ML19316A488
| ML19316A488 | |
| Person / Time | |
|---|---|
| Site: | Oconee  |
| Issue date: | 06/03/1970 |
| From: | Cady K US ATOMIC ENERGY COMMISSION (AEC) |
| To: | Long C US ATOMIC ENERGY COMMISSION (AEC) |
| References | |
| NUDOCS 7912200742 | |
| Download: ML19316A488 (8) | |
Text
{{#Wiki_filter:.Cer.Jht n-2 6 7 '"T '? Unit D 07ATc3 ~ /: f.'. ATCMiC ENZ.7GY COMMiS3MN .,_ L -) 9{' wars t. :GTON. D.C. 20:43 ,6 %:s vl ~ ~ W .. Lu N:EP June 3, 1970 ]!' Charles G. Lon;;, Chief, PWR Project 3rcnch 2 Div..sion of Reactor Licensing STZl01 LINE RUP'IUZE ACCIDENT D'JE-CCONCE (DOCKETS 30-269, 30-270 AND 50-287) We have revicued the steam-line-rupture accident for the B&U once-thru steam generators. Accentence Criteria The applicant has proposed in the FSAR that three criteria be met: a. The core will remain intact for of fective core cooling assuming minimum tripped rod worth with a stuch rod. b. No steam tube loss of primary boundarv integrity will occur due to less of secondary side pressure and resultant temperature gradients. c. Desee '-ill be ithin 10 CFR 100 limits. We agree with these criteria, but believe that they must be met assuming that non-safety systems, such as the integrated centrol system and the operator, either f ail to act or act in an adverse manne r. If criterion b on tube rupture cannot be demonstrated, then we believe a much more strin-gent criterion, such as no DNB, must be act. We have several serious concerns which must be resolved: a) Will the turbine stcp valves trip as needed to isolate the unaffected steam generator? b) Can the integrated control system, a non-protection system, function to isolate feed water flow? c) Can the operator be relied on to keep the integrated control system from reopening the feed water valves to avoid return to criticality? -7912200~7 G v._.
Charles G. Long 2 June 3, 1970 In order to resolve cur concerns, we have asked to applicant to re-do his analysis by responding to nine questions, 14.3.1-14.3.9, appearing in Director Morris' letter to Duke Power Company, March 3,1970. Amendment 12 to the FSAR contains answers to questions 14.3.1 and 14.3.2 on the design and operatien of the turbine stop valves. The valves seen to be a globe or disc valve which may not function properly against reverse flow. The applicant was not responsive to this part of question 14.3.1. The attached discussion of the detailed behavior of the B&W design under a steam-line-rupture accident points out the specife problems which need to be resolved. Hopefully the applicant will be fully responsive to the questions and the results vill =cet the acceptance criteria. / N K. B. Cady PWR Project Branch 2 Division of Reactor Licensing Enclosure cc: R. C. DeYcung A. Schwencer D. F. Ross R. R. Powell Docket (3) PWR2 Reading D**3D ~*]D TY& 3 e w j\\\\ c J\\L S . :1=
a DUKE STEAM-LINE-RUPTURE ACCIDENT i I. ' Refe rences I 1) Final Safety Analysis Report, Duke Power Company, Oconee Nuclear Station Units 1, 2 and 3,thru Amendment 12. 2) Once-Through Steam Generator Research and Development Report BAW-10002 Topical Report (p roprie tary, Augus t 1969). 3) Requirement for Additional Information on Safety Analyses for the.Oconee POL Review, Memo to RSEoyd f rcm DFRoss and KBCady (January 30, 1970). 4) Meeting with Duke Pcuer Company on Oconee Nuclear Station (November 13, 1969), Minutes of Meeting, ASchwencer to RSBoyd (Decembe r 12, 1969). 5) Meeting with Duke Power Company en Oconee Nuclear Station (January 21-22, 1970) Minutes of Meeting, ASchwencer to RSBoyd (February 3,1970). II. Inventore Comnarison Time tuojur dif fetettee between the B&W unee-Linuugh s team generaLor and' the U-tube design used by other NSS's is the water inventory. TheLB&W once-through steam generator has its largest water inventory-at full pcwer, while a typical U-tube steam generator has its caxi-mum inventory at zero power. B&W prasented the following table:5) i 3. TABLE 1 } COMPARISON OF OCONEE 5 A U-TUBE PLANT ' Oconee U-Tube Plant 3) 3 Reactor Coolant Volume (f t ) 11,478 9,088 One SG Inventory at 100% power' (lb). 62,600 84,586 One SG Inventory at 0% power (lb) 20,000 160,000 EOL Mod. Coeff. (ok/k)/*F. -3 x 10-4 -3.5 x 10-4 a) The numbers seem to come 'from the Westinghouse Turkey Point, three-loop' plant and the-160,000 lb inventory seems to be ' based on-a cold water density. l l .,q
