Resolution of Generic Safety Issues: Issue 68: Postulated Loss of Auxiliary Feedwater System Resulting from Turbine-Driven Auxiliary Feedwater Pump Steam Supply Line Rupture (Rev. 3) ( NUREG-0933, Main Report with Supplements 1–34 )
In an evaluation of INPO/NSAC Significant Operating Experience Report 81-17,271 the operators at Fort Calhoun determined that the configuration of their plant made it susceptible to the possibility that all AFW supply could be disabled by a break of the steam supply line inside the pump room. The steam line break would disrupt the supply of steam to the turbine-driven pump and concurrently disable the electric motor-driven pump, if the electric pump motor were not qualified to operate in the steam environment and were located in the same pump room. The Fort Calhoun operators reported this deficiency in LER 82-012.
In the analysis,271 the Fort Calhoun operators identified a sequence which begins with the loss of offsite power and postulated a break in the steam supply line when the AFW pumps are required to operate. Fort Calhoun has the electric motor-driven AFW pump housed in the same room as the steam turbine-driven AFW pump. In reviewing the LER submitted by Fort Calhoun, NRR concluded that the design met current acceptance criteria and that the scenario postulated, loss of offsite power followed by the passive failure of the steam supply line disabling all AFW pumps, was outside the scope of events postulated as part of the current licensing basis and did not represent a credible accident scenario.272
Concerns were expressed in AEOD/T302273 that a single passive failure could result in the loss of a safety system that could be required to bring the plant to a safe shutdown. It was suggested that the pipe break criteria presented in SRP11 Section 3.6.1 be reviewed to determine if additional guidance was necessary. Similar problems were identified at San Onofre Units 2 and 3 and at Arkansas Nuclear One, Unit 1.
A possible solution to this issue was the relocation of the turbine-driven pump to another room separate from the electric-driven pump(s). An alternative to this first solution was locating a full capacity electric pump in a room separate from the steam-turbine pump and any effects resulting from a steam line supply break. A second solution was the replacement of the electric pump motor(s) with Class 1E environmentally-qualified components. A third solution, the alternative implemented at the San Onofre facility, was the addition of a forced lube oil cooling system to the electric pump motor bearings, assuming that everything else was already qualified. Finally, an augmented ISI of the steam lines was also proposed as an alternative solution.
Three event sequences were analyzed and assessed separately for CE and B&W reactors to determine the frequency at which core-melt would be expected to occur, as a result of placement of all AFW pumps in the same room. The sequence initiators appraised were: (1) a break of the AFW turbine-driven pump steam supply line inside the pump room; (2) a transient reactor trip other than interruption of the main feedwater; and (3) a loss of offsite power for longer than 15 minutes or interruption of the main feedwater.
For the first sequence, the turbine driven AFW pump steam supply line break occurs at a frequency of 2 x 10-3/RY. This failure rate was estimated by assuming one AFW steam line break has occurred in 440 RY (which had not happened). Since only 10% of the steam line is in the pump room, the frequency is one tenth of this or 2 x 10-4/RY. The dominant sequence events and their probabilities which follow are: a turbine trip (0.9); offsite power is not lost (~1.0); the electric motor-driven AFW pump(s) fail due to the operating environment (0.9); the operators fail to align the steam generators back to the main feedwater system (0.1); and for only the B&W reactors, the operators are unsuccessful in achieving feed-and-bleed operation to remove decay heat with a probability of P. The frequency of this sequence was estimated to be less than P(2 x 10-6) event/RY for B&W reactors and ~2 x 10-6 event/RY for CE reactors. B&W reactors were given credit for feed-and-bleed operation while CE reactors were not given credit for this mode. Many CE reactors do not have sufficient pressure relief capability to permit injection of coolant water at the lower HPCI pressure provided by the CE reactors. The success of feed-and-bleed is difficult to establish on a generic basis and it has not been demonstrated to be achievable, in that the availability of procedural guidance and training in this operation is undetermined. Therefore, no probabilistic estimate of success is assigned.
The second sequence is initiated by transient trips other than the interruption of main feedwater or loss of offsite electrical power. These transient trips have a frequency of 7/RY. The dominant sequence events which follow and their probabilities are: retention of offsite power (~1.0), a rupture or break of the steam supply line to the turbine driven AFW pump inside the pump room (2 x 10-5 based on ten demands on the AFW per year), failure of the electric motor-driven AFW pump(s) due to the operating environment (0.9), failure by the operators to align the steam generators to the main feedwater system (0.1), and the operators are unsuccessful in achieving feed-and-bleed cooling (a value of P for B&W reactors). The frequency of the sequence was estimated to be less than P(1.4 x 10-6) event/RY for B&W reactors and ~1.4 x 10-6 event/RY for CE reactors.
