ML20214A014

Safety Evaluation Re Util Evaluation of Chlorine Release Into Primary Coolant Sys.Evaluation Provides Useful Info That Could Be Used to Reduce chloride-induced Corrosion of Reactor Components

ML20214A014
Person / Time
Site:	Fort Saint Vrain
Issue date:	11/13/1986
From:	Office of Nuclear Reactor Regulation
To:
Shared Package
ML20214A003	List: ML20214A001 ML20214A014
References
TAC-57248, NUDOCS 8611190155
Download: ML20214A014 (5)

Text

'

j g pracg[g UNITED STATES NUCLEAR REGULATORY COMMISSION g p,

rj WASHINGTON, D. C. 20555

\...../ SAFETY EVALUATION BY THE OFFICE OF NUCLEAR REACTOR REGULATION RELATING TO THE EVALUATION OF CHLORINE RELEASE INTO THE PRIMARY COOLANT SYSTEM PUBLIC SERVICE COMPANY OF COLORADO FORT ST. VRAIN NUCLEAR GENERATING STATION DOCKET NO. 50-267

1.0 INTRODUCTION

By letters dated January 31, March 18 and 26,1985 and July 9, 1986, Public Service Company of Colorado (PSC) submitted for our review their evaluation of the role played by chlorine in the primary coolant system with respect to corrosion of different safety related components. The evalua-tion was performed for PSC by a contractor. The PSC submittals consist of two parts. In Part I, the licensee evaluated different sources of chlorine and estimated the amount of chlorine ingress into the primary circuit. In Part II, the licensee evaluated corrosion effects produced by this chlorine. This safety evaluation covers Part I of the licensee's evaluation including the January 31 and March 18, 1985 submittals and a portion of the July 9, 1986 submittal. Our safety evaluation on Part II will be provided in the near future.

The licensee's contractor performed an analytical and experimental study in order to determine possible sources of chlorine. The contractor deter-mined experimentally that the water incoming to the primary circuit did not contain chlorine and, therefore, the sources of chlorine existed somewher0 within the primary coolant system. By making careful screening analyses, the contractor was able to determine the major sources of chlorine as well as the mechanisms by which this chlorine was introduced into the primary coolant (helium).

Three mechanisms of chlorine transfer into the primary coolant system were identified. These were: 1) water leaching of chlorides, 2) releases of hydrogen chloride (HC1) to dry helium on heating, and 3) release of hcl to moist helium on heating. Ths contractor has established that the majority of chlorine was introduced as hcl on heating different chlorine-containing materials in contact with dry helium.

In order to obtain quantitative estimates of the amount of hcl present, the contractor performed tests and from the results of these tests was able to estimate not only the amounts but the approximate time of release of hcl as well. Two types of tests were performed: water leaching tests and heating in contact with dry and moist helium. Also, the contractor measured concentrations of Cl-36 radioisotope in the chloride deposits on different components of the primary coolant system. From the results of 8611190155 861113 7 DR ADOCK 0500

1 these tests, the contractor predicted that during the total time of plant operation, 300g of chlorine were released into the primary system. Of this amount, 200g of chlorine were released at the beginning of plant operation, mainly as hcl from the heated core and graphite moderator. An additional 50g were released after reloads from new fuel, and finally, 50g were released by hydrolysis of chlorides by moist helium.

It should be recognized that the above estimates are very approximate since they were extrapolated from the test data to full scale plant conditions. Therefore, they cannot be expected to possess a high degree of accuracy. Despite these limitations, the information is still very useful in determining the environment to which different components in the primary coolant system were exposed and thereby permitting a finding as to the probable cause of their corrosion.

2.0 EVALUATION An estimate of the amounts, sources, and times of chlorine ingress into the primary coolant system during plant operation was required as a first step in performing an evaluation of the corrosion of safety related components at Fort St. Vrain. The licensee's contractor approached this issue by performing analytical and experimental studies with the objective of providing answers to the following questions:

(1) What were the sources of chlorine ingress?

(2) What were the mechanisms by which this chlorine entered the primary coolant system?

(3) What was the amount of the chlorine ingress?

(4) At what times during plant operation did different amounts of chlorine enter into the plant's primary coolant system?

In order to answer these questions, the contractor performed several tests. The first test consisted of analyzing incoming water for chlorides. The results of this test indicated that there were no chlorine compounds in the incoming water, and hence the sources of chlorine were within the primary coolant system or within the systems remaining in direct communication with it. The possible candidates for chlorine-containing materials were: concrete, graphite, fuel, ceramic insulation and titanium sponge. Chlorine-containing compounds could be released from these sources in three different ways:

(1 As chlorides by water leaching; (2 As hcl by heating in contact with dry helium; (3 As hcl by hydrolysis of chloride salts in contact with moist helium.

To determine the magnitude of water leaching release, the contractor performed a test in which crushed materials were boiled in water for two hours. Except for concrete, other materials released only relatively small amounts of chlorine. Although concrete constituted a major source of leached chlorides, other materials like graphite or ceramic insulation,

1 1

.. 1 i

i i

1 because of their large masses, could also become significant contributors.

Because leached chlorine was in the form of chlorides which are nonvola-tile and therefore not readily transportable by the circulating helium, the contractor implicitly considered this mechanism as having a very minor effect on the overall chlorine ingress. We consider this judgment to be somewhat speculative, although lack of volatility of the leached chlorides makes this a logical assumption.