2 4 Making a.hcat balance for the two cases where the potential cool-down is largest (100% power for Oconce, 0% pcuer for the U-tube . plant) B&W ccmputes a potential cooldewn reactivity increase of 2% for Oconee and 8% for the U-tube plant. This approximate analysis is based on the assumption that the cooldown is limited to one steam generator volume, and can therefore only be taken as a crude, relative conparison of the types of plants. It does, however, show that with effective feed water isolation, the return to criticality is aggravated for an accident occurring at 100% pcuer for Oconee and for an accident occurring at zero power for a U-tube plant. Without ef fective feed water isolation, the situa-tion is not clear. III. Full Power Accident Description B&W through the applicant has presented partial results } for an 1 accident occurring at EOL when the moderator coefficient is most negative, and full power when the ateam generator inventory is largest. They use a " minimum" tripped rod worth of 3.46%ak/k (a little more than the power defect of 2.3%2k/k and a hot shutdown margin of 1%ak/k) and the additional assumptien that the reactor trip signal causes a power demand runback in the integrated control nenrem, wnten in turn.Mnc-5 tba feel "eter velves Le isolata the a'f fected steam generator. From an oral presentation by B&W ) this 5 runback takes 25 to 30 seconds af ter turbine trip; however, this is not in agreement with the 10-second closing time (0% FW flow 16 seconds af ter the break) listed in the FSAR on page 14-22 }. In 1 addition the applicant has assumed that the operator overrides the ICS (integrated control system) in less than the 50-second stated time to blow dry the af fected steam generator. The assumed opera-tor action keeps the ICS from reopening the FW valve to the affected steam generator when the water level goes below 2 feet. The apparent sequence of events and results are described below: 1) The reactor is operating at 100% pcwer with an EOL Doppler coefficient of -1.2 x 10-3 ak/k)/*F and an E0L moderator coefficient of -3.0 x 10-4 ((ak/k)/*F. 2) The break occurs as a double-ended rupture of one 34 inch steam line and both steam generators begin to blow down at a steam flow rate of about 400% (cral communication at _ meeting . January 22,-1970) of-the nominal -full power s team flow rate. The feed water flow is assumed constant at 100% until reactor trip occurs. This is only an approximation because many other affects.may occur: a). The decreased turbine header pressure inereases feed water-demand which may cause - the main feed water valves to open . wide r., f, f T
3 b) The increased feed water demand and the decreased aP across the FW valves both may cause the FW pumps to speed up. c) The drop in pressure at the discharge of the main feed wcter pumps causes the auxiliary feed water isolation valves to open supplying uater to the s team generators through the auxiliary feed header. The same signal opens the steam supply to the emergency FU pump turbine (from either the main steam line or the auxiliary s team header). The total result is that the main feed water pumps are supplying feed water through the main 14-inch ring header, and the auxiliary feed water pump (7-1/2% capacity) is supplying 90* F feed water through the auxiliary feed water. See Figure 1, a schematic diagram of the feed water supply sys-tem. d) Even if the integrated control system does not open the main FW valves fully, the feed water flow may go up because of decreased pressure in the steam generator. / 3) The increased steam flow in both steam generators increases heat transfer and causec a cocidown of the reactor conlant. Tnis cooldown causes a rise in reactor power demand (because of the constant T control system) and 'a rise in the actual reactor av neutron power (because of the negative moderator coefficient of reactivity). Depending upon which rises faster, at a 2% mismatch the control rod drive system would either drive the control group in or out. The applicant assumes no control rod motion until reactor neutron power reaches 114% and a trip signal occurs. 