ML20062H293

Forwards Requested Summary of Telcon

ML20062H293
Person / Time
Site:	Crane
Issue date:	04/24/1979
From:	Schwarz H BROOKHAVEN NATIONAL LABORATORY
To:	Budnitz R Office of Nuclear Reactor Regulation
References
NUDOCS 7908230008
Download: ML20062H293 (4)

Text

{{#Wiki_filter:. ll f\\/ i.M_ g,yp i } ( ~ BROOKHAVEN NATIONAL LABORATORY (( (( { ASSOCIATED UNIVERSmES. INC. Upfon. New York 11973 Decenmt cf Cer. (516) 345-4330j .Y. / April 24, 1979 g. ,p M 6 D-5&O r Dr. Robert J. Budnitz f Office of Nuclear Regulatory Research \\ Nuclear Refulatory Cor.. issien

• ' ashing:rn D. C.

20335 =

Dear Beb:

I a. en: losing the su:.ary of our telephone conversations and further th ughts on :5e accident which you requested. Sincerely, ./ / s c. i. w Harold A. Schwar: Enclosure MAS:bv 790 83 30 cc51

O Radiation Chemistry of the Three Mile River Accident 3 The bubble formed during the accident contained 1000 f t of hydrogen at 1000. psi at around 400*T. This gas was above (and assumed to be nearly in equilibrium with) 10000 ft of cooling water at nearly neutral pH, about \\ of which was in the reactor. The water was being irradiated by dissolved fission products and by T-rays from inside the fuel rods. The question was whether or not hydrogen and/or oxygen would continue to be formed by water radiolysis. Nearly all the radiation was absorbed in the water phase, consequently it was only necessary to consider the cooling water itself. The concentra-tien of hydrogen in the cooling water was about 0.046 soles per liter, based on the pressure over the solution. In addition, the water might contain I fission product iodine at a concea: ration as high as 0.1 ppa and several other fission products at comparable concentration. Part of the Zirconium cladding was oxidized to 2:0 which is essentially insoluble in water at 2 neutral pH. Terric and ferrous iron could conceivably be present from reaction with the core vessel, but ferric solubility is rather limited, probably about 1 ppe, and anyway there should have been no more iron present than is there during normal operating conditions. It was assumed' that there was no copper present, other than fission product amounts. Consequently it would appear that the total concentration of dissolved species which were very reactive towards the free radicals produced in water radiclysis was about I ppa, certainly no more than 10 ppa. () Water radiolysis by fission products produces molecular hydrogen and hydrogen peroxide in email quantities and larger amounts of the hydroxyl radical, OH, and reducing radicals. Consider first the molecular hydrogen and the hydroxyl radical. They are formed in yields of 0.4 and 2.8 molecules per 100 ev, respectively. The hydroxyl radical reacts with hydrogen OH + H H+HO 2 2 and this is the only reaction which needs consideration in, determining whacher or not more hydrogen will be produced. If 1/7 or more of the OH radicals react this way, then no more hydrogen will be produced. e 4 e --w-

2 The rate coefficient for this reaction is 9.2 x 10' exp -1620 ~ k = 0 ~I -I or 3.0 x 10 M see at 200 C (400 T). The maximum possible rate 10 constant (diffusion-limited) for any OH reaction would be about 3 x 10 at this temperature. The rate of the reaction is the product of the rate coefficient and the concentrations of the reactants [k z C(OH) x C(H )l' 2 so the fraction of OH radicals reacting with hydrogen is at least 8 3 x 10 x 0.0t.6 8 10 3 x 10 x 0.046 + 3 x 10 C where C is the concentration in soles per liter of reactacts with maximum e possible rate constants. This fraction is greater than 1/7 if C is less than 2.8 x 10" M. For an atomic veight of 100, this would be 280 ppa. With less than 10 ppu of reactive impurity present, essentially all of the hydroxyl radicals reacted with hydrogen to produce water and hydrogen Thus all radicals produced by radiolysis vill and up as reducing atoms. radicals, and whether they are all hydrogen atens or half of them are hydrated electrons (as initially produced) is immaterial. Both react with oxygen at diffusion-limited rates. Oxygen could be produced by decomposi-tien of the hydrogen peroxide with a yield of abcut 0.3 molecules per 100 ev and a steady state concentration is determined by reaction with H atems. The total radical yield is 6 radicals per 100 ev. Four hydrogen atoms are required to reduce one oxygen molecule to water, so the steady state of 0 vill be reached when 202 of the H atoms (or electrons) react 2 with 0. Impurities which would compete with the oxygen are ferrie ions, 2 and many oxidized fission products. As before, the sum of their concentra-tions was at most a few parts per million in the reactor, so the steady-state of 0 should have been about 1 ppm or less. This was about 10 of 2 the hydrogen present in the water and since the distributions of 0 and H 2 2 between solution and gas are nearly the same, there should have been less than one part in 10 of 0 in the hydrogen gas at steady-state. 2 If significant 0 was present initially it should have been reduced 2 to water with a yield of one molecule disappearing for every four radicals O o

g produced, or with a yield of -1.5 molecules per 100 ev. Based on a dose rase of 6 x 10 rads / hour given for Sunday, April 1 (which seems 3 very low to me) about 1.8 peunds of 0 Per hour, or 0.5 ft at 1000 pai, 2 would disappear, until the steady-state level was reached. In the above analysis which has the benefit of checking some literature references, one figure is different from those given you over the phone, namely that the OH + H reacti n pr coeds at 1% the maximum 2 theoretical rate at 400 F, not 10% as stated. I did not remember the correct value for the activation energy. On the other hand, a safety factor or 10 was included, so the answer, that one or two hundred parts per million of oxidizable impurities would be required to prevent the ( back reaction was the same. _ The above analysis is based on what is actually dissolved in tha water, but the radicals could possibly diffuse to the surface of a solid and reaet. Most of the solid is Zr0 and so the hydroxyl radicals would 2 find nothing to react with. The hydrogen atoms could diffuse up to 0.01 cm in their lifetime (though impurities probably limit this distance to ~ about 10 em). One might expect some reduction of 2r0, perhaps to a 2 hydride of some sort, in which case hydrogen could slowly disappear. Most of the parameters used in this analysis are accurately (210*) known in the' region of 0 C to 100 C. Extrapolation is required for 200 C, but should involve little error, certainly no worse than a factor of two. f Lack of knowledge about impurity levels is the most doubtful part. [ I \\.. i 1 l l I l i

8 l i y7 1pg i f I M:MORANDLM FOR: File ~ FROM: Saul Levine. Dirsetor Office of Nuclear Regulatory Research

SUBJECT:

PRINCIPAL CDNTktC MADE WITH EXTERNAL ORGANIZATICN3 DURING TMI 2 ACCIDENT 1 l 1. I spoke with Robert Ritznann of Science Applications concoming H2 and 1 02 generation rates in TMI 2 vessel in the period March 31 - April 1. He informed me, that although an increase without considerations of a reforention rate of H2 and 02 due to bubble back pressure, that the 1 percent rate was probably too high. He also said that he felt the rate was probably no higher tMn 0.1 percent per day and could be zero. but that he did not have the data to calculate an explicit rate. 2. I spoke with James Proctor of the Naval surface Weapons Center about the effects of a Hydrogen explosion on vessel integrity. He said the cylindrical position of the vessel would be subjected to about 6 percent strain, which should not break it, and that it would also g he subjected to a lifting force of about 1.5 I 105 lbs.. He could not. calculate whether the main loop piping could hold the vesgel dann when subjected to this foece. since he did'not have 'deta11e8 info'riation on plant layout. ~ i OrHr.al signed SJ Saul Lavinef A Saul Levine. Director Office of Nuclear Regulatory Research Dist: Subj Ci nc. Chron 1/ RBudnitz rdg

f.,, v'...

Stevine rdg fv j t L h \\ { RES:0 ..S.k Ay /.hC.._ o. c. 6M../. 7 9........ . ~....... '....... -... .m

• j h 30. 38 He M 3 EROS ate.

8u- -- . ewess e esse e age m,...,',........ i i s . L a -,,..._--_,,,.m_-__m- ,---,-_,--_.-.._._m-m. ~

/ / ~ s

1. peag%

q UNITED STATES {' l  ; NUCLEAR REGULATORY COMMISSION { E wAsmwcTom. o. c.2oses k,.'.v / JUP F 1979 MEMORANDUM FOR: Saul Levine, Director Office of Nuclear Regulatory Research FROM: Robert J. Budnitz, Deputy Director Office of Nuclear Regulatory Research

SUBJECT:

INFORMATION ABOUT INDIVIDUALS AND ORGANIZATIONS CONTACTED BY ME DURING FIRST FEW DAYS AFTER THE TMI ACCIDENT ( I am responding to the memorandum of May 31, 1979, from E. X. Cornell, " Request for Information from Presidential Comission." I have gone over my log book for that period, and have found seven outside individuals with whom I had substantive contact. In each case, my contact was the only or the pri'aary NRC contact. Besides these individuals, there is a large number with whcm I spoke but for whom the primary contact was you or T. Murley. I assume that you and he are assembling your own lists, similar to mine, and that you will cover those other individuals. For each individual, I will indicate their organizational affiliation, address and telephone number, as well as a brief description of what infonnation was furnished. 1. Dr. Richard L. Gamin (I.B.M., Yorktown Heights, NY 10598,(914) 945-2555). On Saturday morning, March 31. I was called at home by Dr. Garwin, an old friend, and he provided a number of ideas to me about things that one might attempt to do to eliminate or reduce the pressure from hydrogen within the primary system of the TMI reactor. His ideas included putting a snake-like tube into the vessel, and using chemical means to combine hydrogen with other s ubs tances. He also gave me some insight into how important the back reaction is in calculating the shock pressure in a fast burn or detonation of hydrogen in a vessel like the TMI reactor vessel. He referred me to Dr. Harry Petschek of AVC0 (see below) for es:istance on the hydrogen combustion problem. Later that date, and again on \\, Sunday, April 1, I talked with Dr. Gamin by telephone, to follow y up on his understanding of pressure shock waves, something about which he had extensive advice. 2. Dr. Harry Petschek (AVC0 Everett Research Laboratory, Everett, MA \\ 02149, (617) 389-3000). On Dr. Garwin's suggestion. I called Dr. Petschek on March 31, finally reaching him at home in late ,.O .~ V r

l Saul Levine 2 morning. He responded imediately by indicating that he and some ~ colleagues could assist in understanding the issue of hydrogen combustibility and combustion kinetics in a reactor vessel such as at TMI. Later that day and through Sunday, April 1. I spoke, two or three times, to Dr. Petschek and one or two of his colleagues. They worked on the questions of what concentration of oxygen in pure hydrogen would be the threshold for combustion, particularly at the temperatures and pressures thought to be present at TMI (about 1000 psi at many hundreds of degrees F), and he reported back sometime Sunday on those. Dr. Petschek also referred me to Dr. Bernard Lewis in Pittsburgh, who turned out to be a highly-regarded expert in just these same issues. 3. Dr. Bernard Lewis (Combustion and Explosives Research, Inc.,1016 Oliver Building, P1 ttsburgh, PA 15222, (412) 391-3633). I finally reached Dr. Lewis, on referral from Dr. Petschek, on Sunday morning, April 1. He acknowledged expertise on the combustibility of hydrogen and oxygen; indeed, he is the coauthor of the definitive textbook i on this subject. He and an assistant, reached at home on Sunday morning, worked through that day and part of Monday, April 2, and gave important advice on the issues that governed the physical behavior of hydrogen and oxygen burning in conditions such as were thought to exist at IMI. He gave infonnation about the mixture of oxygen in pure hydrogen that would be a combustion threshold, talked at length to me about the physical difference between combustion and explosion, and what would be the impact of gaseous impurities, t He reported back his pre 11minary conclusions sometime after midday on Sunday, April 1, and his final conclusions in midmorning of + Monday, April 2. He calculated pressure ratios (pressure within a j fast burning situation vs. starting pressure), detonation thresholds, i heat release, flame temperatures, and other parameters. His insight l; was valuable in providing a perspective on which parameters were, and which were not, important in modifying the result of what was j calculated using approximations. 4 Dr. Harold A. Schwarz (Brookhaven National Laboratory, Upton, NY 11973, FT5 Tel. 666-4330). Dr. Schwar was referred to us by Dr. H. J. Xouts of BNL, who called several times during the TMI incident to provide advice. Dr. Schwar: worked much of the weekend of March 31 and April 1 on calculating the production and recombination rates of oxygen in the TMI orimary coolant water. He did these calculations at home mostly, I think; telephone contacts with him during the weekend were at his home. He reported on the considerations that were involved in his calculations, and showed definitively that oxygen generation from radiolysis would not result in much oxygen in the gas phase, because of the recombination reaction with the assumed large hydrogen gas overpressure and the associated dissolved hydrogen. Se were apprised of the preliminary results of Dr. Schwarz' work early on the morning of April 1, in my memory, but it was not firmed up until sometime shortly after midday on that day. Dr. Schwarz continued with his work for several days after Sunday, April 1, and filed a description of his calculation with NRC on April 24. A

1 1 Saul Levine 3 t l 5. Dr. Heinz Heinemann (Lawrence Berkeley Laboratory, University of California, Bernaley, CA 94720, (415) 486-6000). I telephoned Or. Heinemann in early morning of Saturday, March 31 to follow up a suggestion of Dr. Garwin that the oil companies might have expertise in snake-like methods for extracting hydrogen from a pressure vessel like the TMI reactor vessel. Dr. Heinemann is a chemical engineer at my former laboratory in Berkeley and is a colleague and frieno there, who spent most of his life working for Mobil Oil Corporation. Dr. Heinemann referred me to Dr. J. Penick of Mobil, whom I called subsequently. Dr. Heinemann also discussed with me the question of addition of catalytic chemical agents to reduce the hydrogen in water solution. Dr. Heinemann gave me the names of several catalysis chemists who might have expertise in this matter, and also enlisted in advice of two Berkeley colleagues. We talked several times over the weekend of March 31-April 1, but I turned over the entire problem of cherrical hydrogen removal to others in MRC, and did not concern myself with the issue directly. 6. Mr. Joseoh E. penick (Mobil Oil Corporation,150 E. 42 Street, New York, NY 10017, (212) 628-9757). I contacted Mr. Penick on Satuday morning, March 31, on referral from Dr. Heinemann. He said that he thought Mobil could assist NRC with advice on the availability of snake-like devices to extract gas from a TMI-like pressure vessel. He called back later during the weekend (I recall his return contact as occurring on Sunday, April 1) and indicated that devicas such as we sought were not readily available in the Mobil Corporation, and unlikely to be available elsewhere in the I petroleum industry. The problem was that the path into the reactor vessel from the outside to the upper dome was too tortuous for the use of the devices that did exist, and the fabrication of a special device would be quite difficult. 7. Dr. Laura Cherubini (17 Pandover Road, Billerica, MA 01821, (617) i 667-9699. Dr. Cherubini called me on her own on Saturday, March 31, l with a suggestion of chemical means to reduce or eliminate hydrogen dissolved in the reactor coolant water. I do not know how Dr. Cherubini i received a reference to me. The method was to use algae that trap i hydrogen from solution by presence of free electron acceptors. Since I was not expert in this matter I turned it over to others at NRC for follow-up. However, by the time anything more could be done with this ( suggestion, the perception of the impcrtance of a " hydrogen bubble" had diminished, and I think that no further follow-up occurred. f. Robert J. Budnitz, Deputy Director Office of Nuclear Regulatory Research cc: T. Murley e e -,~ ,g w mv-- -e --,,,e---,-,,n ---g-- e -w n-- --r

i f %g umTuo :TATES y NUCLEAR HEGULATORY COMMISslON Jk *) wAssmorow,o c.moess o Y."... / JUN 131979 MEMORANDUM FOR: The Files FROM: Thomas E. Murley, Director Division of Reactor Safety Research

SUBJECT:

RECORD OF ACTIONS, THREE MILE ISLAND ACCIDENT MARCH 28 - April 6 This memo records my major activities and lists the individuals with whom I had substantive contact during the Three Mile Island Accident and its immediate aftermath, March 28 - April 6,1979. On Wednesday, March 28, I learned about 9:30 a.m. that there had been an accident at TMI. Not much was known of the details except that there were high radiation levels in the plant and a general emergency had been declared by the utility. On Thursday morning I attended a br.iefing of the Commission by NRR staff where I gained the impression that the situation was generally under control, although the high radioactivity levels in the plant clearly indicated extensive fuel damage. On Friday. March 30, while on annual leave, I received a call from Saul Levine at 4:00 p.m. asking me to come to Tony Buhl's office to get some tests started to help resolve some problems at TMI. On the way to the office I heard a bulletin on the car radio that NRC had announced there was a danger of core melting at the Three Mile Island Nuclear Plant. The following activities are listed by topic and are generally in chronological order. Removal of the Hydrocen Bubble Upon arrival in Tony Buhl's office I was informed that measurements at the site indicated there was a noncondensible hydrogen gas bubble in. the reactor vessel having a volume of 1000-1500 cu. ft. at 1000 psi and 280*F. There was some concern that the bubble was growing and might lead to uncovering the core and potentially to fuel melting. One option being considered was to open the relief valve on the pressurizer ~ and try to vent the hydrogen bubble out the valve to me containment building. It was recognized this would be a tricky maneuver since it would mean that the one pump operating at the time would have to be 'e' shut off and there was no assurance that it or any of the other three pumps could be restarted if they were subsequently needed.

• W

\\.

s e The Files. At about 5:30 p.m., we called INEL (Larry Ybarrondo, Nick Kaufman, Hank Ziele and others), described the problem to them and asked if there were meaningful tests that could be done in the Semiscale facility to help decide whether and how to vent the hydrogen bubble. We described the TMI primary system layout and gave them the important plant dimen-sions and elevations. Later in the evening on Friday, they called back with the following information and suggestions: A test could be run in Semiscale using nitrogen gas to simulate the hydrogen venting maneuver in TMI. They reconnended against such a maneuver in TMI, suggesting it was better to keep the plant in its then stable operating mode (this suggestion was relsyed to the NRR personnel in the Incident Response Center). After working all night to set up the Semiscale facility, the INEL staff ran a test on early Saturday morning (Enclosure 1). The results showed that the Semiscale system could be depressurized by opening the pressurizer relief valve and turning all pumps off but about half of the nitrogen gas in the bubble remained in the primary system -- princi-pally in the steam generators. The electrically heated rods remained cooled during this maneuver. This information was passed on to the IRC and to B&W. During the remainder of Saturday and Sunday, Wayne Lanning of the RES staff worked with B&W engineers and INEL 'to establish conditions for a second test using a larger helium gas bubble and injecting coolant from the High Pressure Injection System. This second test was run early t, Monday morning, April 2 (see Enclosure 2). The Semiscale tests provided the following general information: They showed that the noncendensible gas in the bubble would not all vent out the relief valve -- in fact about half of the gas would remain in the primary coolant system. They showed that it would be difficult if not impossible to remove the gas from the primary system by subsequently restarting the primary pumps. Since the pruece of gas in the high points of the steam generators could prevent natural circulation ecoling, this information was a strong argument for leaving the gas bubble in the top of the vessel. They provided useful data for S&W in establishing the appropriate apis flow rate for-their proposed emergency procedure in the event all primary pumps were lost.