A more important mechanism for chlorine release is heating of certain chlorine-containing materials in contact with helium. The importance of this mechanism is due to the fact that in this case, chlorine is released in the form of volatile hcl which can be easily transported by the circulat-ing helium and deposited on various components resulting in their corrosion.

The contractor performed a series of experiments with different materials which remained in contact with hot helium in the plant. The experiments consisted of heating these materials to temperatures between 370 C and 950"C and simultaneously passing helium over them. The released hcl was collected and its amount measured. The experiments were performed by initially passing dry helium over a sample and then, after a certain time, passing helium saturated with water vapor. The results of these tests indicated that a considerable amount of hcl could be released by this mechanism. Dry helium removed the hcl contained in the material tested.

In general, this release occurred at the beginning of the heating period with only very small amounts released afterwards. When moist helium was passed over the sample, chloride salts were hydrolized and additional hcl was released. This release was also relatively small when compared to the initial release. The contractor extrapolated these results to the condi-tions existing in the Fort St. Vrain plant and determined that about 100g of hcl would be released from the core and an additional 100g from the graphite moderator. Titanium sponge would release only a very insignifi-cant amount of hcl, somewhere between 1 and 3g. Later in the life of the

plant, an additional 50g would be released from the new fuel after reload

, startups and 50g after moisture ingress. The total amount of hcl released l

would be slightly over 300g, resulting roughly in 300g of chlorine. In reviewing this part of the analysis, we found that the hcl release in these l experiments reflected fairly accurately the general nature of the release I mechanisms existing in the plant. However, application of the data by simple extrapolation to the plant conditions at best could only produce very approximate results. This is due to the fact that in the actual plant, release of hcl was controlled by much more complex mechanisms. The reported figures for chlorine release should, therefore, be treated as a very gross approximation which was obtained without spending a large amount of time and effort. It is, however, still a very useful body of information for evaluating chloride-induced corrosion in the plant.

In addition to the quantitative studies of chlorine release, the contractor performed a study to determine the approximate times at which different amounts of chlorine were released. Information was obtained by measuring the concentrations of radioactive isotope Cl-36 present in chloride deposits

1 1

on different components of the primary coolant system. Cl-36 is formed by neutron activation of Cl-35, and for a given neutron flux its concentration is proportional to the time of exposure. The contractor measured the concentration of Cl-36 (expressed as ACi Cl-36/g Cl) on the following compo-nents: control rod drive cable wire, moisture monitor valve, circulator duct, circulator bolts, plateout probe, HPS knockout pot and reserve shutdown system balls. The measured concentrations of Cl-36 on these components confirmed that the major source of chlorine was the active core

and that the chlorine was not uniformly released throughout the life of the plant. Rather, most of it was released upon initial plant heat-up with much smaller a

nounts released afterwards. Low concentration of Cl-36 isotope on the plateout probe compared with other components indicated that the greatest release of hcl occurred during the later part of cycle 3 when the probe was already removed and the reactor was operating with high moisture content in the helium coolant. This confirmed the previous findings based on experimental data and gave the approximate time when the release of hcl due to moisture ingress occurred. We believe that this information is very important in evaluating chloride-induced corrosion of the primary system components.

Measurement of activated species were also carried out for the deposits on the failed control rod drive (CRD) cable. Examination of these deposits indicated that the concentrations of Cl-36 and Cs-137 were much higher near the failure point than at the upper elevations of the cable. Two possible explanations of this finding were offered by the contractor. When the rods were fully out of the core, the part of cable which failed was exposed to helium while its upper portion was protected by the CRD mechanism housing.

High Cl-36 content on the portion which failed would indicate that the rod was in the out-of-core position when the contamination occurred and the failure must have occurred at a time late in the plant life when highly activated chlorine was released to the primary coolant. In their first explanation, the contractor attempted to correlate this observation with the rod insertion sequence, but in our opinion, the correlation was not very convincing. The contractors second explanation was that during rod insertion, contaminated moisture condensed on the cable and dripped down its length until it reached a temperature where it evaporated leaving the deposit. Since moisture ingress to the helium coolant occurred relatively late in the plant life, and high concentrations of Cl-36 were observed, this second hypothesis appears much more plausible.

3.0 CONCLUSION

S Based on the considerations discussed above, we conclude that the licensee has provided a thorough evaluation of the available data utilizing suitably designed experiments, and a generally satisfactory explanation of the sources and mechanisms for the chlorine ingression in the primary coolant system. However, since most of the quantitative information was obtained by extrapolating the test results to full scale plant conditions, the l

1 t limitations of the results should be recognized because of the large approximations. However, despite this shortcoming, the methods used provided useful information, and we consider the reported work to be a significant contribution toward understanding and eventual correction of the chloride-induced corrosion problems at the Fort St. Vrain plant.

Principal Contributor: K. Parczewski Dated: November 13, 1986 i

3 J

, - - - . , , . ,-.m- , - - . . - , . . . -, - . , , , , , _ _ _ _ _ . . _ _ _ _ , _ . - , _ . . , _ _ - - . . _ - , . , . - , _ _ _ . . _

_ , , , - _ . _ . _ _ . _