4

uniformity: "These [fire barrier cable] penetrations must be sealed to achieve a degree of fire resistance equivalent to that required of the barrier that is pierced." 45 Fed. Reg. at 76608. The retention of this requirement in the final rule is evidence of the Commission's special concern that a rule of uniformity, rather than a more flexible rule of reason, be observed for cable penetrations. The degree of the Commission's concern may well have been a result of the fact that the fire at Brown's Ferry started at a point where cables penetrated a barrier.

Thus, the staff's proposed interpretation, insofar as it treats properly evaluated non-uniformities other than those where cables penetrate barriers as being in compliance with the Appendix R, accords with the Commission's intent in promulgating the requirements for fire barriers. Indeed, it appears that the staff has had no strong legal basis for requiring exemption for non-uniformities other than those where cables penetrate barriers.

However, standing alone, the staff's proposed interpretation, by the generality of its language, appears to include non-uniform cable penetration seals in the class of non-uniformities which might be treated as complying with the Appendix. The proposed interpretation thereby appears not only to run counter to the intention of Section III.M of the final rule, but also to the explicit meaning of the text of that section. To make clear that uniformity as specified in sec. III.M is still the rule for cable penetration seals and that a licensee wanting to deviate from the rule must apply for an exemption, the staff proposes to insert the following sentence between the last two sentences of the interpretation: "However, if certain cable penetrations were identified as open SER items at the time Appendix R became effective, section III.M of the rule applies (see 10 C.F.R. 50.48 (b)), and any variation from the requirements of section III.M requires an exemption."

The interpretation presents a difficult interpretive question in one other respect. As the interpretation now stands, it would leave open the possibility that an incompletely sealed barrier could be in compliance with Appendix Indeed, the principal aim of the interpretation is to leave open this possibility. In support of this interpretation staff could argue that when the final rule dispensed with what the accompanying comment called the "excessively restrictive" requirement for uniformity in barriers, the final rule also dispensed with the requirement in the proposed rule that the barriers be completely sealed. Section M of the proposed rule had said that "Penetrations in these fire barriers . . . shall be sealed or closed to provide fire resistance rating equivalent to that required

5

of the barrier." Nothing from this sentence appeared in the final rule. The staff could argue that had the Commission wanted to promulgate a requirement that barriers be completely sealed, the final rule could have simply dispensed with the phrase "to provide . . . rating equivalent. in the quoted sentence and retained the clause with the words "shall be sealed" in it.

However, arguments can also be made to the contrary. First, all that can be concluded with certainty from the fact that the quoted sentence was not retained in the final rule is that penetrations need not necessarily be sealed to the rating of the surrounding barrier, not that penetrations need not be sealed. Second, although the final rule does not explicitly say that barriers must be completely sealed, such a requirement is a fair implication of the content and length of certain comments accompanying the final rule. "The best fire protection for redundant trains of safe shutdown systems is separation by unpierced fire barriers . . ." 45 Fed. Reg. at 76608. Virtually the same wording occurs later on the same page. The comment does allow for openings, but not, apparently, unsealed ones: "Even fire barriers with openings have successfully interrupted the progress of many fires provided the openings were properly protected by fire doors or other acceptable means." Id. (emphasis added) In context, the phrase "acceptable means" is likely to refer not simply to the absence of a fire hazard near an unsealed opening in a barrier, but rather to devices such as are mentioned by either the proposed rule or the final rule, e.g., fire door dampers and barrier penetration seals. See Id. at 36089; see also section III.M of Appendix R. Note also that in both the proposed rule and the final rule, one whole section is given over to describing means for assuring that fire doors are kept closed, thereby evidencing Commission concern about openings in barriers. See 45 Fed. Reg. at 36090; see also section III.N of Appendix R. As to length of comment, the technical comment given in justification of those parts of the final rule which are about barrier penetrations is longer than the technical comment given on any other subject in the final rule (understandably so, given how the Brown's Ferry fire started), and nowhere in that technical comment on barriers did the Commission suggest that barriers in compliance with the Appendix might have unsealed openings. In the face of the evidence in the final rule comments, the Commission's concerns about barrier penetration, and the lack of an explicit allowance in the rule and comments for unsealed openings, the better legal argument concludes, we think, that the Commission's intent in promulgating Appendix R was to require that openings in fire barriers be sealed. The text could be modified accordingly to make it

6

clear that although in some circumstances a non-uniform fire barrier may be in compliance with the Appendix, no barrier with an unsealed opening will be, and that a licensee wanting to install such a barrier must apply for an exemption.

However, the contrary argument, based on the deletion of the proposed rule language, is not totally lacking in legal merit. We believe that the relative weakness in the contrary argument incorporated into the interpretation could be substantially overcome by a strong technical argument that unsealed openings do not, as a general matter, necessarily pose fire hazards, and that there is nothing inconsistent, as a matter of safety, in allowing unsealed openings but enforcing stringent requirements on fire barrier cable penetration seals and fire doors.

The staff's proposed interpretation of the time limits in Appendix R relating to cold shutdown equipment appears to us to be in accord with the text of the pertinent sections of Appendix R and the intention behind them. However, our reasoning differs some from staff's, and, we hope, clarifies the intent of the time limits.