JUN 13 I"79 The Files. Oxycen production Rates On Saturday, March 31, we received a question from the staff at the IRC whether there could be sufficient oxygen gas in the primary system to form an explosive mixture and thereby constitute a threat to the reactor pressure vessel. The answer to this question proved to be elusive. I discussed this question with staff members from INEL (Sid Cohen, Ron Ayers and Jack Liebenthal). Concurrently, Saul Levine called Bob Rit:mann of Science Applications Inc., and we understood that Bob Tedesco of NRR was contacting staff at KApl. The information I received from INEL was based on reported data from the Cooper plant (a BWR) and was scaled down to the power level of 25 MWt. Their conclusions, which \\ they stressed were extremely conservative, were that the hydrogen bubble contained about 2.2% oxygen and that it would take at least 4 to 5 more days to reach 5% oxygen concentration. I was later given some data from the Advanced Test Reactor (ATR) that was purported to support the data from the Cooper plant. I found it very difficult to piece together all of the information into a consistent story. The Cooper BWR data were not directly applicable to TMI (a PWR) although there was some boiling in the TMI core. Similarly, the ATR is a low pressure (150 psi) reactor and was also not directly applicable to TMI. Late on Saturday evening I received a call from Rob Rit:mann who reported that he was not having much luck in calculating the oxygen concentration, although he believed it was below the flamability limit. Some time after midnight on Sunday morning, I went to the Incident Response Center where Roger Mattson asked what we were finding. I told him that the picture on oxygen concentration was confused, but that a conservative estimate seemed to be that tne cxygen concentration in the hydrogen bubble was increasing at the rate of 1% per day after reactor scram. Later that morning (around 9:00 a.m.) Roger Mattson met with Saul Levine, Bob Budnit: and ne at the IRC prior to leaving for the TNI site. Chairman Hendrie, Comissioner Gilinsky and Comissioner Kennedy came and went throughout this short meeting as I recall. Mattson sumarized the following information as the distillation of all of the input he had received: Flamability limit ? 5% 02 in pure H2 Detonation limit 2 12% 02 in pure H2 Comoustion limit 218%02inHgsteam 02 production rate x 1% 02 per day in H2 bubble Current 02 conc'entration :s St 02 in H2 bubble

The Files - After this meeting, I spent little further time on the oxygen concentration question. ~ Potential Pressures from Hydrogen Explosion On Sunday morning someone suggested that I collect information on what pressures could be generated if there were a hydrogen explosion in the pressure vessel. I found that Vince Noonan of NRR was the focus in NRC for these analyses and I therefore was involved only peripherally. I received infonnation that Dr. Nonnan Slag of Picatinny Arsenal had made calculations showing a sharp peak pressure of 12,600 psi for the case of a detonation of a 1000 cu. ft. bubble containing a mixture of 83% H > 2 12% 02 and 5% steam. This pressure appeared to be consistent with infornation received by Bob Budnitz from Dr. Bernard Lewis of Pittsburgh that pressures could reach 5 Po for deflagration and 13-14 Po for detonation (where Po = 1000 psi is the pressure of the bubble and the system initially). This information was passed on to Vince Noonan and I had little further involvement after Sunday afternoon. -l Hydrogen Gas Behavior From the beginning of my involvement in the TMI accident one of the tasks being investigated. principally by Bob Budnitz, was how to eliminate the hydrogen bubble by mechanical or chemical means. During Saturday and Sunday I received a number of unsolicited suggestions from people calling in-to NRC. One of the most novel suggestions came frem N Roger Billings, President of Billings Energy Co. in Provo, Utah. Billings specializes in hydrogen research and technology,and they had done work for DOE under a contract from 00E-Idaho. His suggestion was to inject into the primary coolant a large number of 2-4 micron diameter evacuated glass microspheres. The idea was that the hydrogen in solution would diffuse easily through the glass and be trapped in the spheres, thereby gradually deplenishing the hydrogen in the systein. Billings claimed to have experience with these microspheres. I felt the idea was worth a look, so I authorized INEl. on Sunday night to have Billings start an experimental program on hydrogen behavior. The Billings staff began work on Sunday night and ran their first hydrogen solubility tests on Monday morning. Although the microsphere idea proved to be not feasible, Billings did some very good work over the follow-ing three weeks that proved very helpful in understanding hydrogen te-havior in TMI (see Ecnlosure 3 for a complete report of their work). , +. - N c

= - ^ ~ ~ JUN I 31979 The Files, I Some of the conclusions reached frem hheir work were the following: Adding nitrogen gas to TMI coolant to try to form ammonia (NH3 and remove hydrogen would not work (reaction was too slow Adding microspheres to trap hydrogen gas would not work (takes weeks) Adding hydrogen peroxide (H 0 ) as a source of oxygen to combine 22 with hydrogen in solution would not work (02 gas evolved within minutes) Measured H2 solubility at TMI conditions Measured H2 evolution rate if TMI were to depressurize provided a means for detemining the pressure at which H2 saturation was reached if TMI were to depressurize (pressure reboundeffect) During the week of April 2-6, Billings was put in direct touch with the NRC team at the TMI site, and they continued to provide information to NRC during the following weeks. Hydrogen Degassing On Monday, April 2 I was asked to calculate how long it would take to remove hydrogen from the coolant by degassing through the letdown system ~. ( and through the pressurizer spray system. I called Glen Jenks of ORNL for infomation on hydrogen removal through pressurizers, and he pro-vided me with some useful data and several references to look up. My calculations showed that it would take.from 1 to 2 weeks to reduce the hydrogen in solution from 1600 sec/kg (saturated concentration at 850 psi and 280*F) to 300 sec/kg (saturated concentration at 300 psi and 140*F). These calculations also showed that it would take nearly a week to reduce the hydrogen bubble by degassing through the letdown system. The only way the hydrogen bubble could have been reduced from 1000 cu. ft. to zero in 2-3 days is by assuming high flow rates through the pressurizer spray line (15 gpm) and a high efficiency for hydrogen removal in the pressurizer (90%). I am skeptical that these flows and efficiencies were attained in TMI. I have not heard a convincing story of how the 1000 cu. ft. hydrogen bubble was reduced to nothing in 2-3 days. h

The Files JUN 13 579 Measurement of Water Level in pressurizer On April 5, I received a call from Tony Buhl to help with one of the instrument problems he was working on. The problem concerned what would happen if all the water level indicators were lost in TMI. Buhl was considering whether the resistance temperature device could be used as a backup level indicator by running a high current thrcugh the RTD and observing the change in temperature when the water level drops below the RTD in the pressuri:er. I called Hank Ziele at INEL and put him in touch with Bob Shepard of ORNL to get the details of the RTD design. The INEL staff then ran a scoping test in an autoclave and determined that the technique could be used in principle y (- to indicate water level in the pressurizer. I subsequently learned that there were practical problems at TM1 that made this approach not feasible. Follow-on Work Since April 6. my staff and I have been involved in extensive analyses of the TMI accident. These activities are documented in the formal transmittals from RES to EDO. Thomas E. Murley, ' .to r Division of Reactor Safety Research

Enclosures:

( l. "Semiscale Pressurizer Relief Valve Venting from Three Mile Island Type Conditions" 3/31/79 2. "Second Semiscale Relief Yalve Venting test from TMI Conditions" 4/2/79 3. "TMI Reactor Simulation Final Report," Billings Energy Corp., 4/20/79 cc: S. Levine R. Budnit: K. Cornell J. Cummings

'~ ~ ~ } ^ 0: W. D. LANNING ( L L = f PRELIMINARY TEST RESULT 3 SEMISCALE PRES $URI2ER RELIEF VALVE VENTING FR TkREE MILE ISLAND TYPE. CONDITIONS r b s %S 9 + J ,5 March 31, 1979 L g ,,,, r p "; r "-' ""- ~ ~ A-i.,6y$L.< l ?' /pAU . c, 4.. 1 4 3 g. g .g:pv : .., ; ' 3 .. A.- y;. < Y, ~ -Qa s.- y-l l , C. " 7.- l. ~~ ()- _ =. .,y

~ -\\ -1 = =:= ' ~ ~ L _ m._.,- g i ( h l A. SIN %Rf On March 30, 1979 NRC sanagement personnel requestad EGAG Idaho ' personnel to help evaluate alternative coursas of action for securing the Three-Mila Island Plant (TMI). We conducted our evalu several recoausendations, we proposed conductfag a venting test o i primary relief valva (PRV) in Samiscale from present TMI conditions to check the accuracy of calculations we performed on the responsa of TMI to such a venting condttion. We conducted the proposed test from 8:55 a.m. to 9:47 a.m. en Merch 31. 1979. Two-hundred forty channels of data were recorded. The test was successful. We believe thL test results may ha of use to MAC , in evaluating the probable TMI plant responsa,1f venting from present conditions is attempted. The r=mainder of this report is divided into three sections. Saction 8 presents a ccaparison of TMI and Samiscale significant parameters as best we know them. Section C providas the sequence of experimental events and signf ficant pehnomena correlated with the tfee at whfch they occurred. Section D prasants the calculated TMI plant responsa dur1ag venting from the 9RV from the inttfa1 conditions provided by MC. Section E presents our conclus!ons from pressurizar rettet valve tests in Semiscala. k _ -n.wm.c. wAggg.

_.. ~ ~ g e .8. .(ON,PARISON OF SEMISCALE AND THREE MILE ISLAND (TMI) $1GNIFICANT PARAMETERS i (Vol vesselVo1 esmoonent ) 1. VOLUME RAT 105 8&W 5emiscale $$/8W Pressurizar 0.374 0.454 1.21 Cold Leg (one side) versus broken loop 0.118 0.255 2.14 e liot teg (one side) versus broken loop 0.122 0.144 1.18 9[ Total Loop (beth sides) 0.974 1.596 1.839 F 2. ELEVATIONS (from $ Nozzle HL) 88W Samtscale Top of upper plenwn 14 ft - 5 in. 13 ft 1/2 in. Top of core g ft - 0 in. -5 ft - 0 in. Surge line connection to liL 6 ft 1/2 in. 4 ft - 0 in. Y surge line vertical drop (not from nozzle () 12 ft - 8 in. Il ft - 0 in. Piping vertical height dtotal) (including pump suctionj 58 ft - 0 in. 51.3 ft (SL) 21.3 ft (IL) ~ [ Nuzzle to top of tube (or pipe) 46 ft - 0 in. 41 ft 11 ft surge line belcw top of cara 3 ft. J-1/2 in. 2 ft - 0 in.

w o'- e 3. $1GNIFICANT DITTERENCI5 l i SSW 5ectseele Cold Les 2 cold legs / sfde 1 cold leg / side (1 side scaled for 3/4 flow) Not Leg 1 per side 1 per side .. Upper Plenus Vent Valves No Vent Yalves Not Leg / Cold Leg Elevation Difference Mone 8-1/2 in. (Hot liigher) SG Elevations (difference) None IL 9 11 ft - 0 in. ~ SL 8 41 ft - 0 in. l l l 1 (

T ' ~ - ~ ~ ~ = ' I,.' ( b l-( C. EXPERIMENTAL SEnUEf!CC AND $1Cff!TICANT EVEllTS A description of the initial condItfons for the THI plant response test and a table of significant events follows. Also included is a description of the significant Semiscale system conffguration or operating conditions that may not be typfcal of the TMI plant. Initial Test Conditions Nitrogen bubble initially established at a level o'f K3 fn, above s. hot leg pipe upper invert (about 0.5 ft ). The elevation nf the 3 bubble is higher than expected for the full scale prant. b. No secondary side water was added to the steam gener'ator. This 4 lack of water would cause a lesser amount of energi to be trans farred [the primary flufd. ( [ Pressurf ter steam dome was estah11shed at 30% of the pres mrfier volme. c. i ?" d. Leak rate of the system was estahltshed since the flow out the sisiulated pressurtzer relief valve flow was of the same ragnitude I 2 as possible leak rate. I System was heated to 410*r. utilizing the core and an initial pressure v. [( of 7.24 kPa was established. L m se=e

• N
• 4*

.,g-T--

I g q ( Int tial Test Condt tions (Contd.) f. Pugs were coasted down and an initial power of 7.64 kW was established. i This power was about 2 kW above the scaled valve to allow for energy ..l losses to the structure. f ( l e

q t

1 t I P -r - -;; -,.__,_ v,.,, _,__ r.

~~ g e ( _ TABLE OF 5!GN!r! CANT EVENTS I Time Event 0 Experiment initiated (pressurtzer steam bubble venting starte Initial clad temperature - 420*K (295'T) Initial core outlet water tenperature - 411*K (280*F) Initial pressurizer pressure - 7.6 MPs (1100 psi) Initial inert gas / water interface at 1.35 m above (53.15 f cold leg centerline Initial heat flux 4.1 kW/m2 (1300 Stu/hr-ft ) ? 0 - 1000 see i Nitrogen bubble in upper head expands downward, pressurfrer water level rises as steam is fonued and vented. Pressure falls, temperatures rise slightly. 9 \\ 5 1000 see Pressurfrer water level reaches vent connection Clad temperature - 435'K (323*F) Core outlet water temperature - 420'K (295'f) Pressurizer pressure - 4.7 MPs (680 psi) Eas / water interface - 0.36 m (14.17 in.) above enld leg center 1tne. 1000 - 3800 see Nitrogen bubble continues downward expanston, pushing water out pressurizer vent connection. Temperatures continue increasing, pressures fall. 3800 see Nitrogen / water Interface resches top of h' ot leg openf ng 0.25 m (9.84 in.) above cold leg cantarlfne clad temperature - 442*K (336*F) Core outlet fluid temperature - 427'1 (309"F) Pressurizar pressure - 2.5 MPa (380 psi) w i - - - ~ ~

.~ g e ( _TA6LE Or $1GNIFICANT EVENTS (Contd.) i r 1 Time Event 3800 - 5000 sec Nitrogen expands into loop pf ping. Steam generator drafns through cold leg and downcomer forefno cool watar fnto lower part of core. Clad temperatures geaarally decrease as well as pressure. 5000 sec Steam generator tubes empty completely. Peak clad temperature - 456'K (361*F) Core outlet flufd temperature - 444*K (339'r) Pressure - 2.2 MPs (319 psi) 5000 - 7000 see Clad temperatures increase as core riew stagnates, flutd temperatures rise as pressure continues to fall. 7000 sec Flutd temperature reaches saturation s.od bult bnfitng r. begins in core. ll( Core outlet fluid temperature - 452*r. (354"r) Li. Peak clad temperature - 456'K ( 361"r ) Pressure - 0.8 NPa (116 pst) k. t 7000 - 9000 see Low vofd fraction fluid riser in core, clad teeperature decrease, fluid temperatures rosstn at naturst fon as pressure falls, 9000 sec Test Shutdown. Nitrogen has expanded into stor.m generator but pressurf rer appears to stt11 be filled with lf quid or very Icw yetd fraction f1std. Pressure - 0.34 MPs (49 psfa) Clad temperature - 426'K (307'F) Fluid temperature - 4:3*K (302*F) Core bulk bofling occurred when saturation condit'fons were resched. Max cled temperature - 456*K (361'F)

g c 8* e' { . In several instances the Semiscale syster co conditions were not typical of the TM1 Plant. instances include: The most significant (1) Potential for structures to provide excessive c i ooling nf the fluid. (2) t Lack of sinviation of the PWR vent valves (3) Core elevation effects. (4) Possible stypicality of power operated reifef ^ valve flow due to effects of sfre. (5) Use of charging punps to account for leakage from (8) Lack of steam generator secondary water, pump seals. The structures in the Semtscale Mod-3 system have area which will cause atypical energy transfer during the excessive surfac e temperature transient. course of a flutd During the simulation of a pressurf 2(r relfef transient the structures wf11 absorb excess energy wh incr6ase the depressurfzation rate and p t f-region. rovide cooler water to the core An atteept was made to provide more typical fluid by increasing the core power level by about 35% conditions m. t above the scale valua whfch was arrived at by determining the rate of energy t rans fer to the vessel structure and increasing the core power appropriately i The Samiscale system cannot provide a particularl of vent valve actuation during a pressurfrer relief d y good simulatinn t - rain test because of elevation difforences between the het and cold legs f i However, thfs inability ches not strongly influence the Seniscale t est results since the influence of the vent valves on the TMI pl 'P Ifmitad due to the cold leg geometry. ant cnid leg behavior is No adverse effects on test results are expected in Semiscale pressurizer relief valve drai of the lack of a vent valve simulation. n testing because . wp -

?- - - - -- A ~ .... ~

.: ~. - - ----

~ -- 4' e' 6 1 s A particularly signirtcant dif ference' ! rtwacn tfic

• e-t 5,

'a '1Iis and the TF11 Plant is related to the location of the. top of the enre relative to the vessel nozzles. In the TM! plant the top of the core is ...c.=1 .s.1r as sh. .t...et.. se sh. v... 1 ..i s,eh. s et.e.1. sy s t spi. however. the top of the core is serrootset 1y los en meneath M*.3010.*U!.';,0**.000.,"y'32.IO*;* 20'4* 74.00'%%"O *,i: I*, the TM1 plant than in the Semiscale Mod-3 systesi. F1mt through the Seniscale pressurizer relief velve sirulation is not expected to completely duplicate, on a scaled basis.. that.which might i occur in the Tit! Plant. Critical flow thrnugh small ori f tr.es (e.g., the 0.030 in, diameter Semiscale flow ares) has been show& to be dif ferent ~ from that experienced in 2-in. disneter pipes, so that vent flow / pressure relief characteristics might reasonably be expected to dif fer somewhat. between the two facilities. j Because of considerable leakage of psep seals (and other miscellaneous small leaks) it was necessary to provide makeup liquid to ti.e systeei. The HPIS punp was run fnr brief periods at fixed intervals thronohnut the test 2 to supply the additional liquid necessary to account for the pump seal leakage rate. Although the makeup rate was small compared ta tha discharge rate through the simulated pressurizer relief velve. te 1"tal j amount of HPTS liquid injected into the systen over the dur ation of tb. test was a substantial menount. Thus, considerable additional subcuni te was added to the primary system liquid inventory through uma n' the WIS As a result the core thernal response pey have been It'..t severe 1 pump. than would have othenvisa' occurred. 11 . % '.A-

-~ g-g- ( e The secondary side of the stesse generator was dry. i.e. no mntive fluid for heat transfer. Thfs condition minimized the influence nf secondary heat trans fer on the course of the PRV trans ient in.temit t el e. If the ' temperature on the secondary side of the steam generator is lower than the primary system temperature the subcooling on the pep suctf on .[ 1eg will be increased, thereby f acreasing the subcooling to the core. () Conversely. if the secondary side temperature is higher than the primary .1 ' side tegerature the pump suction density will be reduced thereby reductno the core inlet subcooling. For a PRV transf ent the seconifary side teeperature should be less than or equal to the primary s fde teeperaturr. Based on previous experirnents conducted in Semtscale, stene generator heat transfer f s not expected to have a signt ficant influence o.1 the pressurfrer rel f ef draf n tests. E w' t IbGf h I t i P \\ ~.-- -. 1

a~~- - - - - - ~

{;

g" g' d ) r D. PRELIMINARY ASSESSMENT OF 7E TMI Pt. ANT RESPONSE DURING VENTINS Analysis i Simp 1f fled calculations were performd to evaluate expected response of TH! to the PRY release mode of depressurfration. First, the hes tup rate of the core fluid was calculated to be 480*F/hr if no circulation occurred (heating the core liquid volume only) and 170'F/hr if the total ' reactor vessel fluid volume were to be heated. (Neitherenlenistion i ~ fncluded fuel or metal mass heet capacity). From this it vos concluded that makeup should be provided to assure core coverage as heat is remved by steemir.g. About 60 gym wet cniculated as the required rate to meet the steam generation needs. The expansion of the- (then assumd) 1500 ft 3 of gas to fill the hot Teg, steen generator, and pressurfrer no reach the l point of gas venting and twee rapid depressurization would reach this point at about 300 psi. At assumed 1tguid relief rate of sno gpm and steam rate of 110,000 #1hr this was esiculated to take about one hour. The samtscale system depressurtzed slower; reaching about 350 psi in one hour wf th a smaller relative gas volume. Integrating the high velocity ' gas' ref fsf showed that the het If qufd in the pressurtzer - flashed to steam and separated yielding a larger total volume of steam to be relieved prior to the time of liquid reitef. The data also showed that the pressurizer and surge Itne renefned liquid full, thus not making that volume available for gas expansion. These features are being added to a more conplex model. Initial indication < are that a reasonable description of the pressure transtant and volume change will result from this podel, and it should be app 1feable to TMI.

~ ~ ~ ~ ^ ~ - ~ S' (' ~ E. Ct*CLUSIONS (1) The sent cale system can be depressurized via the proposed rathod to a level at which the RHR pumps can be activated and used to l ) renove residual heet from the core. 5 Tf (2) In the Semiscale system noncondensible gas did not vent easily or uniformly wf th the proposed method. h The noncondens thie gas buhhie entered the hot leg at approximately 3000 seconds. T bp (3) Core unenvery in the Semiscale facility did not occur until after + ^ s point in ticse at which the RHR pumps could have been activated if desired (thus preventing core uncovery). (4) The Seniscale results suggest that if ECC fluid is in.fected into the systen at a rate comosrahle to that at which the sys/en is being vented, significant benefits in the overs 11 systen response and core cooling may be reaf f red. (5) The heater rods in the sentscale test remaf ned in a mnde of good cooling during the proposed trans font and rod temperature rises were minimal. (6) Depressuriration from 1050 psia to 49 pete was accompitshed in the Semiscale test in approximately 3 hours. Sof1fng in the core did not occur untti approximately 6006 seconds after the vent P relief trans f ant was initf ated. p.

m .= .--. _ = -. ~... w.- TO: W. 3. UW8t!NG, IIRC I ( RECElvED ./ .I, 2 APR i 21 11 21 U.S. NLt*' F *

• FO.AIMY c:: e n u Msat, & MEssa saANCH dii..

PRELIMINAAY TEST AESULT3 i,CDiu SECOND SEMISCALE RELIEF YALVE VENTING TEST FROM THREE MILE ISLAND CONDITIONS I April 2. 1979 / M

• e.'

4 - y .g 9 t,' (, e, C'-- / s [I.< t\\ e 4 0 ~ 'a w_ m N

~ l. ... _. =.... -w --- - ,1 9 i I( A. su m utY Results from the inttial Semiscaf a Mod-3 test conducted at conditions sistlar to the Three Mlle Island (TMI) plant provided afsnificant insight into the growth and movement of an upper pianus bubble during continuous pressurizar reltaf valvs (ptV) operation. On April 1.1979 NAC management requestad a second test to be conducted at condf tfons stattar to the inittet test except ihr a larger gas bubble volume and injection of a scaled asovat of coolant from the MP!5. The test was conducted between 1:18 a.s. and 4:35 a.m. en April 2,1979. The test appears ta be a success although some loss of data was expertenced as a result of a malfunction in the data acquisition system. A prettuinary evaluation or results indicatas the test $ehavior was sfallar to the fattial. j tes t. Injat. tion of coolant from the Mp!5 maintained cooling f a the core region, pressurizar reltaf flow was taminated and recharging of the systen with a high MPIS flow was (af tf atad den the systan pressure reached. about 270 psfa. The systas was returned to a stable stata where the primary coolant pump could be used to circulata flow withf a the primary systen about ena hour aftar taratnation af pressurizar reltaf flew. j e 'b ( s f e

~ =.... .....z_ ,y g e 3 3. INIT!Al. TEST CONDITIONS AND $EQUENCE OF KYENTS A descrfptfon ef the inttf al conditiens for the second TN! plant raspeasa test and a table of sf pf ficant event fo11ews. A table of the f aittal operating condittoms Table T. and a description of significant Samtscele operating conditions are also facluded. A descrfption of the systes conftguration that may not be typical are documented in the report en the initial Semiscale pressurizar re11ef test. INITIAL TEST CONDITIONS a) A Nelfum bubble was initially established at a level of 61 cm above 3 the hot leg pipe upper f avert (equivalent to 0.8 ft ef Nef fus). Nelfun was chosen flor this test sface its properties are relatively close to Nydrogen. The bubble size was chosen to provide a larger bubble volume than used en the inttf a1 test. b) The diameter af the orifice used to simulata the relief valve mes O.031 cm. This diameter was slightly larger then the diameter used in the initial test (0.079 cm) and was adjusted based en resulta from the initial test to provide a steam flow more. typical of the speciffed relief valve flow for the TM! plant. c) The NP!$ system was initiate'd at rupture at a constant rate of 12.4 al/s (0.20 spa). This rate was scaled to an average injection rate uhtch was befag considered for the TN! plant ever a pressure range of 300 to 1000 psf. Additional f ajection was included to make up for the normal leakage from the Semiscale system. d) The steam generatar in the same leap as the pressuriser had a water level aquivalent to 2/3 the total tube elevation. The fattial steam generator flufd temperature was 415 K. The second staam generator was run with a dry secondary side due to difficulty te datarstaine a consistant set of initial conditieas. 3 l

• ??,

Y.~

. - - ~ ~ -~ ~ .i -;.......-~ ~ w-%:..7&,cQ.=}:] - s.. w,; .g a) An eriffca was included in the intact loop het leg to compensate fIsr '6 J.] the differenes in size, and he'nce elevation of the top invert (Figure 1), '- of the fatact and broken loop pipes. This elevation diffeitase mes 4,,.. belfaved to have caused the intact and broken loops to behave differeag)I. fa the initial test. .g.Mk... . a.,. r .c o:~a; '*f f,'. t f) The core power was held constant at 21.5 kW throughout the test. 'This' 51 < is substantially above the 3.9 W 1evel necessary to simulata the kamme ' :- t THI conditions. Th's additional 17.5 W was included to make up fler. ambf ent and structural heat losses which are in ancass of the TMI heat j.- loss es. The Semiscale heat losses were determined ampertsentally us f ag intt1al test condttons. It fs expected thf s excess power would j cause the Semiscale core red and fluid temperatures ta be sanservatively hfgh compared to the TMI plast. . ;_61,;.1.%js.['$.. $I -. :.a ; g) The initial pressurizar staan dame was about 50% of the pressurizer. '. " ; r-ve1une. .- - l h) 1he systan was heated to 416 E utt11aine the core and as fattial pres' urg h a ( ef 7.24 MPs was astah11shed. ,.s.,..j.[ -A.e

• ~' :.. A 1)~

The pumps were coasted down and the initial power was estabitshed. t l 7 ' ::. G.,.~ ? ' ): ~.. .g - !.{, %k .h A ~-: ;;n.. s a 0 C. ! J. o ~ f d.h.'. i

• • .. w... >

..,. 7. ' '.." j.- g

4.