The third sequence is initiated by the loss of offsite power for more than 15 minutes which has a frequency of 0.2/RY or an interruption of main feedwater which may be expected to occur at a frequency of 3/RY. The events following in the dominant sequence are: the steam supply line to the turbine-driven pump breaks inside the pump room (2 x 10-5), emergency power is supplied by the diesel generators (~1.0), the electric motor-driven AFW pumps fail due to the operating environment (0.9), and for B&W reactors the operators are unable to achieve feed-and-bleed cooling (a value of P). The frequency of this sequence is estimated to be less than P(5 x 10-5) event/RY for B&W reactors and ~6 x 10-5 event/RY for CE reactors.
The sum of the three dominant sequences for B&W reactors is ~P(5.3 x 10-5) event/RY. For CE reactors, the sum of the three dominant sequences is ~6.3 x 10-5 event/RY. These values closely approximate the increase in core-melt frequency resulting from placing the steam turbine-driven pump and non-qualified electric motor-driven pumps within the same enclosure. It should be noted that the sequences involving loss of electric power, which were the original concern, do not dominate the results.
The consequences of these sequences are obtained using the CRAC Code64 for the release fractions and categories of a PWR as given in WASH-1400.16 The calculations assume an average population density of 340 persons per square mile (which is the average for U.S. domestic sites) from an exclusion area of one-half mile about the reactor out to a 50-mile radius about the reactor. A typical midwest plain meteorology is also assumed.
The sequence described would be similar to the T1MLU sequence described in NUREG/CR-165954 for Oconee resulting in a Category 3 release. For B&W reactors, the risk is P(144) man-rem/RY or P(3,850) man-rem/reactor for the remaining life of the reactor. If P were considered to be 0.9, which means that only one in ten attempts to cool the core by using feed-and-bleed is successful, then the risk would be 130 man-rem/RY or 3,460 man-rem/reactor. For CE reactors, the risk is 170 man-rem/RY or 4,600 man-rem/reactor.
It was estimated that an enhanced ISI effort would reduce the frequency of steam supply line breaks and the effects which result by a factor of 5. Enhanced inspection of the AFW steam supply line would reduce the frequency of core-melt accidents involving the failure of the steam supply line from 5.3 x 10-5/RY and 6.3 x 10-5/RY to 1.1 x 10-5/RY and 1.3 x 10-5/RY for B&W and CE reactors, respectively. The core-melt reduction reduces the public risk by 227 and 270 man-rem, respectively, for the 27 years remaining reactor life.
Industry Cost: It was estimated274 by one plant that relocating the turbine pump would cost $13.5M. The cost to install Class 1E environmentally-qualified motors was estimated to be $5.2M. The addition of the forced-cooled lube oil system was estimated to be $2.5M. (While the forced-cooled lube system could permit the continued operation of the electric pump motors in the steam environment at San Onofre, it is not certain that a similar solution would be possible at the other affected plants.)
The cost to perform ISI was estimated to be 0.1 man-year/RY and included efforts necessary to erect scaffolding, remove insulation materials, perform and evaluate the inspections, and to restore the system to an operable configuration. In addition, a one-time cost of $30,000 was estimated to be necessary to make some piping changes to permit the inspection of some welds. The total cost estimated to perform ISI of the AFW steam supply line inside the pump room for the remaining plant life is $0.3M.
As described, the costs for modifications to eliminate the risk associated with this issue is estimated to be between $2.5M and $13.5M. For B&W reactors having a risk exposure of 3,850 man-rem/reactor, the value/impact score varies between 1,540 and 285 man-rem/$M per reactor. For CE reactors having a risk exposure of 4,600 man-rem/reactor, the value/impact score varies between 1,840 and 340 man-rem/$M per reactor. ISI has a value/impact score of 756 and 900 man-rem/$M per reactor for B&W and CE reactors, respectively.
(1) The accident avoidance savings, based upon a resolution which reduces the core-melt frequency by 5 x 10-6 event/RY, is $82,000/RY or $2.2M for the lifetime of the reactor.64
(2) No increase in ORE is expected to result from the proposed solution to this issue. The results of the analysis are sensitive to changes in the assigned failure rates. An order of magnitude change in accident frequency could change the priority ranking.
The value/impact score for this issue indicated a high priority for those facilities that have pumps co-located and suceptible to common cause failure.
A review and evaluation of the criteria for judging the acceptability of proposed plants to determine if similar practices would be judged acceptable was deemed to be of high priority. The requirement for redundancy and diversity to assure the availability of an important safety system was of dubious benefit, if designs are permitted in which environmental or single-fault conditions may render the diverse or redundant components inoperable. This issue was integrated into the resolution of Issue 124.