Resolution of Generic Safety Issues: Item B-17: Criteria for Safety-Related Operator Actions ( NUREG-0933, Main Report with Supplements 1–34 )
This NUREG-04713 item involved the development of a time criterion for safety-related operator actions (SROA), including a determination of whether or not automatic actuation was to be required. At the time this issue was identified in 1978, existing plant designs were such that reliance on operators to take action in response to certain transients was necessary. In addition, some existing PWR designs required manual operations to accomplish the switchover from the injection mode to the recirculation mode following a LOCA. The required time for the ECCS realignment operations was dependent on the size of the pipe break, and the operation was expected to be accomplished before the inventory in the borated water storage tank was depleted. The evaluation of this issue included consideration of Issue 27.
Development and implementation of criteria for SROA would result in the automation of some actions that were being performed by operators. The use of automated redundant safety-grade controls in lieu of operator actions was expected to reduce the frequency of improper action during the response to or recovery from transients and accidents, by removing the potential for operator error. This, in turn, could reduce the expected frequency of core damaging events and, therefore, reduce the public risk accordingly.
Plants would be required to perform a task analysis, simulator studies, and analysis and evaluation of operational data to assess existing ESF and safety-related control system designs for conformance to new criteria. Where non-conformance was identified, modification to existing designs and hardware would be required. For plants at the CP stage of review, changes and additions to the ESF control systems were anticipated, but replacement equipment costs were not anticipated.
In the analysis of this issue the following major assumptions were made:
(a) Operator error comprised 40% of total plant risk
(b) 10% of short-term emergency response actions were taken by operators
(c) 50% of long-term emergency response and recovery actions were taken by operators
(d) One-half of operator actions being taken in the short-term would be automated
(e) 20% of operator actions being taken in the long-term would be automated
(f) the failure rate of automated ESF controls was on the order of 10-4 /demand
(g) the failure rate of trained and practiced operators was on the order of 10-2 /demand in the highly stressed short-term period, and 10-3/demand in the less stressful long-term period.
Using WASH-140016 frequencies, existing estimates of the doses to be expected for the various PWR and BWR release categories, and the projected population and remaining operating life of PWRs and BWRs, a total plant risk of 2.8 x 105 man-rem was determined. Operator contribution to total plant risk (40%) was thus estimated to be 1.12 x 105 man-rem. Of this risk, one-half was attributed to the short-term response period and one-half to the long-term response period.
Using the above stated assumptions on operator error and automated control system failure rates and the portion of short-term and long-term actions allocated to the operator, a short-term potential public risk reduction for completion and implementation of SROA criteria was estimated to be 2.8 x 103 man-rem. Resolution of the issue was estimated to provide a potential long-term public risk reduction of 5 x 103 man-rem. Thus, a total potential public risk reduction of 7.8 x 103 man-rem was estimated, and an average potential public risk reduction of 50 man-rem/reactor was estimated. Assuming an average core-melt consequence of 5 x 106 man-rem/event, a potential reduction of core-melt frequency of 3.8 x 10-7/RY and 5.4 x 10-5/reactor was estimated.
Industry Cost: Designers and/or operators of all plants were assumed to perform a design review and analysis of their existing ESF and safety-related control systems, and prepare modification packages for NRC review and approval. Comparison of existing designs to new criteria, preliminary design, final design, and NRC documentation were estimated to require 1 man-year/plant since most plants were multiple unit designs. Thus, the design cost for 143 plants was estimated to be $14.3M.
Equipment costs were divided into two groups: (1) older plants; and (2) recent and future plants. Recent and future plants were separated because of existing requirements for the automation of ECCS switchover to recirculation and automatic initiation of AFW systems. Backfit equipment and installation costs for older plants were estimated at $500,000/plant while the newer plants were estimated at $250,000/plant. Using the above breakdown on newer and older plants, a total equipment and installation cost of $53.7M was estimated. No additional recurring costs were estimated for operational maintenance and surveillance of the automated control systems since maintenance and surveillance would have been required for the manual control systems which were assumed to be replaced.
NRC Cost: The FY-1983 RES contract (FIN B0421) with ORNL included efforts by ORNL and its subcontractor (General Physics Corp.) to complete operator task analyses, simulator studies, operational data collection and analysis, and the development and recommendation of SROA criteria. This work was being pursued as part of Item I.A.4.2, "Long-Term Training Simulator Improvements."48 Completion of the above efforts, review of the above, and development of SROA criteria, review and approval of new criteria, orders to licensees and applicants, and review and approval of licensee and applicant responses were estimated to cost $4M over a 5-year period.
Total Cost: The total industry and NRC cost associated with the possible solution to this issue was estimated to be $(4 + 14.3 + 53.7)M or $72M.
Based on a potential public risk reduction of 7.8 x 103 man-rem and an estimated cost of $72M for a possible solution, the value/impact score was given by:
Uncertainties for this analysis were very large due to the subjective nature of the approach to operator error reduction. It was acknowledged that a more deterministic design-specific analysis, which might be performed after the Item I.A.4.2 SROA criteria recommendations were developed, could have altered the value/impact score for this issue by one to two orders of magnitude in either direction.
The value/impact score calculated was indicative of a medium priority ranking (see Appendix C). It was recommended that, after the conclusion of the SROA criteria development efforts on Item I.A.4.2, a more rigorous analysis should be performed to reassess the value/impact associated with the adoption and implementation of specific SROA requirements which were not available at the time this issue was evaluated in March 1982.
In resolving the issue, the staff concluded that the following actions taken by licensees, in response to regulatory requirements issued since the issue was identified, addressed the safety concern: (1) enhanced operator training and licensing requirements, including plant-specific simulators; (2) improved training, based on the Systems Approach to Training for all covered staff; (3) implementation of symptom-based emergency operating procedures; and (4) the completion of the Individual Plant Examination (IPE) Program at all operating plants. Thus, the issue was RESOLVED with no new or revised requirements.1766