. :.r.. Y. :... -T..,s i' l l ..j. . 3 4 . ;;..... ; ~ I t l ~" _ " ..un-.---" < mm.,, __ w- 'g

4 4 4 ~ ( vant i W Vsssal ha Geseaxu Staa.m & arseede ~ 2.S4"I.D. chifika f \\ (

a. - _ _ _.

'g g.......... Beeken La* ~ $3* 2.D.) IntasT. Lp Het La3 - W a Lag (2.u.h"r n) \\ f / Figure 1 Schematic of the intact and broken loop hot legs. I h I e e l ( i S I "T--

I ~1 g ( j TA8t.E OF SIGNTFTCANT EVENT 3 \\ i Tfe* fvent i O < a ria a= ' it' tad (>

== ria st 6 662 a** = atar i Inf tial peak clad tamperature - 442 K (338 F) Inf tial core outlet watar tamperature - 434 K (325 F) Initial pressurizar pressure - 7.3 MPa (1058 psi) Inf tf al f aart pas / matar faterface at 0.92 m above (38.4 ja.) cold leg canter 11ne Initial care powr - 21.s W s - 200 s Nelfue bubbla in upper head espands downward. pressurizar h-- 2 N 5 .atar ievei ris.s as staa is cor..d. and.vaated. er.saura. g falls, tamperatures risa slightly. 40 s Mallus / tratar laterfaca reaches top of ho) lag opening l,, 1CCg 0.25 m (9.84 in.) above cold tog cantar11ma g Clad temperature - 442 K (338 F) ? Care outist fluid temperature - 438 K (325 F) Pressurizar pressure - 4.0 MPa (870 psi) r 1, 200 - 700 s Nelfue continues to aapand pushing wtar into pressurizar and filling intact and broken loop steam ganarators with 7g r @s Pressurizar watar level reachas vent connection Ciad tamparature - 445 K (378 F) 7oO5 =~ tiet -tar t-earatua - 44s = (37= >> Fressurizar pressure - 4.45 MPa (440 psi) s } a

~ - = - - - ( e \\) TABLE OF $fGN!FTCAMT EVEMTS (Contd.1 e 'ine twent it - 2200 s Nelfus bubble continues downward expansfoa, pushfag water 700 - 2,'2.00 $ out pressurizar vent connection. Temperatur1s centinue l increasing. pressures fall. e '00 - 2800 s Intact and bmkan loop staas generators drata through es1d "2.2 00 - 28d05 les: and downcomer forcing cool water fato lower part of cars. Clad tosperatures generally decrease as well as pressurg 100 - 4800 s Data acquisition system malfunctioned during thf a period. A1). l yfgo6./.l7dC3 data was lost. During this time the staas generators complete) drained but some water remained fa the bottaa of the het leg piping. ( 10 0 s Data acquisition systas back on line. yTdO $ Recseding of data continued. 100 - 5800 s Temperaturas continue to decrease as systas pressura decreasey, '800- 5 400 $ (L 100 - 4700 s Clad temperatures facrease as core flow stagnatas, fluid 57o0-(,74d5 temperatures rfse as pressura continues to fail, l i 100 s Test terminatad when vent valve was closed. MP!3 tajectica rgg [.p]O D I increased to 0.83 spo causing systas pressure facrease. I 100 s Systas pressura peakad at 1080 psfa. At maxfans system prasset 1/ oo S the broken 1 , pump was slowly brought up to speed and tettfai I conditions vers re-estabitshed. Natium was bread out of brotei loop but callected in fatact toep. It appeared impossible to restart and achiev,e full flow from both pumps staca Nettum wouli collect la one pump when the other pump was betag startad. ( preventing the second pump from estahlfshing a set positive ^ suction head. 7 l 3

.,a. .t TABLE Or $ffm!F1thiT fyDTS (Contd.) e Time front 12.000 s Test shutdown ir \\ t I i t i t 4 pp Y. ?. n v I L b s b I 6 ' P r F T ~ ~ m ny_ g

-~ - - - - - ~ { t .s ( TABLE I LIST OF INITIAL CONDITIONS CON 0!TIoss FOR SEMISCALI TEST St!! pa! MARY SYSTEM l 7.29 Wa (1058. psta) Pressure Tamparature - 418. K (289.4 F) i, j pRf35URf2fR pressure - 7.29 W a (1054. psfa) Tamperatura - 550 K (548.8 F) Top 400 K (260.8 F) Sottos \\ Level - 505 Heaters - Turned off after pressurizar condttions established. t ( STEAM afNERATORS y Intact Loop Laval - 0. (Dry SeconJary) Irokan Loop l Laval - 1/3 - 2/3 full l Temperatura - 418 K (250 F) l CORE Poutt 3.9 W s Decay Neat 1 17.5 W Heat Lasses e l Total 21.5 W 3

~ .... _, L _,. _ ~.... g 3 TABLE I - List OF INITIAL CONDITICMS (Centd.) UPPER PLENUM / UPPER MEAD Mel1 m Vol me - 1.67x104 + 567 os8 8 (0.59 f,0.02 ft ) 3 3 Water volume above hot leg nozzles - 3400 cm (0.16 ft ) NIGM PRI55URC !NJECT!ON PUMP Location - Intact Loop Cold Les Rate (constant) 12.5 al/ (0.20 gon) CORE FLUID TEMPERATURE DISTRIBUTION TOP - 438 K (325.4 F) ( 430 K (314.5 F) MIDDLE 420 K (296.8 F) SOTTOM LOOP COOLAMT PUMP 5 / \\. Standby condtt1ons (power off) ,e ( P 10 1

f 14 ( ) ( C. CONCLIJSIONS

1) The' additional gas volume and the enfarsument of the size of the staied relfef valve area influenced systas' response somewhat but did set change the general conclusion that the systes could be depressurized to a level dere the 15(R system could be activated without signf ficantly shoctfag the synten.

2) The use of the HP!S pump during the early part of the asperiment caused cold water to be supp11ad to the care. This additional sold water maintafnad high density fluid in the core and contr%ted to sustafaed

( i effective eselfag of the rods.

3) Based sa a compartson of subcooled flow rates the simulated reifef alve flow fe Semiscale is 70 to 90 percent of the subcooled flow calculated for the 1HI plant valyes. This lower flow any be due to the presence of noncondensable gasas or to ort ffce size.

The possibf11ty ef reductog ( the subcaoled relief valve flow when noncondensable gasas are present should be considered in calculationa performad for the TMI valves.

4) The need fbr higher core powers to make up for additiona1* system heat lesses resultad in higher rod:tanseratures and fluid temperatures la Semiscafe than would be expected ta the TMI plant. These higher

(~ temperatures would influence the depressurization rata late is the tast when the pressure reached saturstton levels.

8) After completion of the system depressurizatten attampts were made to repressurize the system and re-establiah tattial operating candittens with both seafscale pumps aparattag.

It was discovered that when either one of the pumps was started gas was ibecad through the loops fate the other loop pump and the rensining pump csuld est be startad and maintafae( at full flow becausa a net positive secties head could not be established. Therefore, it appears that onca Melium is present la the samtscale loop piptag it is un11kely that the gas can be removed from the systes by ( opers' fag both pumps'sfrultaneously. 11 _,_...-w .p__ .-m_ ,y_.-.-_ c.

..} ~~ - v-- g s 9 8) In Semiscale Test 3MI1. differencas between the intact and broken loop pfpe sfzas allowed the expandtag Nelfun bubble to reach the top fevert of the larger intact loop pipe early in the transtant. Thfa produced preferential dispersion of the gas into the intact loop. which m felt to be atyptcal of a pWA. In Test 2!2 an erifice was added to the intact loop hot leg nozzle to produce aquel top fevert efevations la the two het legs. This geometry change produced more saf fbre gas dispersion into the het less and an increased venting of the gas out the pressurizar which is bs11wed to be more typfcal of the TMI plant. 7) 5entscale results are dertaitaly fafluenced by such scaling distortions as geometric size, sna-dtsenstenalf ty, structural heat transfer area. and elevation influences. Caution should be exercised in the fater. pretation and extrapotation of.these casalta D any other size facility. b e 9 e \\ 12 ~~' ~ ' ~'.. '1--~- v m. e - - - m

• 9+e e

0 BILLINGS ENERGY CORPORATION THREE MILE ISLAND REACTOR SIMULATION FINAL REPORT e ~ Billin9s

• 'L'

"**T=*1

• • • -"ag

= a:

.._s-EnerOY nVOrDOration i

, L/ (V u

t STATI?.ENT CF ~40P3 RI? ORT TERIE MT ISI.AND REACTOR SI.WEICM FINM RI?CRT Contract No. R-2265 Schmitted to Depart =ent of Inerrf Nuclear Regulatory Cem=ission W111ste Building Silver Spring, Maryland 20900 Schmitted by Billings Energy Corporation P.O. Box 555 Prevo, Utah 84601 \\ Research Iffer: Spenscred by U.S. Depa-ent cf Enerry Nuclea: Regula:c:f cc. issien IG&G Idaho, Inc. 't ~ Md 5 M-ML.U L Re - a'.d ;cn::all Accer I. 3a hnes 71. residenc of Rese=- '- 2-Esident ~ ~ Agil 20, 1979 e -___.w,,e " ~ ~ ~ '

' esse ACKNOWLEDGE:iEll?5 Personnel frem Billings Energy and Ccmputer Corporations who participated in this research effort are: Roger E. Billings David C. M. Eyerly Reginald Wintrell Darrel Anderson Dr. Ronald L. Woolley Don Larsen Hal W. Tucker Mike Lee Dr. Jack H. Ruckman Douglas Furr V:1ughn R. Anderson Lewis Billings G. Lyn Kimball Dan Garfield Harold N. Simons George Walter: Craig M. Huntsman Wayne Rice Paul C. H/rstncrsh I Personnel from EG&G, Idaho, Inc., who participated are: John L. Liebenthal ~ Arnold L. Ayers, Jr. Personnel frcm Brigham Young University who participated are: Dr. Angus U. Blackham Dr. Calvin H. Bartholomew I Art Uken O e + b i O W D I ~

.~ a TABLE OF CONTENTS Page List of Figures iv List of Tables. . vii 1.0 Executive Sammary 1 2.0 Introduction. 6 Statement of Work 9 3.0 Theoretical Evaluation of Hydrogen Solubility in Water. 15 3.1 Objective. 15 ) 3.2 Results. 15 4.0 Solubility Tests / Pressure Drop Tests, Continuous. 20 4.1 Objective. 20 4.2 Introduction 20 4.3 Assembly of Bench Test Apparatus and Data Acquisition System. .20 4.4 Experimental Procedure 20 4.5 Test Results 26 - 4.6 Discussion 29 i 5.0 Solubility Tests /Depressurization as a Function \\.- of Time. 31 5.1 Objective. 31 5.2 Introduction 31 5.3 Experimental Apparatus 31 5.4 Experimental Procedure 31 5.5 Test Results 34 L. 5.6 Discussion 37 6.0 Catalytic Systems 38 6.1 Catalysis of the Hydregen-Oxygen Reaction by the Reactor Vessel. 38 6.1.1 Objective 38 6.1.2 Observations Regarding the Possibility of Reaction of Hydregen in the Reactor Vessel 38 ^-

m l Table of Contents Page Two ) 4 Pace 6.2 Literature Search for other Catalytic i Systems 39 6.2.1 Objective 39 6.2.2 Discussion of Literature Search Findings. 39 6.3 Material Balance and Thermodynamic Study 44 6.4 Catalytic Reduction of Oxygen with Hydrogen under Water 45 6.4.1 Objective 45 6.4.2 Catalytic Reduction of Oxygen with Hydrogen under Water 45 6.4.3 Nickel Catalyzed Reactions. 47 6.4.4 Results 47 6.5 Appendix 49 7.0 Microshpere Test 51 i 7.1 Objective. 51 7.2 Introduction 51

7. 3-Experimental Apparatus 52 7.4 Experimental Procedure 52 7.5 Results.

54 ( 7.6 Discussion 54 8.0 Reactor Simulator Tests 57 8.1 Objective 57 8.2 Introduction 57 I 8.3 Experimental Apparatus and Data Acquisition System. 57 8.4 Discussicn - Reactor Simulation Tests 58 8.4.1 Experimental Procedure. 58 8.4.2 Test A 69 8.4.3 Test B 71 8.4.4 Test C. 75 8.4.5 Test D 75 9.4.6 Test E 75 - ii - e e cw - -. ~ v

---

c-,

-r

) i I i Table of Contents Page Three Pace 8.4.7 Test F. 75 ~ 8.4.8 Test G. 78 E.4.9 Test E. 80 8.4.10 Test I. 80 8.5 Interrelationship of Tests A, B, E, F, G, and I 84 8.6 observation During Set-up and Procedure. 88 ~ 8.7 Appendix 90 9.0 Hydrogsn Peroxide Tests 97 9.l' objective. 97 9.7 Introduction 97 9.I Experimental Apparatus 97 9.4 Experimental Procedure 100 9.4.1 Test A. 100 3.4.2 Test B 101 9.5 Results 102 9. 6. Discussion 102 9.6.1 Test A. 102 9.6.2 Test B 102 k.._ - f u I 4 O 4 iii e

./. LIST OF FIGURES Pace Figure 1 A Schematic of the Nuclear Reactor System at the Three Mile Island Nuclear Station. 7 Figure 2 Henry's Law Coefficient for Hydrogen in Water 17 Figure 3 Theoretical Plot of Hydrogen Solubility in Water as a Function of Temperature and Pressure 18 Figure 4 Theoretical Data Plot of Hydrogen Solubility in Water. 19 !\\ Figure 5 Experimental Setup to Model Hydrogen Solubility in Water. 21 - Figure 6 Schematic of Data Acquisition System for the solubility Tests 22^ Fi ".re 7 Photograph of Test Apparatus. 7 .- 23 ~ l Figure 8 Photograph of Test Apparatus with Instrumentation Equipment 24 Figure 9 Plot of Bubble solume as a Function of Pressure 26 Figure 10 Modified Reactor Experimental Apparatus 32 I( / Figure 11 Pressure-Time Trace Showing Changes and Subsequent Pressure Recovery as Hydrogen Comes Out of Solution. 34 Figure 12 Interpretation of Data in Figure 11 36 i Figure 13 Pressure-Time Plot for the Catalytic Reduction of Oxygen with Hydrogen Under Water on a Platinum Catalyst 46 Figure 14 Microsphere Test Apparatus 53 Figure 15 Schematic of the Simulated Reactor Apparatus 60 Figure 16 Labeled Photograph of the Simulated Reactor Apparatus 61 l - iv - ( I

I' List of Figures Page Two Page Figure 17 Photograph of the Simulated Reactor Including Instrumentation. 65 Figure 18 Photograph of the Simulated Reactor System Under. Construction. 66 Figure 19 Photograph of the Simulated Reactor High Pressure Pump 67 Figure 20 Photograph of the Instrumentation for the Simulated Reactor System 68 t Figure 21 Depressurization/ Hydrogen Bubble Growth as a Function of Time -- Test A 70 Figure 22 Depressurization/ Hydrogen Bubble Growth as a Function of Time -- Test B 72 Figure 23 Relationship of Pressure Rebound vs. Target Pressure for Water Saturated with Hydrogen. 74 Figure 24 Depressurization/ Hydrogen Bubble Growth as a Function of Time -- Test D 76 Figure 25 Depressurization/ Hydrogen Bubble Growth as a Function of Time -- Test E 77 f (j Figure 26 Depressurization/ Hydrogen Bubble

j,/

Growth as a Function of Time -- Test F 79 !l, Figure 27 Differential Pressure as a Function of Temperature -- Test F 81 Figure 28 Differential Pressure as a Funcrion of Temperature and Time -- Test I 32 Figure 29 Hydrogen Bubble Volume as a Function of Pressure -- Test I. 83 Figure 30 Comprehensive Plot of Hydrogen Bubble - Growth as a Function of Vessel Pressure -- Tests A, 3, E, F 95 Figure 31 Ccmposite Plot on Semi Log Paper of Subble Growuh.as a Funcrion of Pressure -- Tests A, 3, E, F, I 86 6

List of Figures Page Three Page Figure 32 Calibration of Pressure Transducer. 91 Figure 33 Hydrogen Peroxide Test Apparatus. 98 Figure 34 Hydrogen Peroxide Test Instrumentation. 99 J 1 1. t L 4 - vi -

1 LIST OF TABLES Page Table I Theoretical Results of Hydrogen Solubility in Water 15 Table II Theoretical Calculations of Hydrogen Solubility in Water Based on Henry's Law 16 Table III Hydrogen Solubility Data -- Continuous Discharge 27 Table IV Hydrogen Solubility Data -- Fast ,7 Discharge with Intermediate Stops 28 \\ Table V Numerical Values for Step Changes in [ Pressure as Recorded by Computer. 35 Table VI Advantages and Disadvantages of Four Different Catalyst Systems. 41 Table VII Catalysts Prepared by NaSH Reduction. 42 4 Table VIII Results of Activity Tests for Borchydrida Reduced Ni. 48 Table IX Pressure Data for Microsphere Test No. 2 55 Table X Pressure vs. Time for Hydrogen Peroxide Test No. 1. 104 -l Table XI Pressure vs. Time for Hydrogen Peroxide Test No. 2 105 i i L 1 s l l l l

1.0 EXECUTIVE SU!cOJtY Billings Energy Corporation was contracted by the Nuclear Regulatory Commission to run various tests to provide information regarding (1) methods of predicting the maximum amount of hydrogen in the water circulating in the damaged reactor at Three Mile Island, Pennsylvania; (2) methods of scavenging gaseous hydrogen from the reactor system; and (3) the deter- .inatien of the most efficient and also the safest means of depressurization. This section contains a brief summary of the results j-of the experimental effort launched by scientists and engineers Of Billings Energy Corporation to help solve the problem with t-a nuclear reactor at Three Mile Island, define depressurization i procedures, and indicate possible alternatives to aid in the shut-down of the reactor. Solubility of Hydrogen in Water (Theoretical Evaluation). _. l / Calculations were made to determine the maximum amount of l (.. gaseous hydrogan that could theoretically be dissolved in water at various temperatures and pressures. These computations gave an indication of the potential hydrogen bubble grcwth in the reactor during depressurization. It was estimated that the maximum bubble growth in the Three Mile Island Reactor would be 1037 cubic feet at 300 psia (9%.of the total volume), assuming the water in the reactor was cc=pletely saturated with hydrogen gas. . ~ r,- r-. - +, -. -

l Solubility / Pressure Drop Tests An experimental test apparatus was developed on a bench scale to investigate the solubility of hydrogen in water under' conditions anslagous to the Three Mile Island Reactor. Depressurizations of the test unit were performed for two cases: (1) continuous discharge, and (2) fast discharge with intermediate stops. Although the results were higher than the theoretical calculations (a bubble growth of 1781 cubic feet, 15.5% of total, and 1884 cubic feet, 16.4% of total, for the Continuous t Discharge Test and the Fast Discharge with Intermediate Stops i Test, respectively), the bench scale tests esitablished a. base. i-for the Reactor Simulation Tests, as well as aiding in th'e definition of the experimental testing procedura. Solubility Tests - Depressurization vs. Time Using the bench scale test apparatus, the hydrogen bubble grcwth during depressurization was experimentally determined as a function of time at a constant temperature. The results ,\\ of this test showed that a large pressure rebound could be an indicatien of saturation. t Catalytic System Several different schemes were censidered for catalytically causing the gaseous hydrogen in the reactor to react with oxygen to form water and thus reduce the pressure in the reactor vessel as well as reduce the hydrogen gas volume. l (

