NUREG/CR-1682, Submits Observations of RIL-123, Performance of Electrical Terminal Blocks Exposed to LOCA or Steam Environ & Related rept,NUREG/CR-1682
| ML20039D108 | |
| Person / Time | |
|---|---|
| Issue date: | 12/03/1981 |
| From: | Joseph Austin NRC OFFICE OF THE EXECUTIVE DIRECTOR FOR OPERATIONS (EDO) |
| To: | Minogue R NRC OFFICE OF NUCLEAR REGULATORY RESEARCH (RES) |
| Shared Package | |
| ML20039D100 | List: |
| References | |
| RTR-NUREG-CR-1682 NUDOCS 8112310324 | |
| Download: ML20039D108 (3) | |
Text
- - - - - - - -
. [f* ** %q'o, UMITED sVAVES
( ~j NUCLEAR REGULATORY COMMISSION E.
W ASHINGTON. D. C. 20555 h
f *E December 3, 1981 MEMORANDUM FOR:
Robert B. Minogue, Director Office of Nuclear Regulatory Research FROM:
J'ohn H. Austin, Technical Assistant Office of the Executive Director for Operations
SUBJECT:
RIL 123 - ELECTRICAL TERMINAL BLOCKS As you may know, I have had numerous conversations with your staff and a Sandia representative regarding RIL-123, " Performance of Electrical Terminal Blocks Exposed to a LOCA or Steam Environment," and the related Sandia report, SAhD 80-1957 (NUREG/CR-1682). The discussions have centered on whether the data in the Sandia report support the recommendations and on whether the method used to calculate failure probabilities is correct. As a result of these. interactions, your staff has requested one of its consultants to review the analyses and interpretations in the Sandia report and the RIL. Having read the RES response to Commissioner Ahearne's memorandum of September 28, 1981, I would like to offer my observations on the subject.
Sandia measured surface leakage currents on terminal blocks and then calculated a failure probability based on a simple Arrhenius equation.
Sandia says the surface breakdown current, I, can be characterized by B
IB IB0 exp (- AE/kT)
(1) and that the probability of failure is given by (2)
K-I/IB P
=
. here K is an arbitrary constant (1.4 was used), I is the measured surface w
leakage current, and AE is the activation energy for surface breakdown.
Experi-mentally, Sandia measured I at 186*F and 110 F for: a) clean terminal blocks and for dusty tenninal blocks at 100% relative humidity, and b) clean terminal blocks exposed to steam containing electrolytic solutions. The terminal blocks examined were either unprotected or were enclosed in a protective box with a 6 mm weephole.
Sandia made no measurements at LOCA condit. ions (325' F). Rather, Sandia and the RIL extrapolate the low temperature measurements to LOCA conditions through the model of Equation (2).
Coimercial data obtained in 1978 in pass / fail type experiments at around 325 were added to the reported results but that data apparently has no bearing on the model in Equation (2).
(That commercial data may contradict the model; see page 18 of the Sandia report.)
RIL-123 states that the Sandia mo_ del has been validated by the Sandia experiments.
Is the model, Equation (1) and (2), valid? That model assumes a singular acti-vation energy for surface breakdown, independent of variables such as specific Nkh$k E
R. B. Minogue December 3, 1981 electrolyte or its concentration. The Sandia report points out that the low voltage surface breakdown phenomenon is very complex involving many variables.
Is there any reference in the literature indicating Equations (1) and (2) are valid or generally accepted for characterizing surface breakdown?
Sandia apparently determined the value of AE in Equation (1) from a 1947 report j
that compared theoretical surface conductance to a measured value at room 5
temperature; that ratio was reportedly found to be 104 to 10.
From that ratio, Sandga calculates the activation energy for surface breakdown to be 0.3 eV (10- = exp ( aE/kT)) and used this in the probability calculations.
Is this a valid determination of an activation energy?
Does the data support the recommendation to clean terminal blocks in operating reactors? Figure 17, page 50 of the Sandia report, contains the main results shedding light on this recommendation. That figure, for unprotected blocks only, and at 186 F, ranks " impurities" according to their effect on surface leakage current. There are only three categories of tests contained in that figure:
clean blocks, dusty olocks, and blocks exposed to steam containing electrolytic solutions to simulate accident conditions. The data indicate that dust caused on the average a 20% to 30% increase in leakage current over clean blocks (the leakage currents of dusty blocks are actually within the experimental scatter of clean blocks) while electrolytic solutions in steam (accident conditions) can increas~e leakage current by a factor of 3 or so. Thus, if the Sandia model is valid (i.e., failure probability is proportional to leakage current), clean-ing terminal blocks would, based on that data, seem to result in little improve-ment.
Is this interpretation of the data correct?
Does the data support the recommendation that protective box "weepholes should be eliminated, ' decreased in diameter (e.g., by flow retarders), or equipped with some kind of bimetallic closing device which automatically activates at high temperature"? The relevant Sandia data are in Figure 16, page 50, and Table 3, page 52. Those data indicate there was only one failure (under unspecified test conditions) in 112 experiments at 186 F, dust again had little effect on leaka.ge current, and electrolytic solutions increased surface leakage by a factor of 2 or so. These data are for a 6 mm weephole. The commercial data, limited as it is, indicate larger weepholes may not present a problem even at 325* F.
Is it valid to perform measurements at one degree of isolation and then extrapolate that data to total isolation? How does RES reconcile the difference between Sandia data at 186 F and conmercial data at 325 F for various weephole sizes?
Apparent omission of Sandia data from the presentation of overall results pre-sented in RIL-123 Figure 1? Table 1 of RIL-123 tabulates the Sandia breakdown data at low temperatures and 1978 commercial breakdown data at high temperatures.
Those data indicate that, at 110*F, Sandia observed 2 breakdowns plus 4_ multiple breakdowns for open terminal blocks. The summary figure for breakdown proba-bil.ity at 110 F in Figure 1 of the RIL apparently shows only 2 breakdowns. Were the 4 multiple breakdowns omitted from Figure l-of RIL-123 and, if so, what was the justification?
If all observed breakdowns at 110 F were included in that Figure, how would Figure-1 of the RIL be changed?
d R. B. Minogue December 3,1981 As an editorial coment, RIL-123 leaves the impression that Sandia experimentally determined the reported high failure rates for terminal blocks under LOCA con-ditions of 325 F (see, for example, page 6, paragraph 1, on "results of the experiments"). Actually, Sandia performed no experiments at such LOCA conditions.
Figure 1 of RIL-123 indicates the reported experimental probabilities were determined by the'Sandia model given in Equation 2.
That is not so.
- Finally, Figure 1 contains error bars. How were those error bars determined, and do they-apply over all temperatures in that Figure?
John H. Austin
~
Technical Assistant Office of the Executive Director for-Operations o
6 e
9
.