rod pulls used in a BWR, there is no analog to the PWR rod bank withdrawal error.)
RDAs and probabilities are discussed in an RDA Statistical Analysis.7 The basic sequence starts with 2 x 103 rod withdrawals/RY in the startup range. To get an RDA, a rod must disconnect (maximum 2 x 10-4), become stuck (maximum 10-2), and become unstuck and drop at the appropriate time (maximum 6 x 10-3). This does not imply that the rod is out of sequence or will lead to high worth. This translates into 2.4 x 10-5 RDA/RY requiring a reactor scram. The maximum improvement the SRM scram channel can make is (1 - 0.0001)(0.0001)2 or 10-8. The F involved is then 2.4 x 10-13 event/RY; this is a negligible frequency. Even if such an event ruptured both the vessel and containment and completely melted the entire core under the worst conditions (i.e., BWR-2 release), the maximum public risk would be 1.7 micro-man-rem/RY.
Short period events are more common. As mentioned earlier, IE Bulletin No. 79-126 and IE Circular No. 77-075 list 5 such events in 155 BWR-years, a frequency of 3.2 x 10-2/RY. These events will happen with the IRM channels set on their first range and the SRM scram, if functional, would not anticipate other scrams but instead would provide a backup to the IRM scram. Rod blocks are largely ineffective here since the high incremental rod worth is tied up in one 6-inch notch.
Historically, these events have occurred just at the point of criticality. Since the operator is using the period meters to detect criticality, a short period event is easily noticed and these events have generally been terminated by operator action, not by the RPS. If the operator does not intervene (assuming a probability of 0.1) and the IRM and APRM scrams fail (probability = 10-8), the reactor core would ascend into the power range where the usual reactivity coefficients would turn the transient around. The only consequence would be some cladding failure due to PCI. The consequences to the public can be bounded by those of a licensing basis RDA in which 770 fuel rods fail. These consequences are 0.007 man-rem/event. Thus, the net risk from short period events, even with no credit for the SRM scram is, at most, (3.2 x 10-2/RY)(0.10)(10-8 )(0.007 man-rem) = 2.2 x 10-13 man-rem/RY. Again, this is negligible.
Based on the calculations in (1), (2), and (3) above, the core-melt frequency estimate was [(5.5 x 10-15) + (2.6 x 10-7) + (2.6 x 10-7)]/RY or 5.2 x 10-7/RY. The public risk reduction was estimated to be [(5.5 x 10-9) + (2.3 x 10-2) + (4 x 10-8) + (2 x 10-1) + (0.2)] man-rem/RY or 0.42 man-rem/RY.
Approximately 20 PWRs would be affected by the proposed action. Action on BWRs was unlikely to be approved since the potential safety gain from the SRM trip was so small.) Assuming an average remaining life of 20 years for the affected plants, the total remaining operating life was 400 RY. Thus, the total public risk reduction associated with this issue was approximately 170 man-rem.
Industry Cost: TS on anticipatory trips would probably be similar to those in the PWR STS. This would involve the following for Source Range Neutron Flux: (1) channel checks every shift except during power operation; (2) calibration every refueling outage; and (3) analog operational tests monthly and prior to startups. For Intermediate Range Neutron Flux, this would include: (1) channel checks every shift while in startup; (2) calibration every refueling outage; and (3) analog operational test monthly and prior to startup. For steam generator low level/steam-feed mismatch this would include: (1) calibration every refueling outage; and (2) analog operational tests monthly. For turbine trip, this would include trip actuation device tests prior to startup. It was assumed that there would be 10 startups/year, 2 months refueling outage every 18 months, 5 days to go from cold shutdown to power, 10 minutes for channel checks, 4 hours for calibrations, and 1 hour for analog and trip device tests. Based on a cost of $100,000/man-year, the industry cost was estimated to be $4,000/RY or $1.6M for the affected plants.
NRC Cost: It was assumed that 2 man-years of effort would be needed to develop and carry out an action plan for the issue, including developing CRGR packages, etc., and that monitoring licensee compliance would take 8 hours/RY of inspection time.
Total Cost: The total cost of implementing the possible solution to this issue was estimated to be $98,000/reactor or approximately $2M for all affected reactors.
Based on an estimated public risk reduction of 170 man-rem and a cost of $2M for a possible solution, the value/ impact score was given by:
The above calculations did not include credit for averted cleanup costs to the licensee. It should be rem that avoidance of PORV opening (and secondary side safety valve opening) is a major reason for installing anticipatory trips. Inclusion of averted cleanup costs as a credit against the cost to the licensee could significantly raise the priority score. Moreover, the occupational man-rem averted by preventing PORV opening (and possible rupture of the pressurizer relief tank rupture disc) might also be significant.
Based on the low risk reduction potential and low value/impact score, this issue was given a low priority ranking (see Appendix C) in August 1984. In NUREG/CR-5382,1563 it was concluded that consideration of a 20-year license renewal period did not change the priority of the issue. Further prioritization, using the conversion factor of $2,000/man-rem approved1689 by the Commission in September 1995, resulted in an impact/value ratio (R) of $11,765 /man-rem which placed the issue in the DROP category.