The first consideration was to analyze the materials in the reactor vessel itself and look at the possibility of these materials catalyzing a reaction between the gaseous hydrogen and oxygen. It was concluded that, although possible, it would be highly unlikely that significant reaction would take place. ~ A literature search identified some subitances that could possibly be used to catalyze this reaction. This survey indicated that hydrogen forms complexes such as ReH, HCo(CN)5 ^^d ~ EPtSr[P (C H ) 3 2 F ur cther materials that would possibly 3 25 ( , catalyze the reaction of hydrogen with oxygen are: (1) colloidal , dispersion of sodium borohydride reduced nickel (or platinum); (2) a finely ground alumina-supported nickel (or platinum); (3) a hemogeneous Co (CN) 3-complex; and (4) catalyst coated glass microspheres. To define the identified catalysts more cdequately, f, several experiments were conducted. ~ of the several catalyst systems considered, only the catalytic reduction of oxygen with hydrogen underwater on a ' ' ~ ~ - platinum catalyst, and the colloidal nickel boride system showed positive experimental results. Microsphere Test Another alternative that was considered which might reduce the amount of hydrogen in the reactor vessel was the possibility of introducing glass microspheres to the circulating water system. The specifications that were received en the m -. -.. ~. - ~,.. - - -,,

~ . ~ - - microspheres indicated hhat they would have a hydrogen scavenging effect. The experiments on the bench scale test apparatus indicated that there was a 10.6% decrease in pressure. The results did indicate that although the microsphere scavenging effect would not be an Omnediate solution to depressurization, the long ta=n ef*ects of the microspheres would have a significant result on the pressure of the system. Reactor Simulation Tests A pilot plant unit was constructed to simulate the reactor ( at Three Mile Island. The simulated reactor system consisted of the following major components: the steam generator, the high pressure circulation pump, the reactor vessel, and the pressurizer vessel. The pilot plant was fully' instrumented to allow the monitoring of temperatures, pressures, flow rates, and volumes. The purposes of the reactor simulator tests were: (1) to determine the effects of pressure and temperature reductions ( upon the behavior of the reactor system filled with water containing various a=ounts of hydrogen up to and including being saturated; (2) to obtain a ecdel of these characteristics so that the degree of hydrogen saturation and bubble size might be ascertained through pressure and temperature and (3) to determine the mest efficient and safest method for cold shut-down of a nuclear reactor system believed to contain a hydrogen water solution. -4 w

The various tests that were conducted defined the most critical condition which could exist for the reactor vessel at various tamperatures and pressures. These results are plotted in graph form in Section 8.0 of this report. It was concluded that the most efficient and safest procedure for cold shut-dcwn was first to reduce the temperature of the reactor systam. As shewn in Figure 3 in the text, as the temperature is decreased there is simultaneously a decrease in the amount of hydrogen in solution, or, in other ( words, hydrogen evolves from the water. Since unsaturated water is added to the system during this period the actual amount of gaseous hydrogen dces not increase significantly. Two phencmena cccur during this period: (1) hydrogen is ccming out of solution because of the decrease in tamperature; and (2) hydrogen is going into solution because of the addition of unsaturated makeup water. After the temperature of the system has been decreased, the pressure can then be reduced. }- Hydrocen Peroxide Test s An experimental apparatus was set up to evaluate the feasibility of introducing hydrogen perexide into the reactor system which, upon decemposition, might react wd'" *ke hydrogen i I and thus reduce the pressure.cf the system. The experiment conducted indicated that the decc=pesition l of th.e peroxide takes place at a much = ore rapid rate than the combining of the oxygen and hydregen in the system. The results showed a net ine,rease in pressure rather than a de-crease in pressure. 6 m ~

u.u. _T ~

2. 0 INTRODUCTION As a consequence of the failure of a nuclear power reactor at Three Mile Island near Harrisburg, Pennsylvania, technological expertise was sought to assist in defining and solving the problem of safely cooling the reactor as a necessary step prior to shut-down, clean-up, and repair.

Based on the premise that a bubble of hydrogen existed above the water in the reactor, Billings Energy Corporation i was asked by the Nuclear Regulatory Commission to investigate u-means of removing the hydrogen so that the reactor could-be. safely cooled and depressurized. Information supplied to Billings Energy Corporation. hy the Nuclear Regulatory Commission relative to.the damaged i 4 ~ reactor indicated that the reactor contained approximataly f..' 12r000 cubic feet of highly radioactive and slightly acidic... water, which was circulating at the rate of 95,000 gallons )/ per minute. Apparently some debris (presumably frem the damaged tN' core) was also circulating with the water. The pressure and temperature inside the reactor were 1000 psig and 280 F, respectively. Various esti=ates of the size (as of April 1, 1979) of the bubble above the liquid in the reactor ranged frem 350 to 1500 cubic feet with some indications that the bubble was slowly decreasing in size. It was presumed that the hydrogen bubble resulted from a reaction of circalloy and water at the elevated temperatures l 1.

. ~.. N U C:1.3U U E N ( h m_@ @s_ t _g i i a y W 9 l 7 o e ( e eJ k ~ [tSillirm1 1. Relief valve _2. Reac.cr vessel j 3. Stea= Generater 4. Landcwn Cccler 5. Pressuricer 6. Cen ain=en: Vessel 7. Injecticn Syste= Figure 1. A Sche =atic of the 5cclear Reac::: Syste= z.: the Threa 'e d ' = sland Ncclear Sta icn. W 6

which are thought to have been produced when a portion of the core was exposed. It was considered possible that the bubble contained not only hydrogen but also oxygen, and there was concern as to whether or not there existed or could exist an explosive gaseous mixture during depressurization. It was also possible that the bubble contained no hydrogen. The reactor vessel was inaccessable, because of high radiation levels in the containment area. Compounded with ( inaccessability of the reactor was the lack of or unreliable reactor -instrumentation. Consequently, there was some diversity. of opinion as to the nature of the problem and the po'ssible solution. A quick solution was necessary to minimize the danger to the surrounding area and inhabitants. Resolution of this problem may prevent similar problems from recurring or, at least, may provide a quicker solution should the same problem occur elsewhere. b' Figure 1 shows a schematic diagram of the nuclear reactor and containment vessel. A bubble of a gaseous hydrogen / oxygen mixture in a reactor chamber presents two major problems: (1) If the void space at the top of the reactor increases sufficiently to expose the reactor core to the gasecus mixture, cooling will be lost to the exposed portion of the core, resuIting in possible damage to the core with possible release of radioactive material into the ccoling water; (2) The second major -a.

l concern with a hydrogen / oxygen mixture above the liquid in a reactor chamber is the potential for an explosion of the hydrogen / oxygen mixture. To delineate the problem and provide possible solutions, the following statement of work was drawn up: Task 1. Calculations of Hydrogen Solubility Using known data determine the solubility of hydrogen in water at conditiens similar to the reacto'r's present I operation and proposed depressuri=ation program. Task 2. Laboratorv Test Unit Develop a laboratory bench scale test unit to-investigate the solubility of hydrogen in water under conditions analagous to the Three Mile Reactor i.e. 1000 psig to 300 psig at 280 F. Task 3. Hydrogen Solubility Continuous Discharge Determine, using the bench scale test unit, the hydrogen solubility in water for a range of pressures ( from 1000 psig to 300 psig at a constant temperature of 280 F (138 C). ~~ Task 4. Hydrogen Solubility - Fast Discharge with Intermediate Stoos Determine, using the bench scale test unit, the hydrogen solubility in water for a range of pressures frcm 1000 psig to 300 psig at a constant temperature of 230 F. Task 5 Decressuri : tion and Bubble Grewth vs Time Determine, using the bench scale test unit, the hydrogen bubble growth during depressurizatien a: constan: temperature.

.=

x...

Tank 6. Catalytic Systems 1. Investigate catalysis of the hydrogen-oxygen reaction by stainless steel. 2. Undertake a literature search for other possible catalytic systems. 3. Assess using material balances and a thermo-dynamic study, the present reactor conditions. 4.. Investigate the catalytic reduction of oxygen with hydrogen under water using a platinum catalyst. 5. Investigate the catalytic reduction of oxygen with hydrogen under water using nickel based catalysts. Task 7-Microseheres Investigate, using microspheres, the possibility of scavenging hydrogen in the reactor. Task 8. Pilot Plant Unit Develop a ~ pilot plant unit to simulate the reactor-k-- pressurizer-steam generator systems for operation under pressures and temperatures analagous to the Three Mile Island Nuclear Plant. Install all equipment necessary to measure and record full data requirements and to simulate present reactor conditions and standby cooling conditions; i.e., Pressure, Temperature, Flow Indication, etc. Task 9. Simulation Tests Using the pilot plant undertake a series of tests to simulate reactor conditions during the depressuri:stien of the reactor-steam generarion system. . e ,,w _ _ - -,., _,. ~.. _..

~.. a-Test A. Hydrogenate system to simulate saturation l at a pressure of 300 psig. Lower reactor pressure frem 1000 psig to 300 psig by discharging water medium maintaining constant temperature. During discharge lower system pressure by increments of about 100 psi and correlate discharge pressure, bubble volume (water volume extraction) and time. Test B. Hydrogenate system to simulate saturation

s at 1000 psig.

Lower reactor pressure from 1000 psig I to 300 psig by discharging water medium maintaining constant temperature. During discharge icwer system pressure at increments of about 100 psi. Plot and correlate discharge pressure, bubble volume (water volume extraction), and time. Observe pressure rebcu=d following each incremental discharge. 1 Test C. (. / Calibrate differentir.1 pressure gauge. t Test D. Lcwer reacticn pressure frem 1000 psig to j 300 psig by discharging water medium mainta'ining constant temperature. Partially hydregenate system. During discharge continuously lower reactor l pressure maintaining an even water ficw frem the system. Plot and correlate discharge ' pressure, bubble volume and time. l Test E. Lcwer reacter pressure fren 1000 psig to 300 psig by dischargi.g water medium maintaining. 7.-

constant temperature of 280 F. System to be fully saturated with hydrogen at 1000 psig and 280 F. During discharge continuously lower reactor pressure maintaining an even water ficw from the system. Plot and correlate discharge pressure, bubble volume,and time. Compare bubble volume in reactor and steam generator. i Test F. Repeat Test D discharging water from pres-surizer as an alternate to discharge from vessel. I Plot and correlate discharge pressure, bubble volume,and time. Compare bubble volume in reactor l and steam generator. Test G. Install a Bailey Differential Pressurs Gauge from central reactor location to reactor discharge line. i Bring system up to a temperature of 240 F s / and 40 psi, release pressure and degasify system. l Pressurize system to remove vapor bubble. Bring total system up to 1000 psig pressure and 280 F temperature. Depressurize system using depressurizer dis-charge. Measure discharge volume to reduce system pressure to 300 psig. This test to be undertaken without hydrogenation of the water medium. t l

.._. u n During the test the following data are to be recorded: Time, System Pressure, System Temp-erature, and Differential Pressure (ins. W.G.). Appropriate plots are to be made for comparison with prior data and observations are to be made regarding the differential pressure reading against temperature changes and pressure changes. Differential pressure is to be recorded on a strip chart recorder to observe and document noise signals. Test E. The depressurized and unsaturated system following Test G will be saturated using measured volumes of hydrogen. The system will be maintained at constant temperature of 280 F. Following each pressurization step the pressure decrease will be observed to establish degree of hydrogen saturation. During Test E the following data are to be recorded: Time, System Pressure, System Temperature, ( '; and Differential Pressure. Appropriate plots are to be made for comparison with prior data and the differential pressure recording plotted against temperature and pressure to assess any changes. Differential pressure shall be recorded on a strip chart recorder.. Test I. The proposed secuence of reactor temperature and pressure changes.to bring the reactor down to 4 me. y . 7 _,eg-

standby cooling condition is to be simulated on the pilot unit. The sequence of testing will be as follows: 1. Bring the pilot unit up to pressure and temperature to simulate the system at 1000 psig and 280 F. 2. Saturate the system with hydrogen. 3. Drop the temperature from 280 to 130 F at a constant pressure of 1000 psig. 4. Drop the pressure from 1000 psig to 300 psig at constant temperature. During the above sequence the following will be recorded: Time, Pressure, Temperature, and Differential Pressure. The results will be recorded and the appropriate graphs charted. Task 10. A full report will be made on the above tasks. The report will document equipment used, tests ( undertaken, calculations and graphs made, observations, conclusions, and recommendations. The following sections summarize the specific experiments conducted by SILLINGS, as outlined above. The objective, experimental precedure, experimental apparatus, results, and discussion for each section are included. 4 y-n

."t. O THEORETICAL EVALUATION OF HYDROGEN SOLUSILITY IN WATER l 3.1 objective To determine the maximum solubility of hydrogen in 1 water at different temperatures and pressures based on Henry's Law.

3. 2 Results The results of calculations are shown in Figures 3 and 4. The tabulated values are shown in Tables I and II.

Corresponding results are shown in Figures 3 and 4 respectively. TABLE I Pressure Temperature Moles H /kg H O LBM H /ft -F Q _.. 2 2 2 y 1000 psia 280 F .0714 .00891 200 F .0551 .00698 l 130 F .0483 ,00604 250 psia 280 F .0179 .00224 i 200 F .0138 .00173 130 F .0121 .00151 \\,

Reference:

Dr. Angus Blackham 4 e S 15 -

- EWE l TABLE II Theoretical Calculations of Hydrogen. Solubility in Water Based on Henry's Law 3 (All Gas volumes are ft Based on 12000 ft of Liquid volume in 3 M.I. Primarf System) VOLUME OF GAS EVOLVED ~ 1 AT LOWER PRE 33URE VOLUME OF AITD 280 F C VOLUME OF GAS EVOLVED PRESSURE DISSOLVED AT STP av PER AV FROM ,f PSIA GAS AT STP av PER STEP S"TP 1000 PSI j 1000 1.9 7 E4 Base 2.00 E3 51.7 900 1.77 E4 51.7' 2.00 E3 58.0 800 1.57 I4 116.1 1.90 'E3 62.9 700 1.38 E4 195.2 L 2.00 E3 76.9 s 600 1.18 E4 303.9 N,7j 1.96 E3 90.0 - 2 500 9.84 E3 453.0 e N s 1.97 E3 112.3 400 7.87 E3 674.6 N 1.97 E3 149.0 300 5.90 E3 1037 REFERI2iCZ - WAPD TM 633 2 sh e l N.

---

e

---

n

+

.-..n_ - - ~ _ _ . ~ - _ _... _. - - I i I i 1 i l 1 2 me ramm *mK-

4. A e - -_"

--# -r i n - ..T.. .s y -me ; .rr 4 g. .. em.... ,a. - c. y,..._,,..t. y .p =, - _. - +. g r..... _ sw - _s + ~ .vt 4 6 y.9 ;,. m. ..,..,---ry.-.s ... w,.c .. _. ~. ._;. -[g a.=.,a.m. _.:n. _..c -.x i e. - -~ -- "X,.

-P 3.7 _==

• -w a JM, ~. M.;.

.M. ~ **. - m.e e f' j O a_ _. - e. + =~ m %.,.x ~_-_,...,<.4www, .r. .-a.....n,- r p $ p v.ge==ta .c y..a.... 3 ser m e,4 ~ hf'- %.a.r M w r a a s.., a., m---..;.p g.sre.gne f .i s r r -"4.Sem,e m an =. _ -- _ wr-?< gs g w n.. ar-m- --w-- e %.;, ,. r w < w.s n y Nhm=s.amas % +=0., A, 1 _. s. + e. Q H.7&im.w?.sW-weC e e xa.i'" ' -. e'.---b ~_ -_ - M..T

• • • e *s. 19. A g % 3 n m.. m. i.

-n. w.- e.

f I _ s.- _.;;c w g.y n; 4 g 3 -1 Jar va ; -. -.1 L rzer...= 9 . m - - w --,.. r.m e.-c.m s, man-=w ec r. =- f r.w.N h g 2 p e _ma m a e.u -.e c; - i 4.m m - -.- s. f w.e e M'r W ' _- _ --' W' _ - - - _;_w.,ar s<-mp-sea 4. e m-f.. m=l y

• { A n - -

we_ - e_ a a.ai.r _aw -f y n s e 'M._, c fj 4.c.p g 3 = _- ~ ~ ~- . ~ %.m w-~ <,, w,_- -; ar,- - -_ ra, u -..m 2 ,3s= % -f e _w s= 4 ed 4.0 -- IE1=M8'8Eh 'sas-buKWTN W A n _ g f4 2 4 ds.-d'5 As app =*sg-- 3 g Q ~ m 3 g W m = l n=e M m. . M e . - m m.- tad - h _m.- .I I 3 ,3 1 ~ k 4 0 e r-s* 3 bng-Am _~4._ m$ ,,_.n. ies M 1.eA84WhmMD_.er h"_3M" T _' _. - m"FMies,a 4

• r m m-

= x.

h .a - ad .r.v r # ---- - C -c -m. _m_ se ,...w m.nu, w - g .. ~.

• • 4 l

Q .c er. n i.4..-. l Q r,. ..w... .e Q I W = .s \\ -- _ - %ru _ _ m. c., xmu, st-- e%~ .L,. _.m.w w. Jet-Es r6 PT'L E -M_MBr mE.- s.- s.m. m as,- o 4-h4C C _m1= negr *-%r =4sr_- m gg3 e _ _- r -wrr -- - em ww-_ % -, c. v s -- Q _ _ _ _,, we. -sw-: st : .c _en + a m e ~ n - e -. -.v_c_.

r m.

m _ 1 _-. j = s we - _s.r.J m-f#'- C _i N ~?'*

• • '*#-'3"a->*'~~

r 4 ete-K '._.4- - _ w-m. - 4 _ _ _ __--- mc_T.gies.N'T% 'W * = _~wVas :> ) z -- - _- --%-; ^ i *sT 4 =. a' ' _ wegie - = c rT.rg/ _ --~ -== 1 3 gy a L N =w n..Le6

issa.e.

o8 w,.c .._w.. x.r-m.w,.w ; g m_ --_-me. _;mor.m:7. ..,,. a o aarf ni m ,..m-, . m _. w. - 2 m-u = r-. --a.-c.-,-..._- yy - -2 .m, e._ ca. -.. va.--%. v e.. _ n---- y_whaiA".. "l awig y.. ;, sf/..ww ? _.'*T* l ,j l*-

• r ~,

W-v Z:*_ . g-~.-_ - e b ] ,3 ", t _- ;,i,anw n -

• l1: ' YP">:'M E gg k #.= = - ' Dr _ -

.a?*w.nN p 3;i.* T W u erje.e-!f a w a w. -._a -..g wg y-a. - m-w = _.e's. w-- = ..__..M "* p 3 e g 4.J,%. 3 47.- . {e J __f.. vi M 3Q L.,

a. %.

- - - a= af R rtr ? -.is 18' ' 'O A.-*'m8-~~->f -_.=-adY%ii'***+"-*.-** '2 = e= .P

• * ~ el an' w=

S-*** >- J 2 ,O, o*= = t- - - - :.p.m.4 _ - -, - p. e - -, A. .4 e -.m. go..- n,-_. w _ J L - '.,-.3--4 \\'..---- It. t~ -. = 2 l E._ a _ _s.._L _ _ -N N

m. >=L-N

-*-a.-A g h."- y h. . zw _.,.. -; T =,.., c4.,-- .a..A .on D / ,,_,,a,3_.- ..x v c.-. - 1.4 ibew-M s_4., t-, caeayf -en.'r, 4 m."-'s. * * \\N. 'esH ~ ~ e .~ - N y. .,,.,,.u .4 s... .e4 W__ -- y,__ p - o. ,..~.--u- -r

• ,g4 C

C e W T N O = C' C c . 5 a was L'c o is $ Woo (msd

• H 1.

y ,s a' r ir 3 s / em 6l e.m ~ s O l .i k 1 ?.

t j: ,i 4 00 . *i 3 1 I { F a M_ r e ta j W 7 n / i y A A t e lE i r I / S l PW e P i i s 0 >s 1 MI lue O 0 0 O 5A 2F oP l r 2 li - * ;* nn d / E ea R goe U rr T dt i A yt l a P l, g nu I' r,'

1I E

f c P o >g m M: ! t, e E

o. t OT t

l i.. Ql' f l:l -e i I' I l i an co ii tt m ec rn oa i eP h Ta 3 e r O 0 8 6 4 2 O g a i 10 0-0 0 0 l 0 0 0, 0 P g N O x l[N I s1J. m n 'u s 4 ';I ,J4 I

s ~ i i SOLUBILITY CURVES - Tile 0RETICAL 20 10 g 18 17 eigure 4 " ' riieotetidal oata elot of inydrogen l n S lubility in Water la-16 L' 15 ~ o 14 I i -i ai 13 a o 12 ca ii (0 10 v 0-d 8 g 1 i.: ._i 6 m (I t 6 a m 4 sN N~~ 3 2 'sg,N l O~ Qa I I 1--.- a 1. l I _. l h s 3. g g g g j_ a = a O I 2 3 4 5 6 7 8 0 10 11 12 13 14 15 PRESSURE, (100 PSIA) e

l 4.0 SOLUBILITY TESTS / PRESSURE DROP TESTS, CONTINUOUS

4.1 objective

To determine experimentally the amount of gaseous hydrogen that comes out of a saturated water solution during depressuri-zation. 4.2 Introduction Experiments were conducted to determine >the amount of hydrogen that would dissolve into water at various temperatures. and pressures. An apparatus was constructed which allowed monitoring of the following temperatures and pressures: ~(1) the-temperature of the hydrogen gas in the void space above the water, the temperature of the water, the ambient temperature,..the pressure inside the vessel, and the barometric pressure.. The experimental procedures used for Tests 1 and 2 are described in Section 4.4 4.3 Assembiv of Bench Test Accaratus and Data Accuisition System See Figures 5,.6, 7, 8. ( 4.4 Exnerimental Procedure - Test 1 Test 1 was set up to give pressure / temperature information for continuous discharge. The experimental procedure is outlined below: 1. Fill vessel chamber with deionized water. 2. Drain water from vessel. . Measure in graduated cylinder. 3. Fill vessel chamber again with deicni=ed water. Drain out 50 ml. 4. Eeat to 137 C..

p I i i w 1 .s. N Motful Flunctor AppuruI;uu -1 Ituaction Cleasataer Haximum Working Pressure 6000 psia Internal valuene 1095 al e Material - 316 ss a l (A 1.ength - 34 coloutside dimensional e i 0 i + planeter - 9.5 can (outside dimension) p n s a i

1. Sournes Pressure Transduces

'o ( b Hodel 82900 'd Input Pressure 0-2000 psia 2, ( ]' 3 ] l~s i Haximuan Pressure 4000 psia ps ~t l Input Voltage 10.0 vde rt A Output Voltage 100.0 mude

2. Tliarmocouple il

~. - - - I type J - Irou-Constantan 1/8 incti stainless stool prol>e Manufacturer-Imue Controls t

3. Yliermocou le 82 e

h type J - ron-Constantan ta s 1/8 incia stainless steel prot >o Id xxx x x xxxxxxx s t s,. Nxxxxxxxxxs' Hanuf acturer-I.ous controls i j e, ) I

4. Thermocouple 83 i

type J - 1ron-constantan l 1/8 inch stainless steel prol.e i j i Hanufacturer-Loue: Controls 'p ( y.