Section III.L.5 of Appendix R requires that fire protection features be capable of limiting fire damage suffered by alternative or dedicated means for achieving cold shutdown so that such means can be repaired, using on-site resources, and cold shutdown "achieved within 72 hours." (Emphasis added.) The same 72-hour standard for the capability to achieve cold shutdown also appears in subsection 1 of the same section, III.L.

Section III.G.1.b, on the other hand, requires that fire protection features be capable of limiting fire damage suffered by normal means of achieving cold shutdown so that such means "can be repaired within 72 hours." (Emphasis added.) Again, the same 72-hour standard for the capability to repair within 72 hours also appears in the chart in section I (Roman numeral "1") of Appendix R, with the addition that, here too, repair is to be accomplished with on-site resources.

The staff's second major interprecation proposes that this distinction which the Appendix consistently makes between repair in 72 hours and achievement in 72 hours be taken at face value, and thus that Appendix R not be read to require that plants going to cold shutdown after a fire get there within 72 hours. SECY-85-306, Enclosure 4, at 1.

7

The May 1984 Memorandum claims that there is a "generally accepted view" that cold shutdown has to be achieved within 72 hours after a fire, and that the staff's literal reading, without any stated safety basis, divides reactors into two classes: a class of those which must be able to achieve cold shutdown within 72 hours, and a class of those which need only be able to repair cold shutdown equipment within 72 hours. May 1984 Memorandum at 3-4.

The staff answers first that it knows of no "generally accepted view" that cold shutdown has to be achieved within 72 hours. Indeed, the staff points out, it is recognized that the more stringent rule that alternate equipment be able to achieve cold shutdown within 72 hours has not sensibly applied in some areas of some plants, and that exemptions have therefore been granted in those instances. SECY-85-306, Enclosure 8, at 2-3.

The staff further argues that Appendix R itself makes the division to which the Memorandum objects, not, however (the staff adds by way of clarification), between plants, but between areas within the same plant, since a plant with alternate shutdown systems does not necessarily have an alternate system for every normal system in the plant. Id., at 3.

"In any event," the staff says, "there is little, if any, safety distinction to be made between having the capability to achieve cold shutdown within 72 hours and having the capability, at the end of 72 hours, to proceed to cold shutdown.

Id..

The distinction between the capability to repair in 72 hours and the capability to achieve cold shutdown in 72 hours is distinctly stated in Appendix R. Section III.L speaks consistently of repairing alternate systems and achieving cold shutdown within 72 hours, while section III.G. speaks consistently of simply repairing normal cold shutdown systems within 72 hours. The May 1984 Memorandum is correct in asserting that the Commission, in the face of industry opposition, retained in the final rule the proposed rule's time limits relating to cold shutdown equipment. However, what the Commission retained was not, as the Memorandum seems to think, a single rule applicable to both normal and alternative systems. The proposed rule's requirement that

2Moreover, we are

informed that nearly every, if not every, plant to which Appendix R applies has at least one alternate system in some area of the plant.

8

equipment be able to achieve cold shutdown within 72 hours applied only to alternate cold shutdown systems. The proposed rule stated no time requirement applicable to normal cold shutdown systems. See sections III.G. and III.L of the proposed rule, 45 Fed. Reg. at 36089.

Thus, from the start, the Commission apparently intended to distinguish between the fire protection to be given normal cold shutdown systems, and the fire protection to be given alternative systems. Therefore, in reasserting the final form of this distinction, namely the apparently distinct requirements in the final rule, the proposed interpretation does not contradict the Commission's intention in promulgating Appendix R.

Indeed, if one insists that the Commission intended there to be a single requirement applicable to both normal and alternative shutdown systems, then it must be noted that there is good evidence that the Commission intended that single requirement to be the less stringent one that fire protection for cold shutdown systems ensure that they could be repaired within 72 hours. First, as the staff points out, the proposed rule also required that hot shutdown equipment be able to maintain hot shutdown for 72 hours after a fire, and thus acknowledged that cold shutdown equipment might not be available until 72 hours after a fire. See sections III.L.1 and III.L.3 of the proposed rule, 45 Fed. Reg. at 36089.

Second, the final rule contains an analogous requirement whose open-endedness suggests that all the time limits stated for cold shutdown are merely strong suggestions, and that in every case what actually applies is the rule of reason, "Achieve cold shutdown when you can":

If the capability to achieve and maintain cold shutdown will.not be available because of fire damage, the equipment and systems comprising the means to achieve and maintain the hot standby or hot shutdown condition shall be capable of maintaining such conditions until cold shutdown can be achieved.

Appendix R, sec. III.L.4.

Last, the final version of the supposedly more stringent rule, that alternative equipment be able to achieve cold shutdown within 72 hours, is paraphrased thus in a comment accompanying the final rule: "Fire damage to cold shutdown capability is limited to damage that can be repaired within 72 hours to provide a margin in achieving cold shutdown." 45 Fed. Reg. at 76607 (emphasis added).