5. Delonized Water p

C 3 C j

6. StydroeJon Buldsle

7. Ileater j

Input voltago 0-130 volta ~ / Output 0-1500 watts S. staowool Insulation + + 5/a" Llitck i OU j 8 I e

9. Ilydrogen Inlet

10. Ilydrogen Outlet to sians flow mater

11. Water Inlet Figiare S Experimental Setup to Model liydrogen

12. Water Outlet - to drain Solubility in Water l

u. support I

• q

9

• m m

f' ~g (E i r p 4 i t \\ ,~ 1 I i i / l [ h h f5 r' y <L \\ r e lO t g, .. l 6 =, y ,~ l m J t pf. @ 9n v ten inosr 1. Reacecr Cha=ber (see Figure 5) = 2. Ther=cccuple Switch ~ Lcue Centrols Corpcratien Mcdel Nc. 101 3. Digital Eyrc=ecer f Newper: CO. Mcdel Nc. 267 4. D.C. A=p'4=- Dyna 4cs Corpcratien 5. A to D 3 card 3.4 _' _'.4.e. s C~~ec..=._ s* *. c.. _=. _'.. t _T _') %_ _4 - Cen*7ersien cime 40,000/sec. 16 ch=~"el 6. Cc=puter 3.' _' _' _'.c..a C-r- u..=.. C rc_-=...'...~. Mcdel 3-100 ma. l 3 _' _' _'.*. c. a-r ~ u..a.. C e.=. = _'.. ~~ Mcdel 701 3.

• nt eracti*79 Digital Electer

m. G.s

., s. e -- e, % w = 7 w -. 4. w. w... b 4. 4 q,.= n f' . q.b.e., s.=. 4 mA q .S. g. 3 Aw5w_43.4 4.g.. = y 3 an .Jm .bo "g *.. b 4 4.7 9 .w-u. w Tests. 22 -

mmyw+en w y:?'.. ;M;xw% ..m:v+;;.. v r-g- -u i %...,yr~ y r.y -.s..: w :.e..~;,~e-. v~_g..n-- w. ~ p WMh*k. M !*I b k m+m'; .A =,,&.n=e bas: . _,.,_~.meq3jjlg __c h,._y M _' " _' _$Y E5 y- {- h

• 4W mp+--

.m __69' n N, m 43 a w s ~~ a-44 L ^ $ xn - - - =; MM -% J g{ pg; - _ Mt4 ) ~ b c ! b 4 ~~~ , o,, .4 ~- - o y__ %-j._ ~--- P " ~ ~ = ' > - ~, Q g gg-p_ga. .h,w,t [ " Y .j M~ b- -[_: s y---t p. w w- _~ G l ' ~N'.% WN .g y .e >. p .m r h ^

d

'L- ..,{ h W2.Q_; , N' ' ~f. _ n.' .i a~ g~ M*? ?- ? nh '**a ~ L .,,i'? W ', p brii. 2gbir7 . ~t '..,.* jg"*- .~ --d, , :'. - ~ >%

n -~. 9.~

y._gd* j' 4,,*.. >.- M. Q..; g,-i.,.. s m 2__ =

,y[y W ee g-y.

. ;:.*;cg;g P .kI[.h' ; G,

- ['

Y 2 ps;.?. ~.-; .p J., et *;' W kgp@W ?> *e r f tf. y n ~~;nh ';.9 *, :-..- ~ * -:523 :

f. /m k's.::

u, 'f l,....A-F '~~

t*f,'

t-r.

i (4 ^.'< IF, [ G ""*f G. M 'w~ M CYW 9.^- rSK.t - - Y..' 27: n,..,. A + w._ Fig're f ?hOtOgrap. Cf 7950 Ap?P!1~*15 l 23 -

a---._ n.- 4 -+exama. -ee m, ,a m g' gg - g 'N.- ? 1 ~ m, 1 wm I j , ~. \\ 4 e.s c 1 j .m a g j, P s.S ",'eph nusw n sse.m a # m -~ kWhM 46f wfZ 3u gs ;d? g q _- y > 7 n. .g g e .i )e_ 3, ~ c, s .&^ [^ _ _~ A W __ L ^ - ~ ~ } $y owm--- w_ f <C. 3 =s? t _, - D_, Q,p"~ c. e '- w M f.L1 - w_ w_ _ g,. 2 . ~=.,.maq 2 e >a w e, . Aw &- ht: g A o$ g. .a. .I**~p .e.

, 4,,,, 1%eg_s,Njc' !'*g% ;P_ ;

. -r ..,d Y j ~' WS* - p 4{. ~< gy. .O O t y lf 'I . ' ', " ' m= -A.,,.St ~~.5~; w e.i t. $ V @W

M

.. i.gS - ' r-< 3 ", l:~~jfg'T!.Y4.hhk,s. INT ^ ~3c '* ' , [v'$, ~ ' .g ~ 1,,.'."- . ;r.,, prA.T '::7 *i.%~**W ' *- 'W' E A-I g ~~" {v?.Ca.n,,J,,.. w:, 4 yg .x.-.. - s i ~. -W e *. s . ? .?y,W h[..y* d . ~.. -i-545p_ 3 p'. ;~'j,..,.y.-=*,=-f"."m"~ f@.}. ffg'sQ7l-' ;'

• f.*

~ ' f g 625 -[ t i s a d D'": y.. ;a y. K 5 = sa m d s. c m g'=.- A_ y' 24 -

5. Pressurize vessel with hydrogen to 1050 psig. 6. Rock vessel to help saturate water with hydrogen. Adjust pressure to 1050 psig. 7. Begin pressure drop tests by opening hydrogen outlet valve which is connected to a mass flow meter and then to a water displacement vessel. Allow pressure to drop slowly. 8. Record temperature, pressure, time, and volume of displaced water at 50 psig increments. ( Experimental Procedure - Test 2 - 4/4/79 Test 2 was set up to give pressure / temperature information for a fast discharge with intermediate stops._ The experimental-procedure is indicated below: l 1. Retain water in vessel from Test 1. 2. Maintain temperature at 137 C and pressurize with-- hydrogen to =1050 psig. 3. Rock vessel to help saturate water with hydrogen. Adjust / (_/ pressure to =1050 psig. 4. Begin pressure drop tests by opening hydrogen outlet valve. Rapidly d:cp pressure to predetermined value and hold for several minutes (until pressure is stable). 5. Record temperature, icwest pressure achieved, pressure at end of hold period, hold time, time, and volume of displaced water at 100 psig increments. 6. At ccnclusion of test drain water frca vessel. Measure in graduated c?linder. Measure temperature of the water..

\\ SOLUBILITY TEST DATA 4/70 - IlEC 20 u ~ IO \\ 2 18 5 \\. I7 3 n n g l,j l6 6 Il' l5 et o i4 ni 13 3 0 12 Test 1 Test 2 O (( i e g 10 m 0 8 7 tst _s 6 s. h) h1 5 3 h1 4 3 2 - 1 8-0 F*^"' 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 PRESSURE, (100 PSIA) Pigure 9 Plot of Dubble Volume as a Punction of Pressure O

_ TA01.11 I I { ~ Huti 1. 4/3/79 h,.lolubility Data - Continuc,. Discharge l'res su re Temp HGas V sample V Hubble Time (psia) (OC) Z (Ibil I I"II II /kg (ft ) 2 2 23:02 1033 136 1.034364 .000 0 0 0 23.03 1000 136 1.033344 .0002528 10 -36 (0) -2 7 (0) 23:05 950 136 1.031795 .0002405 125 -9 (0) -6 (0) 23:06 900 136 1.030239 .0002282 230 10.1 7.6 23:07 850 136 1.028678 .0002159 415 95.6 70 000 136 1.02711 .0002035 650 222.5 167 750 .136 1.025534 .0001910 904 365.0 273 700 137 1.023915 .0001782 1150 498.2 372 650 137 1.022327 .0001657 1429 661.3 493 600 137 1.021732 .0001532 1785 888.3 647 550 138 1.019102 .0001403 2348 1285.5 955 500 138 1.017489 .0001277 2769 1566.3 1162 450 138 1.01586 .0001152 3129 1796.1 1330 400 1 311 1.014225 .0001025 3469 2009.1 1485 350 139 1.01256 .0000896 3832 2239.8 1653 300 138 1.010899- .0000771 4160 2443.5 1781 "The ratio of the Voluine of gas at standard conditions leaving solution to Llie mass of solution. I i I i i I

g TABLE IV ItuN '2 11 Solubility Test - Past Discharge with Intermediate Stops 3 laressure Temp M Gas V sample V Dubble Time (pula) (OC) Z (Ibil ) hal ) cc il dMI (f t ) y y i 24 1055.0 136 1.035042 .0002663 0 0 0 24:53 978.5 136 1.032678 .0002s75 225 83.3 62.7 24:56 933.2 136 1.03127 .0002364 690 408.6 307.2 1:13 799.6 137 1.02705 .0002029 1160 613.9 459.6 1:18 755.7 137 1.025674 .0001920 1690 994.77 743.7 1:41 645.1 137 1.02217 .0001644 2359 1398.9 1042.3 1:50 543.1 137 1.01890 .0001309 2993 ' 1784.8 1325.6 2:04 446.1 137 1.01575 .0001144 3582 2139.5 1584.1 2:14 334.0 137 1.01205 .0000860 4207 2501.9 1845.7 2:24 306.0 137 1.011107 .0000788 4320 2556.4 1884 I i I ei t.

_~ ~ 4.6 Discussion The results of Tests 1 and 2 seem to indicate that more hydrogen can be dissolved in water at 1000 psi and 280 F than suggested frem theoretical calculations. This difference may be due in part to certain experimental conditions which might have contributed to the high results obtained. Those conditions were: 1. Deionized water was used, but no effort was made to degasify it prior to hydrcgen absorption. 2. Some air bubbles could have remained in cavities in the pressure transducer and gauges. 3. A small amount of water was retained on all inside surfaces of the vessel and tubing during vclume measurement. Since the water was not degasified by boiling and since i' air is more soluble in water than is hydrogen, more gas would evolve from solution than if hydrogen had been the only gas Ii in solution. Ls Any air bubbles initially trapped in any of the small tubes would be evolved during the tests, thus giving high results. Trapped air bubbles would result in the assumption that the vessel has a icwer volume than it actually has. Water retained in the vessel during a volume determination would indicate a lower vessel volume than actual thus contribu-ting to high results.

4 l 1 l It is estimated, however, that these experimental errors can account for probably not more than 20% of the difference between the experimental and theoretical calculations for the solubility of hydrogen in water. 6 1-C_.-) 1 6.

5.O SOLUBILITY TESTS - DEPRESSURIZATION AS A FUNCTION OF TIME

5.1 Objective

To define the parameters and general characteristics associated with the depressurization of water saturated with hydrogen.- 5.2 Introduction The experimental apparatus, used to determine the solubility of hyd:cgen in water, was modified such that depressurization and bubble growth could be measured as a function of time. (Compare Figures 5 and 10). Summarized below is the operational procedure carried out under controlled experimental conditiens in order to determine solubility rates of hydrogen in water. 5.3 Exnerimental Accaratus See Figure 10. 5.4 Exnerimental Precedure:- 1. Fill test chamber with boiled, deionized water through Valve L with make-up water unit discennected. 2. Rock chamber until water fills all spaces. Tilt chamber frcm side to side and upside down. 3. Repeat steps 1 and 2 until cnly water cc=es out of chamber. 4. Heat chamber to test conditions (137 C). 0 S. Pressurice to 1070 psig with hyd:cgen through Valve L. 31 - ~

1. stlei punctos-A 'laus ut,taa -Il (4

6 a. Va.nt Valwu e La es. 4 eagu l's essesas es Ta aeeadau;er 86.enuta t uses - ekmas sia lagoas t s eemiss es 0- 3004 gem 8 3 [tj 8 000 ol.m C. l'amanuse Tsaneduces 3 ?' N.snuf act us es - tsuess na j ikkle.l 840 2900 tg 4. Impiat Pressur es 4-2008 ges t a j ( m m8 mum Pressese doce pela 3 ls*gast Voltage 10.4 wJe D ) g3 Cast put Voltage 100.8 aivde 1 gt i Ll Id at TB.e s mucinas.le il 3 -a I plana.f ace user - luwe Coat rals Caerp. q } Tyseen J - Ison-constantan p 1/8 lach statalues steel psuAee g) o E. Fleermuuougile 82 i e pg I saanutect ures - 34 we cous sula Corp. gg l Type J - t run-Con.t aa. l au gg l I/t ina;te st eamloma na v.I lisotee i gg r. MyJsamaan asutJ.Be n G. hollmJ DelunlassJ W aer OO i as. Thesmo out.le i l - d Bauus T*:sapos at us el 0 l essoustacturer - novo const rale Cusp. I'3 T sus J - I s a.se-con st ant en h I & dancas st atulums steel prutse C E 8. Support Steactuse i g 7 J. M.shouse Wat er Unit l E. Akset air inlanat Vult 4.Je 8-4 I e wolt e ( Out gaat 0-1544 u4t a m

8..

Intut Habuup Water Velve 80. that l e t Valve O N. Inlet Valve I-east e s agg es.tus tu enado manute 41 44 s l.as ayais es ceae O t .Ilmee=inuct.J t.etu.en n and se O. Cyllu.lur Watus falsplaereununt Sasaple Cylismaus (umed lor aiuasuring clos ma== uf uator le wing reactaas wummel.3 e sausaat; e es. 4 6 t e o.a Isuaa;s us t'apus loweet a l as.4s et = - assual fas p. ps a g es Valve slutueelutmas time t omadanse **f layJswaum atlanneeleat tua in u.ater. u. Inlet Israiss>Juss Psessessiaer Valves as. t*s emmaes t acJ asyJs emamen G.se 3. Sk.gami t t ud asyJo emja.se Set eerat ona 86ahmeage Wales T. Emanesmal Immulet 44.se T ~ @!Ili1HD 9

6. Bleed hydrogen through Valve L into chamber at 1070 psig centrolling bleed rate and pressure with Valve A. 7. Periodically stop ficw of hydrogen and rock and shake vessel. 8. Repeat Steps 6 and 7 until pressure remains constant after rocking and shaking. (Hydrogen saturation point). 9. Disconnect hydrogen pressure line and connect to make-up water unit J. (J centains hydrogen saturated water). 10. Pressurize system through valve Q to force water into ( test chamber through Valve L while bleeding hydrogen through Valve A. Continue until water discharges fran valve A. 11. Close all valves. 12. Connect water displacement sample cylinder o as shown in Figure 10. 13. Begin PRESSURE DROP TESTS. ~ 14. Open valve P and crack open Valve M allcwing pressure to drop. ~ i N.J 15. Cool connecting line between vessels. 16. Discennect water displacement sample cylinder and j weigh cylinder with water sa=ple. l 17. Repeat Steps 14 through 16 until 300 psig is reached. O i I

l PRESSURE DROP lEST dlATER SATURATED d11Tilil12i S 1076 PSIG.,. 280 ID I Y' i i i e i i i,,,( e i e i i i i ui I100 3 .i ct gggg Pleial lli hubble volume

o (a t "300 "psig"Yressure 8, 200*P),

O 000 Apgwoz. Saturation 4.57s f inte.ial water volume, c ll f/ y [ 000 a Y C r I u) 0) f 500 b 400 ~ s 18 1 300 u) u) sl 200 a ggg 68 ! .i 11. 0 -10 0 10 20 30 40 50 60 70 80 00100 110 120130140150160 .s l TIME' HINUTES @!llinasi Pigure 11. :Pressuru - Time Trace Siiowing Changes and Subsequent Pressure itecovery as liydrogen Comen Out of Solution. i i 4

TABLZ V Numerical Values for Step Changes in Pressure as Recorded by Cceputer Time A**- S tar'- Of Test '1 p e Per Cent n 2 -3 Pressure (Min) (Psici (Psig) (psig) _ g e73 11.153 301.6 - 696.6 755.0 44.4% 21.733 753.0 603.3 737.5 11.5% 43.466 737.5 605.4 702.4 .26.6% 53.25 702.4 599.6 673.4 23.3% 75.433 673.4 601.3 634.6 57.1% 33.016 634.6 196.0 534.3

36. 3%

103.633 534.3 402.0 465.4 '6 5.'?% 119.631 465.4 396.1 426.3 .55.E% 129.5 426.3 303.5 342.2 63.6% 140.517 342.2 293.3 293.4 59.6% \\ l P 1 e o AP achieved 3 a l AP sce; o P2 l O 1 l de = ~ l

~ 1 m b l >0 = M. i f 5 4 4 4 ~~~'** ^ b 5 E 5.

=

O c c c . l a-C c ,c c e m m 0 0 4 3 n .N .N N ,, N c = o 4 O 4 O O

c I

I I I I i l i i = -= r.4 = m =s = = =. c z = u \\. = m a o m = = -w s =. - v _=., n =_ =.. = b 1. = ~= 8

e u. m N = =. = s = o = = t m. '2 c-. m.- = = .. a = s- = 3 e w u = =.- = = a 7 = =- = u = z l

• f3
• =

v: u ? 5 C l = ,= x m a = = = = = N N sH o u = m I. I l i I I I I I i s 4 = = = = = = = 0 " dOO S.7 4 N c Pe^aTqou 'c2Ian21 Y douc rdnsssud. d7

02. CIAEIEDY dCEC EYnSSIEd d0 lNEDEEd e

l 1 -

5.6 Discussion

This test represents the reactor depressurization procedure, in that a liquid-full system is depressurized by venting water. In the experiment the water was apparently not initially H2 saturated. Once saturation was reached, a large pressure recovery took place after venting. 10 to 20 minutes were required for the pressure recovery to line out. Four attempts to reach 600 psi, depressurizing frcm =750 psi, brought the pressure to =630 psi after about an hour. This slow response indicates that attempts to reduce pressure and infer reactor level from pressure are uncertain. Two items seem apparent: (a) The saturation point appears to be at about 770 psig at the start of the test instead of 1076 psig, the pressure at which hydrogen was bubbled through the liquid. (b) The 600 psig and 300 psig data seem to fall on a line connecting the saturation pressure with the m-target pressure for the step change. This is not entirely substantiated by the 400 psig and 500 psig data although they are bounded by the other data. O --

. =.....

6. O CATALYTIC SYSTEMS 6.1. 0 Catalysis of the Hydrogen-Oxygen Reaction by the Reactor vessel

6.1.1 Obiective

To consider the possibility of the H ~0 2 2 reaction being catalyzed by the stainless steel tank or reactor contents in contact with the hydrogen bubble. 6.1.2 Observations Regarding the Possibility of Reaction of H in the Reactor Vessel. 2 (a) The stainless steel head in contact with the gas does not significantly catalyze oxidation of H with 2 0 because 2 1. The stainless steel was previously passivated and the surface consists primarily of Fe O ' 23 Nio, and Cr 0 which exhibit icw catalytic 23 -5 activity (r = 6 x 10 mot,,j,2 ) based on h k. data frcm Soreskov et al. [Adv. Catal., 15, 285 (1964)]. 2. The surface area is icw (100-300 m ) depending upcn the size of the H bubble. Naturally, this 2 cenclusion should be verified experimentally. (b) It is not possible to quantitatively assess the extent of H -02 reaction that might occur in the liquid 2 phase catalyzed by minute particles (of Ni, 3r, CO

• t0*)

2' produced during the temperature transient. The extent of reaction depends upcn (i) the amount, (ii) particle. y - --~, -g yy y e y y r w

size, (iii) chemical state, and (iv) amount of avail-able 0. Unf rtunately, none o# these factors is 2 known quantitatively. It is speculated that most of the particulates are oxides of low catalytic activity produced by reaction of stecm with metals having diameters greater than 50-100 microns.

However, a significant depletion of H via catalysis by these 2

particles cannot be ruled out. 6.2.0 Literature Search for ether Catalytic Systems

6.2.1 obiective

The objective of the literature search is to explore the possibility of catalyzing a reaction that will remove hydrogen gas either by adding oxygen or scme other reactants in addition to the catalyst or by utilizing species already available in solution. 6.2.2 ' Discussion of Literature Search Findines (j A literature search was made for information concerning the catalytic activation of hydrogen. Information in "Ecmc-geneous Catalysis by Metal Complexes" by Khan and Martell indicates that hydrogen forms ccmplexes such as ReH, HCo (CN) 5 ' ~ and HPt3r[P (C H I 3 2532 This indicates that hydrogen might be activated by metals which could be added to the ccoling water of the reactor or which are already there because of the damage to the fuel rods in the reactor. For example, if rhenium forms ahydrideccmplexsuchasReH},thentechnetium, with properties similar to rhenium, may also form a complex hydride'. l 39 - wy

Catalysts which might be added to the reactor to induce reaction of H were considered as "ollows: 2 (a) Four alternative catalyst systems were considered (and are listed in order of preference) : 1. A colloidal dispersion of sodium borchydride reduced nickel (or platinum) [R. C. Wade, Catal. Rev., 14, 211 (1976)]. 2. A finely ground (micron size) alumina-supported nickel (or platinum). ( 3. A homogeneous Co(CN) complex. (B. DeVries, ~ J. Catal., 1, 489 (1962)]. 4. Catalyst coated glass microspheres. (b) Advantages and disadvantages of each system are listed in Table VI. Colloidal nickel boride.is recommended as the leading candidate because.(1) it has been used successfully in liquid phase hydrogena-tion reactions, (2) since the nickel metal crystallites ( are submicron, they will be uniformly distributed throughout the liquid with negligible settling or clogging of the system, and (3) the metal crystallites are stable toward thermal degradation to 350-400 C. Moreover, the chance of explosion with this catalyst is very small and the rate of the liquid phase reaction is i easily controlled by the rate at which oxygen is addcd to the make water. Last, but not least, because heat of reaction is, absorbed by the liquid, formation of het spots accompanied by runaway reacticn is prevented.. ,,,p__ ,.--,a e

TABLE VI Advantaces and Disadvantages of Four Different Catalyst Systems Catalyst Advantages Disadvantages colloidal Extremely well dispersed, preparation complex nickel boride active, will not settle, and slow controlled liquid phase reaction minimal chance of explo-sion supported well dispersed, active, may settle, nickel powder comercially available, may plug portions of minimal chance of explo-system sion Co(Cit)*3' Soluble, homogeneous reacts with H,0, reacts directly with H, rate of reaction relatively stable comp}ex, with H O catalyzed by 2 very little chance of acid explosion HC?l? Catalyst. coated stay afloat could initiate glass explosion, microspheres relatively inactive, difficult to recover l 1 -

The use of finely ground Ni/Al O s a viable 23 2nd choice alternative to nickal boride. It has essentially the same advantages as mentioned for the nickel boride. However, it may tend to settle and plug portions of the reactor system. TABLE VII Catalysts Prepared by N4BH Reduction 4 Catalyst Solvent Used in Chemical Physical Reduction Characteristics Ni-A HO Initially black; 2 turned green because of hydrolysis; coarse precipitate Ni-B HO Black; reasonably 2 fine precipitate Ni-C Ethanol Initially very fine, black ppt. Turned to gray-black and moderately fine precipitate when water was added and the solution was boiled ' ~ Ni-D Isoprepenal Same as for ethanol; precipitate was initally finer than all other catalysts Pt-A HO Heavy black flakes; 2 prepared from aqueous solution [ of H,PtC16 t Pt-B Ethanol Reasenably fine black precipitate; prepared frem Pt l 1 DNS plating solution l l n--

The last two alternatives are not reccamended, 3-the Co (CN) 5 e mp ex ecause it reacts rapidly with water as well as H in acid solution and catalyst-2 coated microspheres because they could possibly initiate an explosion in the gas phase, but would do little to remove H dissolved in the liquid. 2 (c) Estimates of catalyst requirements: Reliable kinetic data for H oxidation in the H ~ 2 2 rich region are available only for Pt [Hansen and Boudart, J. Catal., 5 3,, 56(1978)]. However, work by Boreskov et al. [J. Chim. Phys., 51,, 759 (1954)] suggests that Pt, Pd, and Ni are "the best" catalysts and reasonably close in activity. However, Ladachi et al. [J. Catal., 4, 239 (1965) ] obtained data, showing Ni to be 10-100 times less active than Pt. Leder and -Butt [AIChE J., 12, 718 (1966) ] studied the H -0 r**#~ 2 2 ~" tion on Pt in the oxygen-rich region. Their data shcw reasonably strong inhibition by the product water. \\ Accordingly, the reaction in aqueous phase will be significantly lower because of the high P and because 302 of slew diffusional rates in the liquid. The specific initial gas phase rate of the H -0 2 2 reaction en Pt (see attached calculations) is 5.5 x gm es 10 The gas phase rate on Ni is 5.5 to hm emoles 55 - and fer liquid phase 1-2 orders of magnitude 2 hm 4 icwer in rate.- If it is assumed that 5 x 10 gmeles are present in the reacter sa er and bubble and it is.. .y--

\\n ) desired to remcve the hydrogen by reaction in 5%,0, I.; \\ over a period of 8 hours, the nickel boride catalyst requirement is estimated at 27 lbs. (12,500 q) (see attached cair.ulations). Since this figure cot $[d be high by a factor of 10 and low by a factor of [0-100, it was recommended that the rate be deterained exper-imentally so closer estimates could be made. d 6.3.0 Material Balance and Ther=edvn=4 c Study A rough calculation was made as to how much uranium had. s i l undergone fission since the plant started. This.was calculated- ~ 5 to be 1 x 10 grams. Six percent of the uranium that undergoes fission ends up as technetium-99. Similar amounts of palladium and rhodium are forned. If it is assumed that 10% of.the core P was.. damaged accessing water t the fission products-and that 103 of these products is carried into the water in dissolved. or finely divided form, there could be approximately one pound (600 g) of technitiu=, palladium, and rhodium circulating s with the coeling water. As indicated in the previous.section, these elements are known to interact with hydrogen in catalytic l reactions. Perhaps these elements are in part resgensible for the reduction of the hydrogen bubble in a reaction sequence in which hydrogen forms a complex with these metals and these complexes sicwly reduce scme of the metallic cxides to the metals and water. If a high temperature reaction converted water and netals to hydrugen and =etallic exides, then at icwer tedperatures there would be a thermodynamic tenden" for hydrogen .__,._,,-.,,---,--m--- -no---- - - + - - --- --~' '"^~~~ ' ~ ~ ~ ~ ~ ~ '

_a3g- -m. i 's x\\ ( _N ~{ y ~, ,, u s s

\\

to red'1ce the mitallic"exides. The technetium, palladium, ,..s ~ andsyhedium say have provided,a catalytic path with favorable 4_, \\ s kinetics for\\this.,reducticn.' % s %,. yy \\ i t, c. 6.'4ds t ' Q4talytic' Reduction of Oxycen :with Hydrogen under Water i s 4'. ,El \\

6. '4.15 Objective :

A .?'- 3 ,i + ,s y ( Tu.abtain empirir.11~ evidence as to whether or not platinum w x, '. ,s s. and/or n,ick,31 catalys.ts a'dded to the reactor system might assist ,e . in the re;noval %s hydrogen by reduction of oxygen. o rA 1 q 1T 6.4.2 \\ Catalytic Reduc. tion d2 Oxygen with Hydrogen under Water C,.aIvst .$n a Platinum Exneriment No; i. 54 mg of . dere added to 425 ml of 44 t in a 500-nl g, lass reaction bott.leand'placed in a Parr, wate2-D... ' ;, - y.3 icw-pr, essure hydrogenation system. Th e,. tfore, 75 mis of air ~ 'f ( 1 at one atmosphereipressure were trappei in the bottle. Hydregen N g was adde'd to givdJa gaucje ' pressure sof 50 psi. Therefore, the' s s 3 q relatlve.'prgssure due~f[o 0 9, J, an'd"H are, respectively, 3, 2 2 12,'and 50 units. If all cf the Noxygen ycmbines with hydrogen) n s ( v.there will be a 9 psi pressure drop. The hydrogen necessary \\ i>., ~ . Q N-3 to reduce the Pt02 should \\7ive a 2;,4 psi pressure drop. Shaking ( 1 p of the reaction bottle wu started and the pressure drop during n cne hop was cbserved as*,tadicated in Figure 13. The drop of 11.,5 psi ccmpares closely with what was predicted.

Ecwever, 7

7solub111ty.c,f hydrogen in water was not censidered. N %e /* o 1 + P 9 h r \\,, - + i 45 - ...a

/ s .l SC' ma 2 1 & 4e I.. w a. M I N i i, 35 0 15 30 45 60 75 T!me (minutes) Figure 13 Pres.sure-Time Plot for the Catalytic Reduction of Oxygen with Eydrogen Under Water en a Platinum Catalyst Experiment !To. 2. A si=ilar run with only water. indicated a pressure drop of 2.5 psi due to solubility of the hydrogen. Experiment No. 3. A similar run was =ade with 10 grams \\ of catalytic pellets that had been i=pregna:ed with a platinus n plating solution and heated at 600 7. In 10 minutes the pressure had d:cpped 4.3 psi. Since the platinus had already been i / reduced, this pressure d:cp was due to solubility and to the reaction of E -C cn the cellets which remained on the bettem 2 2 cf the reacticn bet:le even when it was being shaken. Experiment No. 4. The water was pcured off and the conditions of Experi=en: 3 repeated. In 10 minutes the pressure dr:p was 7.5 psi. Thi., mea.s tha 56% cf the Oxygen had reacted ca:a-lyrically en the p ilets at tne bc::c= cf the reacter. 4s.

1 Experiment Nc. 5. Another run similar to Experi=ents 3 and 4 was run for 1 hours with the same pellets. A pressure drop of 3.2 psi was observed. Overnight, withcut shaking,' the pressure d:cpped an additional 2.5 psi. The H -0 reaction en a Pt catalyst underwater is thus 2 2 shawn to occur. 6.4.3 Nickel Catalyzed Reactions Four different Ni catalysts prepared by sedium borchydride ( reducticn of nickel nitrate and two Pt catalysts prepared by bcrohydride reduction of chloroplatinic acid and Pt DNS (plating solution) are briefly described in Table VII. Iach catalyst was prepared in a well-stirred flask containing the =etal salt solutien to which MASH (solid) was added slowly. 4 The reacticn rate was centrolled by ecoling the flask in an ice bath. In the case of catalysts prepared in ethanol and j isoprepanol, the solvent was partially boiled off after the l \\' reductica was complete; then water was added and boiling , / l o continued to a temperature of 98 C. Two of the Ni catalysts, Ni-3 and Ni-D, were tested in a 300-n1 stainless steel reactor for activity in acueous phase oxidation cf hydregen by exygen (512 cxygen in hyd:cgen) at 25 and 140 C and 700 and 1000 psia (50-75 at=). 6.4.4 Results - Nickel - Cataly:ed Reacticn Tests and results of the Ni catalys: experiments are sc==arized in Table VIII. Chromatographic analysis for catalys: 3 indicated cha: essentially all of the exygen (criginally -

TABLE VIII Results of Activity Tests for Borohydride Reduced Ni conditions: 25 or 140 C, 600-1000 psia (50-75 ats) Apparatus: 300-m1 stainless steel bomb Initial 0 'ressure Catalyst Time at 25 C Time at 140 C P Ni-B 3.5 hours 1.5 hours, 675 psig Ni-D 2 hours 4 hours 675 psig Chromatographic Analysis of Products after reaction Catalyst A: about 54 Oxygen about 5% Nitrogen remaindar Nitrogen ~ present in a concentration of =5%) was still present.in the reactor following reaction for two hours at 140 C and eight hours at 25 C. In other words, the Ni catalyst did not catalyze the reaction of hydrogen and oxygen in aqueous phase. Based ou the limited data available at this p'oint, two possible reasons for this behavior are suggested: 1. The finely dispersed nickel particles are passivated or oxidized by hot water. 2. The surface reaction is strongly inhibited by water and the product or reaction d

Evidence that colloidal Ni was oxidized was obtained during the preparation of catalysts. The catalysts prepared by addition of water followed by heating to remove the nonaqueous solvent were a gray-black color, characteristic of nickel oxide, rather than dark black,which is characteristic of finely divided metals. 6.5.0 Appendix Calculation of Rate of H -0 Reaction of Pt, Fe, Ni Metals 2 2 1. Pt at 410K (138 C) and 75 atm \\ 3 ~ - - k = 0. em [Hansen and Boudart] cm S ~, r=kC where C O O 0 2 2 2 3 - -- -- r = 0.1 cm 5 ata -5 g moles 2 = 1.52 x 10 2 cm S C*3 cm S

• t" 82 (400K)

K gmole 2g = 5.5 x 10 2 m h r If reactor were made of Pt i 2 g=o R .5.5 x 10 (100 m ) 5.5 x 10 g moles /h = 2 hm 4 No, moles of H in rea t r is 4.7 x 10 = 1 hour to react 2 2. Ni at 410 K and 75 ats i According to Ladachi the rate of Ni is 10 lower than Pt. gm les r = 5.5 2 hm R = 5.5 (100) = 500 g=cles 90 hours to react at the = n wall ~

NOTE: These are initial rates. There will be strong inhibition b'r H O (see Leder and Butt). Moreover, the 2 passivated surface is probably Cr 0, Nio, and Fe O. 23 23 Calculation of Catalyst Requirement: 4 1. Nuclear reactor with 5 x 10 gmoles H 2 Assume rate for Ni is 1-2 orders of magnitude lower in liquid phase: Y" r = 0.05 - 0.5 2 hm 2 Ni SA = 10 m /g (lower limit) Allow 8 hours for reaction; upper limit 4 Amt. of Cat. = 5 x 10 qmoles x = 12,500g 2 10m (27 pounds) 0.05 gmoles(8 hours) 2 hm 2. Lab reactor with a 10cc vapor space: 3 n = E = 75 ats (10 cm ) (. K gmole 82 (800 K) gm es 1 amt. cat. = = 0.1 g 0.05 gmoles 10 m (0.5 h) 2 hm I 3 e e

by Dr. Rchert J. Teitel (Censultant - Microsphere Specialist) : 1. SI Grade Ratio of Diameter to thickness of wall is 35.3 "'hickness of wall is 1.9 micr= meters Thickness of Diameter is 67 mieremeters "'ha pressure requirement of the mic cspheres bed is i:n=aterial. Exa=ple: 91000 psi, then in 10.1 hours the micr= sphere bed will be at 500 psi. Then at reem temperature the microsphere bed-would lose 1/2 its charge in 1200 hours. 2. 3 M Grade 338/4000 (a) At same pressure conditions, it would take 59 days' to build up to the same pressure as referenced above. Thus, 1000 psi would result in a microsphere bed of 500 psi. (b) at room temperature the microsphere bed would- - loose 1/2 its charge after ten (10) years. ~ 'w 7.3 Ex=erimental A=caratus (See Figure 14) 7.4 Ex=erimental Precedure The =icrospheres test was performed ace =rding to the follcwing,precedure using SI grade microspheres: 1. Place 150 mi micrespheres in 300 ml pressure vessel. 2. Fill with boiled, deienized water. 3. Pressuri e with hydrogen through bettem valve of vessel. Displace sc=e water frem vessel through tcp valve to create " bubble.". e9

7.3 Eye =erimental Apparatus Microscriere Test Accaratus Pressure Gauge o-tsoo psi Gaseous Area; Sc=e Water and Microspheres

i.:,gg'.

• p, ti!4;$df$3N"5

-{$g{ygj fi gN Packed Microsphere e Layer with Water ~ soo mil Vessei ',... - C ', .s Water Slurry with Microsc,heres r..., valve m<Xi / H2 Succly %./ Thermoccucle Lt:$illirx3sl Figure 14 Microsphere rest Apparatus C se

I l TABLE IX Pressure Data for Microsphere Test No. 2 Temperature Pressure Tima and ('C) (psig) Date 286 990 12:10 am April 5 280 975 12: 12 am 279 975 12: 14 am 277 975 12:16 am 277 975 12:20 am i'~ 282 975 12:25 am 281 975 12:30 am 280 970 12:45 am 278 950 1:45 am 275 910 8:25 am 275 900 9:00 am 275 900 9:30 am 274 895 10:30 am 274 895 11:30 am 273 885 2:00 pm = 600 12:00 pm G 4

This was done to see if the amount of hydrogen absorbed could be detected from the weight difference between constant weight at 220 F and the weight after heating to 600 F. The following data were recorded: Drying Time 9600 F 30 min 90 min 3 hrs 6 hrs After Hydrogen Outgas (gm) (cm) Sample A (Gaseous Area) 9.2903 9.2893 9.2886 9.2867 9.2723 .0144 Sample B 9.3730 9.3707 9.3700 9.3683 9.3414 .0272 (Packed Microsphere Layer with Water) Semple C 10.2011 10.1883 10.1832 10.1759 10.1095 .0536 (Water Slurry with Microspheres) The data indicated that a significant larger amount of hydrogen was absorbed by the microspheres in the water slurry than the microspheres in the gaseous area of the test apparatus. It was concluded that additional tests would be required to fully define the characteristics of hydrogen absorption in the glass microspheres. It was however concluded that the S.I. grade responded much more rapidly than did the 3M micro-spheres. e.

8.0 REAC*OR SIMULATOR TESTS 8.1 Objective The objectives of the reactor simulator tests are (1) to determine the effects of pressure and temperature reduction upon the behavior of the reactor system filled with hydrogen containing water of varying degrees of saturation. (2) To obtain a model of these characteristics so that the degree of hydrogen saturation and bubble size might be ascertained through ( pressure and temperature measurements. (3) To assist in the eventual cold shut-down of a nuclear reactor system believed to contain a hydrogen water solution. 8.2 Introduction A reactor simulator was constructed so that the charac-teristics of hydrogen bubble formation in the reactor and steam generator might be observed as a function of pressure, tempera-ture, and hydrogen concentration. Manipulation and documentation of these parameters in accordance with conditions current or anticipated at the Three Mile Island Reactor were expected to reveal means of controlling the effects of hydrogen gas in the reactor system and to forewarn engineers as to what to expect given any set of circumstances relative to changes in temperature, pressure, and hydrogen concentration. 8.3 Experimental Apparatus and Data Acquisition System A schematic of the si=ulated reactor apparatus is presented l I in Figure 15, followed by a description of the components. Figure 16 shows the completed reactor simulator system in service. < d

The simulated reactor and data acquisition instrumentation are pictured in Figure 17. Figure 18 is a photograph of the simulated reactor system under construction, wherein the major elements of construction are visible prior to being covered with insulation. No commercial pumps could be obtained, within the time constraints of this project, which could withstand the pressures anticipated. A Grunfos #25-42 sf pump was thereby encased in a heavy steel shell and immersed in hydraulic oil. The pump casing oil was then pressurized or depressurized in concert with ( the reactor system through a large diaphram to which was attached a lead line from the pump on one side and a lead line from the reactor on the other. The encased pump is shown in Figure 19. The data acquisition system is shown in Figure 20. A Billings B-100 computer with 48K memory and dual floppy disk drive system was used to access and store data from the various thermo- . couples :and pressure transducers of the system, as a function of real time. In addition, the data were printed on paper with a Billings 701 printer. Further backup redundance was accomplished by printing the time and temperature readings on paper tape via a model 9300 data logger by Monitor Labs. 3.4 Discussion - Reactor Simulation Tests 8.4.1 Experimental Procedure For the experiments requiring saturation or' varying degrees of hydrogen saturation the following procedure was ccmmon: 1. Fill the system with tap water from the hot water heater through V 14 with V1, v2, V4 and V8 open. (See Figure 15).

• 2.

Close V2, V1, and V4, respectively, when water begins to flow out. 3. Manipulate pump and valves to dislodge and remove bubbles from the system. 4. Heat to boiling at ambient pressure with circulating pump on. 5. Remove bubbles from system by opening and closing valves V1 and v2 intermittantly. 6. Pressurize through V4 to 1200 psig. Relieve 1/8 - 1/4 k volume of pressurizer, PV, through V7. m .7. While proceeding to heat to the desired temperature 0 (280 F), release hydrogen bubble from PV through V4 while admitting H nto R at V13. Hydrogen bubble 2 forms in R. 8. Saturate with hydrogen through V13 with V8 closed and V9, V10, Vil, V12 open. Vent hy'rogen slowly'at V2 d during hydrogen charging. ( 9. Pressurize PV to 1100 psi and release hydrogen bubble in R slowly ?vintsining at least 1050 psi) through V2. 10. Test saturation by reducing pressure at V7 in 100 psig increments. Look for pressure rebound. Repeat until pressure rebound is observed. Saturation will have been achieved for the pressure at which rebound first occurs. 11. Water and/or gas samples removed frem the system batch-wise or continuously according to the intent of each individual tests. The sampling port differed also ac=crding to the design of individual tests. _

, l I (, I ) s v-l n atort Mack-up l 3 2-V-2 V-3 F BGB

• Btearn Generantor simulator i

g ,m - H.a..... UGS E O P-Pump w' 3 Ts J PV-Presourlaer Vessent 9 -4 GG j . x 0 0 - Ole Olmplira m <g PaV-P,...or.

n...., v....

T.u C R I CT-Clearsing Tank PV O llc - Heatino calt q _ 2 0 - OmuDe '] _ fj l {n y, V-8 PTO - Premesures Tronuducer Gauuo m 4Eg V-T - Tisarmocouples gg EG - Glotic Clama OD R - Rouator l4G - Ifant Unctsanpur i v-a Q y _ v.,,,, v-ol T P"" l' u-a -a-y q us - H=atian es===a= e u u v-io 't - TFf ccc)h v-2 j u - union Papa Fittinua a CT j { j I U I' ' / uc f -13 ' i Figure 15 Schematic of tlie. Simulated Reactor Apparatus 1 in

N,.L9 E - Ar-siu 'w_,,,c,m,,# g,.y;.me.&0--p TC -._ Z' ' w~; --- = >=u n e r w. u m g.=.~ W. -Gf gl w> S@@.>Mh._.::b,kM..,k,g} pas.4 hk,'# 3 y--g y pm -. ~ _ w:v.: n.,....___m.=.. N _ h$- ~ h% ) ~ ~I ../ ~S~; ,' $p, k. ?~~. ~ = ~ ' e r %.. * "T y:7 r- -Ttr.n 5;.g

. 3 -

> c.. - - := - ?" a r-y 3.- , Abi yf,r.. 4. s l } grs S1;f? E E'Y S o EK' -s i r , - LBR pg g,

y "*" mg y

(' W. 8&f g { _ gM-M; 7a,,Q MQl>%r '**. g:::;C. Mir p-

,F ~ =a c.

h*.# e M*' ' I arJ> F_BJ T e PV b @M I k, !.- # *1TX [ w [* 9G ac

+f z

fr A..- g y. - -z_. m.x:. Qg. 4 7

g. y.s..-

+ n,- a f" ? ff ' o @* :. y T r %**% c c =- 7 5,y.gpJT*y. j s a M:: ,%) +- - W. *, m i

M-

] q ~ 4 ( W~+'__ M m~ s -c l

.,, cJ.

P e f p- ~ f G 4 i HG,, _ w. e k Th ] ]!g.,,. g'ea [. -] . e. gp. 'Ew ac -wj4 a 2@dy- - W i W g% @4as ~ , q. . "1. .'F c . f.. g 1 'i n.... w %n:q%f% 1, d. + -2

,,.. - -[ ~~ - '

v c;--.:

• g~

~

JI?-NY.f

.:.-aQ! ~

W.'. **

.-Y ~ 2;c&c.-w, W. _. 1,.... -_ -l a... -sme- ~ 3.'p ~ 5 -~ :s myp;__ g.- ~. g ~C; m,.w.;: i-f.q J ' AM. Vn =. Figure 16 Phot 0 graph of the Reactor Sir.ulator Tes: Apparatus I

Description of Components: SGS - Steam Generator Simulator

== Description:== Government surplus pressure vessel Dimensions: 0.D. = 16" Length - 48" x.875 wall thickness. Function: Simulate conditions in and represent the actual steam generator HB - Heating Band

== Description:== 750 watt 110 volt heating band Function: Heat water contained in the steam generator P - Pump ( Model: Grunfos

1. 25-42 sf

== Description:== 15-20 GPM delivered flow tate Encasament: BILLINGS in-house construction 1000 psi design Function: iEncasament) pressurize outside of pump to equalize pressure pumping conditions Function: (Pump) water system circulation PV - Pressurizing Vessel

== Description:== Luxfer 2,000 psi pressure vessel Dimensions : 0.D. = 7" Length = 15" Function: Simulation of total system pressurization OD - Oil Diaphram

== Description:== BILLINGS in-house construction Dimensions: 0.D. = 10" Length 3" Function: Prevent pump pressurizing oil from entering water system PRV - Pressure Relief Valve Model: Consolidated

== Description:== Standard relief valve 3,000 psi capacity Function: To allcw pressure escape beycnd control pressure CT - Charging Tank

== Description:== Steel cylinder 2,000 psi capacity Dimensions: 0.D. 4" Length - 24" Function: Heat exchange, decreased heating time for water, also sample and material addition

HC - Heating Coil

== Description:== 6,000 watt, 220 volt heat coil Function: General water heating G-Gauge

== Description:== Standard gauge - range 0-4,000 psi Function: Monitor reactor temperature PTG - Pressure Transducer Gauge

== Description:== Standard gauge - range 0-3,000 psi Function: Monitor reactor temperature and interface with Billings T-100 ccmputer ( T - Thermocouples

Description:

Standard type J-iron constantan -1/8" stainless steel t-a Function: Monitor water temperature in steam generator t-b Function: Monitor gas temperatures in reactor t-c Function: Monitor water temperatures in reactor SG - Sight Glass

== Description:== 1,700 psi capacity; liquid level style Dimensicus: width: 4" thickness: 3" length: 13"' Function: Monitor liquid level in reactor ( R - Reactor

== Description:== Government surplus pressure vessel Dimensions: 0.D. 16" length 48" x.875 wall thickness l Function: To simulate conditions in and represent actual reactor HE - Heat Exchanger l

== Description:== Billings in-house construction; coil 3/4" j finned stainless steel tuning, Function: Cool samples HE - Heating Element t

== Description:== 4,500 wat: 220 volt heating element Function: Heat water contained in reactor and simulate actual core conditions

U - Union Pipe Fittings

== Description:== 3/4" union pipe fittings (2 reg) Function: To afford a disconnection of the charging tank i from the main system. 4 V - Valves V-1

== Description:== " ball valve Function: Venting for steam generator simulator V-2

== Description:== k" regulator valve -Function: Venting for reactor V-3

== Description:== %" regulator valve k, Function: Venting and filling valve for reactor V-4

== Description:== " ball valve Function: Venting and filling valve for pressurizing vessel V-5

== Description:== h" ball valve Function: Isolation valve for pressurizing vessel V-6

== Description:== " regulator valve Function: Purge valve for reactor sample 'line - V-7

== Description:== k" Regulator valve Function: Sample isolation valve V-8

== Description:== it" ball valve Function: Bypass valve for charging tank V-9

== Description:== " ball valve l Function: Isolation valve from charging tank V-10

== Description:== h" ball valve Function: dpolation valve for charging tank V-11 Same as V-9 l V-12 Same as V-10 V-13

== Description:== k" regulator valve Function: 31eed valve for charging tank V-14

== Description:== 1" ball valve I Function: l System drain and fill valve for charging tank \\ l 1 . 5 :*,- v..f(g,'v...f"'*~ { _e .#*C""'" r p--1 -. g.3. my: - r-:;.ca;. Q g. 4 y'

• M

?'. :.. g. y,;,, , Q q.3, ,: Q,,.. i: : ? ' **

• E.'$

-~: N E: & f 1l%[5f,,.sy:~...$.,,k C i hli O$..j' ' '. M l ?-? 7

.,' i.,~,' 2 '&

' * '; *

• r' L

~: a [ J.4.;F;UI, J ' j,. ;. $. ^ T.4==== i.Ua 9:..?, :3.,'...;.4: O46 ,W.... p'.u = p .c .r. j?;:; E:i.*~ * '~~~ l>,? *.: ~~ %S at ' ~'

["; f -

s ) ?. b 4 ,s, ,%'Cif :.22 : 74 :l-A$25Qiiichi'j i!M.kd'IM *:[c"-{$.4 %.u,- ~~^ j ?.y. us ~?t Pt -d:'y :n .c W +., n. <1-

,4,*

ItsAryg... g.M{ N1 ~ 74'$.h.1 e u_i - e gj b M' a m. b ~ m. m ] ~ y u R. $. %_. .R L & 7 ^~ .'.n, w

• &(hQhf' ih3-h

..f4 Y q- .e g -. R.: s.. va, s c a e

.^

,_y 2o g g~_.

. M.

wnvu mm ]t? m ' t= p; Q.p 'm e y.g~a + e-n =e - g -=* =. - * > .z.] u.y ~. -, -. w g, r :e n,=c .;p >.. .g q ~ o-n g-au +.4

e A me:.

i+lj,@ 5 ? ) y: a. v-l 's.; l ' f ;. - *.. '. ** ,.j<i.'k'be-.,=*e ;;;t4[.

• F,' '

,,,l:j ?g a 'Jp,

a. T,j p l

g-c;c '" 1,:d - g m' I-l ~"~ 1

~- e 1 e 4st ...c.....u.-. .; w.s_ ..x 3. r.__- w u d-. w t w : mrmn-e m s e z -w, !f

• a w

n$a =~- + 1 Qedt-L ... _ W.F y. * .-u e r nn j!~ ~ I i y f-g -3.s, .mm 2- =hm y,: t-:;%g$__-- _ _.. _ _ sg,q .6. .e-

h

=

."[.$ '

'ha 3;ur. 6:7 i 1 I %6 ( ~ x-M % m#.e4 i m l. 1 m' I gMgf ;q y? # 3 3 n o -w-G%s c p '. i ld y;.

x

=~;Q* P-

c. -

v9 b ~43

g. #

~ 2-s kg h~f. m. h7 f. M $ *[%p $ MT7 NJ -3 p =. - C -- ' ,h e.;c -1: - a !.h.Mx!.. n u v e- _., v-*e _ -- ; q:n t

.;4.2 e.i :: ->g

' g w. ~, .JN. e 3 .,m- -ur q .. :. c.; a ,,~ --i & P4- _ '. n - W.;4., . s! .c ~[ n'% F-L, 'M 2 Dh1 M7 D!O ge,w r & -..L Y.!%,k.I [ ~ W.

p. -

ws s - _-% _ j ( . g. r s - ~ N,. q, ~, 1 5 .s _wc; + 4,t,w..we f. n.-_ -,Ag3 3 -

g.+ 4

?- kp h- . '";.n , r gn: , g, s a gwpA 19 (y- - Y#' ( g+j%$ M I c w w 4.- g: &... g e ,,,h, "'I '95 k .e .a, --h $k hM[. N.. ['l n, ...y. s ~ g-g, -- - g, -."3, w ' ri V;'.~.d.. ..p;...-+.=m#:::. y .n._, .w.a a. -,v. m. .~. ..:=a. 93@yb.e; w?. - }:.-:*-R=+'>&. .. e, ; y; gLt:d$= rXsh g.

._. g.-u.

2 Figure la Photograph of the Simulated F.eac:cr Syste Under Const..c.ctic. ~ 66 -

-p b F],..g%,..;-W*. GA."h ' "' ?. c".-9 '. - f '.i. ~' e'^ LT'A. M.f'5% { 'J 6 h;WSSN[M ',$EY58 p -~7x:lrdi.PM"%d c 8#i. p'$ ages3e> \\ m at _%s -e s i j h .l z t .. y sih4 . 7 Ij

• J E

@*e0%Q4 - ) f:'s' k,4% N^ e 's y2 b h-M ] l '.%,p$ k,, " c

• g w

??.l'~'- h *. YT5*.e. pH ( j., aM s.' e ' ~ ? 'iN .h si,f~ II.',k ' R f/ R ^ $.. ?,.l Y% ~ .l. _ ~- [-

1. k

-[T,'."7~9' f ~- 4- ~ ~ p s-4 f.. a 1.,,, M Kw-r R' ~~.; ~e y l q ~ + - - - n: W. v g-g' c r l 7 q<. p a y .q.re k, e - ~ L I; ETh;:., .r., y ,9. Figure 19 Photograph of the C -"'= ed Reactor Eigh Pressure Pu=p 67 - ~

l I

• . ~ ~

. - +w... -- m pyyy4 w.*=-- rem N i l mM ' [ [ &. ; ~ ~ y e- ?h,_, l Q. p NM ' Y 'l ~ o m. h . ~ '

k. N k n_,_ M TTL

~

• 2=%

? - Q.1G B!?D { ; - = Y ( Wej g C u mm.sd s ,,i !f-h W W W AFf St'd X $'g[ b."' *N y__ ' lbm -; j u e. O [_., .., - 1 g- ~ g ) .S 44 . "J -r e w ^ hN M U h.g l$" X C-@. , ' ~ - h 3:W.m r c ., = ~ [ uu N' iTC ~ am .. =..,. - sew--, s3 ~ oc M. ^ Ya y'* ., y erp r7 ^ s u @ s, m O. 'E.]f $^- 4 j _ sannhska 7 l8 >- N ,,...., Q W )$ % 'fA k'f -;~ $ $,_ 2 5' '. *d h $$5? l?$ o ? >'fA'?. f. g???

• h.ki+'.f,

(== .N Q. ;$~f'1 .g L ;'h! [ [') 'f

S

p

%i s.__-. .a ~ 68 - I

8.4.2 Test A - Acril 6, 1979 The system was pressurized to 1000 psig (ncminal) and 0 heated to 280 F using a primary heating element simulating the core and a secondary heating element around the pipe leg. A band heater on the steam generator was also used. During test, the secondary heater was disconnected. The primary heater was used to maintain 280 F. This heating element was controlled manually in an on/off fashion. Th e data plot ( shows pressure ripples that are coincident with the temperature variation associated with the heating cycle. Water was drawn from the base of the reactor vessel in an that produced approximately 100 psi pressure drops. The amount water sample passed through a heat exchanger La an ice bath to cool prior to measurement. The mean temperature of the water exiting in the heat exchanger was 86 F. The volume of water.. was measured directly with graduated glassware. The plot of Test A (Figure 21) shows graphically that the water was not saturated with hydrogen until a pressure of approximately 300 psig was obtained. At this point a definite pressure rebound, characteristic of saturation, was observed. 1 l The fact that the solution was not saturated at 1000 psig was due to some procedural difficulties experienced in setting up for this first test. The data shcw, however, that in accordance with the laboratory pressure drop tests, the pressure at which saturation occurs can be determined by watching for precsure rebound af t6r an aliquot is extracted. t

4- -m J--A e-s i I X "TCA W31YM W31SAS/70A 37ESnS N3SCWCAH ..t .E N r. w w e N e v 3 o m 3 5-u ,w c ~ N 5 ( O C C =' H N m W r- ~ m y f a 1- $] hk E 5$ MM td o z - P* M f. g-q 5 c .g o Q C w ~ ( gg 8, o< ( au om -c v

. E.

~ 4 O N h 5 e s C Q g Q N m T M N 2 SISd '3 MOSS 3Ed 73SS3A 70 -

The hydrogen bubble volume to system water volume ratio corresponded to 0.4 percent at 300 psig. By dropping the pressure further (to 100 psig) the bubble volume grew to 6.5%. (Note that hydrogen bubble volume is inferred frcm measurements of water removal). No correction has been made in the calcula-tion for volume contraction of the pressure vessels as pressure is decreased. This effect will ae examined in Test G. 8.4.3 Test B - Acril 6, 1979 ( Test B was performed similarly to Test A with the exception that the core simulation heater was controlled via a variable power transformer rather than an on-off switch. As a result, the pressure ripple due to heating variation was no longer in evidence. Step changes in vessel pressure were again caused by taking incremental volumes of water from the base of the reactor. The plot of Test B (Figure 22) shows pressure rebound after the first increment of water was removed. This indicates that hydrogen saturation was achieved at a pressure in excess of 1000 psig. Accordingly, the fraction of bubble volume to water system l volume was much greater than in Test A. At 300 psig, fer instance, the bubble volume comprised 9.55% of the total system as compared with 0.4% at the same pressure in Test A. At the conclusion of the test, the size of.the bubble in l the steam generator was measured by exhausting the gas through a heat exchanger and measuring the volume by displacement of water. At recm temperature and atmospheric pressure, reabsorptien of l l :

i 1 ) i REACTOR SIMJI.ATION TEST-B 4/8n9 1200 12, !!00 11 J 1000 la 7 F w e 900 F H 9 3 Q 8 6 ( 800 F 700 F 7$ ] 7) M N $600 F 6a l ~~ E r 500 J r-5W W J / m D a a D 400 4 388 3Z W Q 200 l 2 m I Q y g' 100 - 1 I I 0 B 0 20 4 68 80 100 120 14 ISO 180 i n m urEs %""] Figure 22 Plot of Hydrogen Bubble Growth vs. Pressure, Test 3 - l 1

i hydrogen in the displacement water was assumed minimal. The bubble at the top of the steam generator was obtained by maintaining the final pressure with the pressuri=er valved into the system. The valve at the top of the steam generator was then opened slightly and the fluid passed through the heat exchanger. When liquid was obtained, sampling was discountinued. The gas volume in the steam generator adjusted to 280 F and 300 psig, conditions which existed at the end of the test, was very small compared to the hydrogen bubble contained in the total system. Of the total hydrogen bubble, 96% was in the reactor and 4% in the steam generator.. Although the steam generator is physically higher than the reactor, there are three effects that may contribute to the placement of the major portion of the bubble in the: reactor. These effects are related to (1) localized pressure. drop in the reactor, (2) preferential removal of hydrogen in the reactor due to heat effects at the core simulator heater, and \\- (3) sweeping action of the water flow which convects hydrogen bubbles from the steam generator to the reactor where they collect in the upper half volume above the exit port. Figure 23 shows the relationship of the pressure re-bound verses the target pressure for water saturated with hydrogen. This figure compares to Figure 12 in section 5.0. During a depressuri:ation period, if the water is saturated, as the pressure is reduced an increased amount of hydrogen will evolve thus resulting in a pressure rebound.. y

s m REACTOR SINULATION TEST-B 4/8/79 188 ..............,... 3....,...................,.............. 9 0 88 f 88 (' 78 0 00 0 n N L A i 0 A u N n n 0 I 48 h ~ t f v 0 0 3 0 t t 0 28 f Ig ~ g........i................i....i....v.........i....i....- 8 188 288 388 48 588 688 780 888 S08 8 180 288 PRESSURE AT START OF STEP, P1, CPSIO 1M!hrX25] I Figure 23 Relationship of Pressure Rebound vs. Target Pressure for Water Saturated with Hydrogen l - J

m,- ;., ,i 1-- h n; r s x 3 ) a i ~ i 8.4.4 5'est C T w w i' 1 . l1.. A' delta 2 me?.ar-was suppfied by E.G.&G. with the intent. \\ .w ,w of measuring the chan'cje,,in. pressure differential as additional x data for ensuing tests.- Th's met'e'r su,nplied was identical to ,s r themeterusedintheHahisburg, Three Mile Island Reactor, t q and was therefore oversized for the simulated reactor system. ,,, s Measurements taken for the purpose of calibration of the meter revealed that the meter would not meet the requirements i of the simulator tests and the calibration test was terminated. .w. 8.4.5 Test D - April 9, 1979 Test D was similar to test B with the exception that the ' water withdrawal rate was continaous instead of periodic. Although the water had' been previously saturated with hydrogen, the' hydrogen saturation pressure had decreased prior to start of test, as is evidenced by the data plot (see Figure 24).. Bubble gel:rathwasonly3%ofthesystemwater' volume. \\ 8.4.6 Test E - Acril 9, 1979 The continuous sample method of ' Test D was repea ed in Test E af ter re-establishing hydrogen saturation at 1000 psig. Figure 25 shows a much more dramatic growth of bubble, size i T t with pressure decrease. a As in all previous tests, the water temperature was held at 280 F during depressurization. 8.4.7 Test ? - April 9, 1979

Test F was a repeat of test E with continuous water removal.

The only change made was that the water was drawn J

\\ REACTOR SDH.ATION TEST-0 4/9/79 1200 12 N III " I 1100 1I J O la 1000 W J em 9 m H m ( n 3 Q-800 8 t W W 2 700 7 F 3 m e W 600 6 s e a Q. O

500 5

W W M J o 40 4 m W m ( 3 300 S -3 Z W 200 ~ 2e 0 g 100 I 0 8 0 2 4 6 8 10 12 14 16 18 20 TIME,MI.NijiES Figure 24 Eydrogen Subble volu:ne as a Function of Pressure, Test D

\\ i 3 REACTOR SIMULATION TEST-E 4/9/79 1200 12 1100 N38Nl 1! j 0 1000 10 g W F e WB 9 H 3 ( m 'M 8 6 r W m o'. 700 7 J m M N m J w 50 6 a E Q. 500 5 W J J w m m m mG 4 3 U m ( 300 3 6 e O 200 2 2 0 3 I 100 1 0' 0-0 10 20 30 40 50 60 70 80 90 100 IIME, MINUTES Figure 25 Hydrogen Bubble volume vs. Pressure, T eS t, T. . G

I from a tap at the base of the pressurizer instead of at the base of the reactor vessel. This change was made to more closely simulate the withdrawal of water from the pretsurizer and also to see if a greater percentage of the hydrogen bubble would form in the steam generator. Results were similar to test I. A slightly greater hydrogen bubble was created in the system at 300 psig in test E than in test F (See Figure 26). The amount of hydrogen volume present in the steam generator was still small in comparison 1 to the total bubble size indicating that the location of the water withdrawal tap was not a significant factor. t 8.4.8 Test G Part 1: Using unsaturated water the system was brought up to temperature while measuring delta p. Pressurization above the boiling point was accomplished by applying hydrogen to the pressurizer. It was assumed that the surface area of f~ ( contact would E' sufficiently amall and the volume flow from x the pressurize

• sufficiently low that the amount of hydrogen going into solution would be slight.

? Part 2: The system was depressurized as in Tests D, E, and F. This run established the system volume decrease I with pressare reduction as the vessels contract. Results of test G are plotted in Figure 30, in which the relationship of tests A, B, E, F, G, and I is shown. Measurement of delta p was taken in both parts. The peak to peak noise level of 'the delta p signal was noted on a strip chart recorder. No difference in noise level was observed. 4 l -

l. REACTOR SI K ATION TEST-F 4/9n 9 ~ i 1200 12 s.- !!00 - 11 a 0 !000 10 g W l .F e 900 9 H 2 m 800 8[ t s F W m E 700 J 7> m m s m J e 600 6o 1 + e 90 i J 5W J l W m m a mM 4J (j W m 300 3j e 0 200 22 a I 100 1 0-' 0-0 10 20 30 40 50 60 70 80 90 100 TIME,FINUTES rigure 2s ayeragen sebble vo_=a vs. Pressu e, Test 7

F 1 DELTA P VS VATE TSP F, TiST-I 4/lln9 l 5000 i 4500 l .~ .V, i, ~ 4000 M F J500 ( J 0 l H 3000 3 a i H I 2500 s Q. <2000 g J W O 1500 N 1000. r 500 Y r L1:2 ilnCS) .,,,i.,,.i,,,,i,,,,' 0 0 58 188 158 200 250 380 l IBPEAIl)Ii5, f Figure 27 Di"e er.tial Pressure as a Fur.cticr. of "'er.perature, r ! t

EACTM SIMILATIM TEST-I 4/11n9 5000 g i ) 1 4500 g DE;.,TA P t 1M 4000

1. 8 m

( 52500,- 358, o [- g H j3000 [ gS H m -3 I g '2500 g a' 6 L 2000 "I.'4PERATURI g J W p h

gg 0 1500

( 1000 jgg 500 - LtslLitxts] g f i i i i i i O 0 0 20 W 68 80 100 120 14 168 180 200 IIME, MINIJTES Figure 2 8 Differential Pressure as a Function of Te=cerature and ..me, Test I

d6 REACTORSIMllt.ATIONTEST-I4/11n9 1200

12 1100 11 D

120 P 10 T = 130 F g 9 0 S00 H g ( M 3 ' %8 8 e s r W W 2 700 7 r 3 pressure rebound observed as P passed y W NO" through the saturation pressure. 6 g Water withdrawal rate was constant. E g L J Q 500 =I=' 5 __.J .W y $400 4 d W e ( 3 300 3 0 z W i 200 2 e O z 100 1 0 I 0 0-0 2 4 6 8 10 12 14 16 18 20 IEE, GUIE3 Figure 29 Eyd:cgen Subble VoltNe~ as a Function of Pressure, Test. I

j 8.5 Interrelationshin of Tests A, B, E, F, G and I Figure 30 shows the pressure vs. bubble volume relationship for tests A, B, E, F, and I. Test G is also shown, which indicates volume correction to be made for vessel system contraction with pressure redaction. Figure 31 presents a semi-logarithmic plot of vessel pressure drop versus hydrogen bubble size for tests A, B, E and F. The slope of each line is representative of the rate k' of bubble growth with pressure drop. The intercept at the ordinate roughly estimates the hydrogen saturation pressure. The bubble for test I is obviously smaller than would be expected in comparison with tests B, E, and F and the rate of bubble growth is lesser. The reasons for this include the following: (a) The starting saturation pressure was less than for - tests B, E, and F. 2 k (b) Decreasing the temperature while holding the pressure constant required that approximately 6% additional water be added to the system from the pressurizer to compensate for the water volume decrease with temperature decrease. The additional water was unsaturated and therefore decreased the saturation i pressure of the system. (c) When the pressure was dropped it was at a constant temperature o,f 130 F. The size of the hydrogen bubble fo=ned was smaller for a given amount of hydrogen leaving the solution because of the lowered )

H2 BUBBLE GETli 5Illi P DROP AT C2ST T 12 . x J 11 o E Ltstillrics1 y g tg TESTS A,B,E,F f Fy T=288F <g 3 2 ( I w 8 F e TESTI:T=138F 7 A N 1 J a 6 j 5 I e e J 4 e i Z 3 w e O z 2 o .y I t-G 0 B 200 40 680 800 1000 1200 YESSE. PRESSURE,PSIG Figureb0 Cce.posite Plot of Tests A, 3, E, F, G, I Showing Bubble Growth as a Function of Pressure -

H2 BUSBLE GROWTH WITH P DROP AT CONST T I i 1 1 i i i i 1000 O 500 - H E N i g -..~~ - w F ( --CE - I g D ,,g m m ss w i N ccc. A 10 0 - _a w m M 50 TEST I: T = 1SOF 5 ~~i~ _ ~ ~ = TESTS A, B, E, F - - ~ ~ _ T = 280F t Ltdillinos1 I I I I I I I I 10 O l 2 3 4 5 6 7 8 9 10 11 ~~ : _ HYDROGEN SUSSLE VOL/ SYSTEM WATER'VOL, Ji ~ s.. Figure 31 Composite Plot on Semi Log Paper of Bubble Growth as a Function of Pressure, Tests A, 3, E, F, I l

temperature of the gas bubble. Figure 30 shows that tests B, E, and F had roughly the same saturation pressure and that even though 3 was a stepwise depressurization test while I and F were continuous, the rate and extent of bubble formation was comparable. Water was withdrawn from the system from the base of the reactor simulator in tests B and E, whereas water withdrawal was from a connection below the pressurizer in test F. The difference in withdrawal ports apparently has little effect upon bubble formation. ( Test A had a beginning saturation pressure of approximately 0 300 psig at 280 F, yet exhibited the same rate of hydrogen bubble growth as tests 3, E, and F. Test I had a saturation pressure of approximately 700 psig at 130 F and demonstrated a significantly lower rate of bubble growth with depressurization. The difference of bubble growth rate can be attributed to the difference in depressurization temperatures. -The advisability of first reducing temperature and then reducing pressure for ( minimum bubble growth is demonstrated. Test G (See Figure 30) shows the extent of apparent hydrogen bubble growth for the system upon depressurization. This apparent bubble growth is in actuality the volume decrease of the system seen as a volume of water that is taken from the system as the vessel contracts with decreasing pressure. Points along line G in figure 29, for any given pressure, should be subtracted from the indicated bubble growth for a given test as compensation for vessel contraction. 87 - (

Delta P measurements revealed changes with temperature but no change in either level or peak to peak noise with change in hydrogen content. 8,6 observation Durine Set-up and Procedures During the final procedural steps before beginning Test A, a fortuitous " accident" occurred. The information obtained is important enough to be included here. The sequence of events is recounted below. No data were being taken at the time. 1. Hydrogen was being bled through the system from below and bled out the top of the reactor vessel. 2. The pressure was near 1100 psig and the temperature was in the range of 110 to 115 C and increasing preparatory to test. 3. The pressurizer was connected to the system so that the system was " soft" with a gas volume in the pressurizer. No gas volume was being maintained I k in the reactor si=ulator. 4. Workmen were installing another heating element on ,the outside of the water circulation leg (HC in figure

15) when the system pressure began to rise out of control.

The hydrogen supply was turned off. 5. Pressure continued to rise and the sight gauge showed the appearcnce of a hydrogen bubble in the reactor vessel that was growing in size. System pressure had reached 1400 psig and the liquid level indicator was f alling rapidly. _ 38 - O +

i 6. The vant was opened at the tcp of the steam generator. This caused the bubble to grow more rapidly, and the pressure would rebound whenever venting was stepped. 7. The heater was turned off. This pemnitted a stable i pressure to be reached at about 1416 psig, but the bubble remained. 8 Some venting from the top of the bubble was tried. ~ ) The pressure dropped during venting but rebounded when [ the valve was shut. s 9. A workman noticed that he had no power to his drill l and it was found that the power plug to the circulation pu=p had been inadvertently pulled from the wall. 10. Water circulation was reestablished. l 11. As the system temperature was observed to be falling, j the heater was again turned on. When the auxilliary heater was ready, it too was turned on. 12. Venting centinued periodically to hold the pressure ( in the 1100 and 1200 psig range. 13. Heating with circulation caused the bubble to stabili=e. Continued rise in temperature caused the bubble to decrease in size and to disappear eventually. 14. Hydrogen flow through the system was reestablished. 15. Bleed flow of hydrogen ecntinued for approximately another two hours at which time the test temperature was obtained. The test was then carried out under the false assumption that saturatica had been reachieved.. _ - _ _ _ _ _ _ _ _ _

8.7 Appendix INSTRUMCCATION LIST 1. Bailey Differential Pressure Transducer - Type BQ Accuracy: 0.25% of span + 0.01% of upper range limit per OF. Less than.005% per volt change of power supply. 2. Sargent Welch Strip Chart Recorder - Model XKR Accuracy: 0.5% of full scale. 3. Billings A to D board (' Accuracy: + LSB, approx. 0.025% of full scale reading. 4 Monitor Labs Data Logger - Model 9300 Accuracy: For temperature measurement -.1 + 0.54 F or 0.30C compensation error assuming no temperature gradient on the iso-thermal block (connection block). For voltage measurements - 0.055% total error. 5. Bourns Indicating Pressure Transducer - Model 2053126050 Accuracy: + 3.5% of full scale. 6. Bourns PSIA Transmitter - Model 2900 (' Accuracy: 0.5% of full scale (total non-linearity & hysteresis). 7. Omega J & K Type Thermoccuples - ANSI Standard l Accuracy: 4 F. conformity error in the range of 0 32 - 530 F Instrumentation - Calibration Prior to running any tests on the simulated reactor apparatus the various instruments were calibrated and checked for accuracy. Shown below is the calibration of the pressure transducers. l 90 -

~ PSIG = A V +B g 9 PSIA = A, V, + B, A = 648.789 psig/ volt A = 612.595 psia / volt B = 85.0 psig a,= +10.86 g 1200 D 4#D Dead Weight Tester -i-PSIG gage g 1000 O PSIA gage n PS mm 800 g (. Oz m_.. 600-PSIG g az 3 'A Ey 400-g

• Pressure Transducers.

for Reactor h Simulation Tests 200 I .o_ y 200 400 600 800 1000 1200 1400 1600 1800 2000 t Figure 32 Calibration of Pressure Transducer 91 -

- ts - z z z o v e o xn m o a M N O N oc a 0 m x o-e o a n n e ? a~ 2 ? a e 8 M a o o = n o aw e a a u o us o e e aU Q W M M C 4 C ~3 o a we m w o = c o> = m e2 % u m o - z o a P4 W =* G 4 o L uO 03 a oC \\ w o a= c o o ao .c u o a r c - * = n c. c. m w 2 wo a s c

• 4 M

= = 0 'e v3 2 < c = c 6 oa = 8 I i i i I i l t 1 i oc ~ 3

• c k 3 03 o c:

4N bu O oo r-e a t A e. a 3L w-a m o y s -, a ( @T h Increase Pressure ga M P wY @:k 18 e 4 o o KI c 4 ga v i I:xtracted irater de:m to 300 psi bg Hydrogen injection to vessel 3 y_ g p __ _ 0 ( EwitWaw water her.enth eressuriser u u e a N o p Sm$ 1 ,i 4 m a-m- 1" 4.25 Pressure Increase to 1100 psia i =o e a= A ( l 3* Temp increase to 280 F and l wu Pressure i= crease oo u.a on z2 h i 3e agrees 40 psi 240 F Flash off dissolved gases I e i r i e i k_.. l A O I eE on M ea " Sus o $8 h 2.54 Time 208 ? Record.1 cm/ min a w M o ci .J O a ~ h S i I ) c c o o e c' c e e O C O C C O s .C .C. = 0 N c5 u)

• =

N e m s. w

a - C6 - - x I t 8 i 3 _. __ *[i , 0 _ U I Ei. a u 0 Start pressurizing { 's, E.s, g

• I again with H 2

.e -- u o= o a oc 1 C 4 0 N g xv: s e o e

n as s e e

o o~ c c - sj m aoau m e L oa M C "e Q QM C a euu a 6 PC:P off L c E.E $ e n % - w = at ea o P==P ors _ 2v253 So N mo e e e l m o = 8 e v N sn ~ ~ j p3 i i o Stcp, bubble gene 277.4 F 493. psig g C o Stai: : to 400 again a w .g- ,u a sr o 1

• End bottle e=pty g%

w e 1 r 512# 277.5 F i e = w e x k e o a: o o o = e 3.y ~ a a o aj, 3 - e o 4 3 6 Start to 400 ,y. 3 3 h e i " !S e 591.3 277.5 F End cf 600 3T-e u 4 > C O start to 600 psig ya x a a : a w o e o 800 psig j @ @ j h, i i g 03. W Z t. Withdraw gas hebble and Measure Start E T**t 2 \\ ie o u "C a C C C C $ 'c C C C C C e c - e = e = c o e j o CD C v N C C3 v N C i N m m m m m I ( l I 1 l l l l i 3 e o c E m 6 N E c c m N C N 4 C = =d 8 c 2 0 E 4 - = c-m = =J c'2 t -e e m C k 1 = C C.."a M -e = = C C1 O O c 0 = = N

== 1 C 03 h O T 2 3 LQ Q w C T R e M 4 C 4 m 5 ) e a w c e w o = o o G U1 4 3 \\ k D 0 m E 8 -.A C 4 h

==4 L C U m E* O O w 23 m O N. W = m C C m> ~ L 0 M i e m : ll 1 L m u c N d3 E S T 3 "o 9 e

l F l Page 3 of 3 2000 mv- _,2000 i t 1000 - -. ~. 100 4 - 10 cm

1600,

.10 cm/ min $ 10.0 cm/ min 160 JL Temp = 200 F ]400 .Y 140 gg /Y&%wrovyWM#v4poH~swhr u rs s ~^.**."- >< vW '+% 3~ Mv-6 1200 .,yy; ~ $ $ f$. so a ts. M 1000 $[' Estimate near lim saturation wille II an e ~ 2 100 Same signal as for pure y p = 1101 psiti ' water. 000 80 We may get somettilng wisen 11 comes liack out of solution. l 600 60 Regin to reapply Hext tent in to drop temp. laydrogen to water 3 o Time scale ) Time scale = Time scale = to 13 P f rom 200 F.... 'I1mn I 10.0 cm/ min wo ill drop pressure from 40 1.0 cm/ min = 0.0 cm/ min' 1050 psig in a continuous mannier. 200 20 0 _ 0 1 April 11, 1970 7:30 [mi HST i 3 i e i 8 i .1 -, i ? _j __. w -.. -.~1 e

e 8 I I l I 3, 1 i g - 20 d 55:40 247.4 o o o e 3 e o o e ~ c' e o s_ e .e e o e o C e o -252.4m c. v s - n m I e-6 ~~ Start te d:cp Te=p @ 20:33 1, {~ ~I l l u-s 1 Completed removal of Pressugi=ed Sc.me cold water ficwed in-l bubble from reae.ucr 278.0 F Should have been close to 2.8 1 since 2.8 1 had been ] J ~ removed frem the system to form the bubbles. h I I I I I i 3 l' o' ~~' o C oC-19 :10 : C 40 C C C o C C C C C-C C C C C .C.D. T N 4 .C CD v N f l Begin removal.of reactor -277.8 F remove reactor bubble airs / pump on~ o ~~-' bubbleS'. G~ bubble l removed h~ eater at pump on SG Bubble - recoved .me Step Pump and Heater l { 18:23:17:8 a:: = 5 7/8 iP =0 ~ st stop - remove bubble frem stream ge' [ Pressurize.into sfstem w 1099 psig. Subble n i-vented frcm tcp. Water displacement meas- .7 lurementofH.,pu=pneededtobeofftoallow 'ascesstobu3ble,consequentlyheaterhadj I 8 I L to be off to prevent E. boil cut in reactmon. - -e o e e o e o e o e el C O C a C C C C C C a w w <rm .o o a w n \\ -g m a t T L i 3[ E [ - 17:15:16 ~ n 1 _h I 4 Y =r l Water saturated with Hydrogen. Synthetic bubble generated at 280 ? by a Y = 100 =v/cm l remov hg l W d frem system at pressurider t I in previcus test. I'.i previcus test H2)I Temp = I"- was re=cved and measured (displace = enc 279.4 F - fr Since bubbles were -- l & stable f re=em pressuricer). cved consecutively'and hydrogen for ~ g this test was added at base of reacter, g 0.1~cm/=.in the bubble space is pred--4.antly 1. 4 l the reacter. i E 5 5 5 5 5 v e e v n C e e v n N m m m a-( m , 5 : s.., :24.0 April 11, 1979 Test I 10 cm/ min Tems d:ce test Part 1. - l

L A N v g; C N l 1 I I I l 1 I I I I Om*o o e C C C C C C C e o o o o o o o o o C o a w w n -o o a w n ce I Denisty change because of tem. diff. i : 1;:=p (gi:: + (K+1)v') Kisfunctionofj , Clearly,Kgaschangedwithtemp. o & g If K p e constant with temp ch L 2 v,0280 F .01726 then 2 2 -.1530 - = ,@l29o.,- .01625- =.3.06215 = a v =.12143-21.4% = = 1 p ". 40 6_.22%. 1 1 - 1260 Shut off H O frem system 10 liter water withdrawn 129.7 F 2 6 liter g l 23:18:26 257.1 psig ]' 1 liter 1 l pressuriser shut of"f frem Start pressure d:cp test 130.3 Fo system a 23:8:4.7 1058 psig 4 cn t know why a ] Test I Part 2 g i g i m'" *== ^ l [ 8 8 _-22:35 40 -151.3$F ~ ] -1 e e e 8 5 o 5 o c> 8.n.i o n o u o a e. w. -22:25:40 162.4 F l .ened pressuri er to system 22:9:04 180.1 F ji, - l { .illing with Water leasing H O frcm pressurizer 2 f L' ressurizer off 22:00:55 182.0 F l ~ l l 01:35:40 214.1 F Pressurizer pressured and cpen to system 230.9 F ~~ k Started filling with water ~ - Pressure off pressuriser 21:12 x 232.4 F ~ O s 1 4 1 I i 1 o i i f 20:55:40 247.4 ? 8 8 8 8 E e e e c m-e w N 262.4 e c o = a w ~ 9, g fStarttodropTe=p920:33 '[ i l t I l l l i { . - 278'T ~Pressurised. Sc=e cold water flcwed in C=mpleted re=cval of \\ - Shouldhavebeenc1cseto2.81Since2.8]> I bubble frc= reacter Y liters had been re=cved from the system ~ i _tc fers the bubbles. l l g 5 t j ~ l i I i I ~ 19: 10:40 I I I 6 e o e = c c c 0 c C C C e 5 c = a o e e e o = y m .c a w n ~ ~ de chan&s with temperat*:re. No cbserved signal change with \\ w 1

9.0 KYDROGEN PEROXIDE TESTS 9.1 objective The objective of this experiment was to examine the feasibility of introducing hydrogen peroxide into an experimental apparatus which, upon dacompositien, would react with the hydrogen in the test cell and thus reduce the pressure of the system. \\ 9.2 Introduction t I Tests were performed by introducing hydrogen peroxide into hydrogen saturated water at 1000 psig and 230 7.~ It was felt that the oxygen resulting from the decomposition of the peroxide would possibly react with the hydrogen to - produce water, thus reducing the reactor pressure. Theoreti-cally this reaction may proceed without the use of a catalyst, but a catalyst may increase the rate at which reaction' occurs. Two tests were performed (noted as Test A and Test B / ( in the text). In Test A, no catalyst was used, and in Test B a platinum catalyst was used (5 grams of catalytic pellets impregnated with Pt DNS plating solucion and heated at 600 F). 9.3 Experimental Apparatus The experimental apparatus used for the hydrogen peroxide test is illustrated in Figure 33. The instrumentation and data system used is shown in Figure 34. The Billings Energy Corporation's Autchydrider was used because of its heat controlled, instrumented sample chainber. -

9. 3 Exterimental Accaratus Hydrogen Peroxide Test Apparatus y+

(X)v-' (X)v-. N ~N ) h Vesseli Vessel 11 PG X v-= V-2 O W m si i t (X) U \\\\ tesillincs1 Figure 33 'dydregen Per0xide Test Apparatus e 4,

--s-M~ -]g-fhT-.%sa mpfr**!**J+.-- 1, g,- - a l F usammasss==mamess.......... Oh f4 i - BE55555Damassassagamussamma O EERBassassamanamanasamassaa L ' y3s% ~~ 1 55555555555555533333553a555 " MJ 3'3 M

• Essassammaassuasassasssasan ;,j'e manunusmamanasammuassessaas

- w{ f 9......Ly . f.y.,2 ~.~~ .a. o: w-

3....%-.

A W~,.W*e.,; f' j jQ' y-.,, .W fY i V.i *}.*E 'W&.~t -d -e ? e r I';g g%g$ y_' " a. ~ h ,,~M ~ W>* E $ &am-M 7 4-hh. t l y-p 5.-- py..pg 1 cy. -3 ;c ~ oy aw ..;ww y -P n @&.,. . &. m b _ ~ y v %^ ._M: S- -Glf-. ks a p

1. @3

~, ~g~~-:9 CD# ;Qp,p.,.,* g 6]4 ,g i+i Cf} / p-AQM ..gg

• g:

. *Y=. < 3,3. %g g J 4 e7 *[t".ny7 _ H 5::2 [ *. f; j QI _.,.l m$ /d .v$. p.:.Y,.i. W ?Y ibN h ~ F: -A a* bA I1 Q. y. n- ~ f* 2 2 e W.a4% ggLA T.*C%S*4.-A,83.yll?'- ph.p'

• T3

, 4Mhi ' & M.4s. G B. W z fy*::w 25 % b E -a

w. wmr--.-.

m -s. -. oL4 3 .,_ g p .s -n ' gE#' ^, : - - [g g4, pq ~ [-- [ 4D \\' v T O M. _.., - 'g('S-1 4J dW % c C ,~, c o o%g gbyw w2 \\ (h' lbp3? g G % ~ q* q n'.M) g% = w% %g% 2 .4 'd. w' Q p ..a h 5 = w.;.3-4 t 1 d 98..%, %. w- ,%.,r.,s.d.&.,c.n*. ' J Akz".Ipi Else W u. \\ .. ew ..:7 :-r-;:-.-q P14PJdG f$$. & ) h h.W$$$ -r PEk..a aWMd Fig.re 34 F.ydrogen Peroxide Test Instr.. :enza:icn

9.4 Experimental Procedure 9.4.1 Test A The uncatalyzed hydrogen peroxide test was performed according to the following procedura: 1. Till vessel I with boiled, deionized water. 2. Pressurize Vessel I to 1000 psig with hydrogen. 3. Remove 50 ml of water through Valve 2. 4. Pressurize to =750 psig with hydrogen. [q,. ' 5. Periodically heat and shake Vessel I adding hydrogen to maintain pressure at 1000 psig and temperature at 280 F. 6. Place in insulated heated chamber with digital temperature readout. Heat to =.80 F. 7. Add 100 ml =10% hydrogen peroxide to Vessel II and pressurize with argon. 8. Connect vessel II to vessel I as indicated in p' - apparatus drawing. 9. Open valves 2 and 3 with Valve 4 open and connected to argon at 1050 psig. 10. Reduce pressure in vessel I to 925 psig by opening needle valve (Valve 1) thus allowing some peroxide solution to enter vessel I through Valve 2. 11. Close all valves. 12. Record pressure and temperature measurements as a function of time beginning i= mediately after addition - 100 - m.

of hydrogen peroxide. 13. At conclusion deter =ine volume of unused peroxide solutior.. 9.4.2 Test B The catalyzed hydrogen peroxide test was performed according to the following procedure: 1. Fill Vessel I with boiled, deionized water. Add 5 gm. Pt treated catalytic pellets. j{ 2. Pressurize vessel I to 1000 psig with hydrogen. 3. Remove 50 ml of water through Valve 2. '4. Pressurize to =750 psig with hydrogen. 5. Periodically heat and shake vessel I adding hydrogen to maintain pressure at 1000 psig and temperature U at 280 r. 6. Place in insulated heated chamber with digital temperature readout. Heat to =280 F. _ gj 7. Add 100 ml =10% hydrogen peroxide to Vessel II and pressurize with argon. 8. Connect Vessel II to Vessel I as indicated in apparatus drawing. ~ 9. Open valves 2 and 3 with valve 4 open and connected to argon at 1050 psig. 10. Reduce pressure in Vessel I to 810 psig by opening needle valve (Valve 1) thus allowing some peroxide solution to enter vessel I through Valve 2. 101 - C ^ - ~, s .s}}