4) Reactor neutron power increases to 114% and causes a reactor trip signal at about 6 seconds af ter the break occurs. The curve of neutron power Figure 14-20 shows 114% power at 8 seconds and peak neutron power of 180% at 9 seconds. The application states that there is a 0.3-second delay before rod' start to drop and an additional 1.1 seconds before they are 2/3 inserted. Because the reactivity insertion rate due to reactor cooldown is quite rapid, this 2-second discrepancy may-be of concern. 5) The reactor trip signal causes a trip of the fast closing main turbine stop valves, but we do not have adequate information on the delay time, closing s time. design, or-tes ting of these valves. - These valves serve as isolation valves for the unaffected stes., generator and their performance is crucial in reducing the rapid - cooldown of the primary system.. Assuming failure of one of these valves to. close, others must close against reversed critical flow. e -n
FIGURE 1, FEED.MTER SUPPI.Y SYSTDI - _ _ -(pj--- lL'.S ['{ j - !.~f; l { } -N'.} -l --- ' feed heade r to :.tn:. N "J:.ru } ^ - m:.: : - FW ~' SG "A" N{ aux. startup k caitrol V isol. V l upstrcam h [ downs t ream ',". l bsd startup ),.' f[..,f--ICS startup isol. V l isol. V ? a tcs-x .'g. 'il- ~ d t o t.ia in j S Q := - ---m :.::.p ('p:: cur =.cc. c h'g ' ~ * ;" ring, h ler main FW main FW l SC "A' I "~ isol. V control V l y 1 I 1 main FW ic.':;..a~ 4 : I 7 ,supp1y h, (1 p "A" shown la normal operating mode l 7 15% f'ow thru startup V aux. FW r ~ @ 400* F r-l k 85% flow thru maia V supply 11,180,000 #/hr g*~.~ l '@ 90* F 7-1/2% flow f gqg,[.t l h ,g. g,3 {3 . _... ~ - ~ ~ bll to main s q ring header main FW main EU N' SG "11" O l isol. V control V W upstream Y / downstream N / l jd} )I'd -\\CE l startup Y startup isol. V .1. i isol. V EEo.) iCG - -{i] (if[--lCS 'I q [hc. -h% .;. [. )] { ['. !,- eil cder
- ;carin..
~ i pg startup aux. FU S f,. ,,r> " g control V isol. V b (1 op "B" shown it aux. FN flow mode) Note: On loss of pressure at dLacharge of main FW pumps t"a ICS causes the aux. FU isolhtion valves to open and the downstream startup isolation valves to close. The s ame loss of presnure signal also causes the aux. FW pump:i to start which in turn allows the upstream startup I. solation valvec to close when the ICS has opened the aux. FW isolation valves. The main. FW control valve may ar may not close depending on whether a pouer demand run-back has occurred. On low liquid level, the ICS would cause the main FW control valves to reopen. l -q l
5 6) The turbine trip causes the entraction steam supply to the main feed water valves to decrease and it is possible that the main steam line supply (start-up line) to the FM pumps opens. We do not kocu. 7) The turbine trip signal causes a power demand runback in the integrated control system.0) This in turn reduces faed unter demand to the steam generator subsystem which closes the main feed water valves. A description of this runbach is not included in the application and we do not kncu if the main FW startup valves are also closad. We have conflicting evidence on whether this runback of the valves takes 10 seconds 2) or 25-30 seconds.5) Even more seriously, the closing of the main feed water and startup valves to the affected steam generator is the means of isolating the steam generator and limiting uncontrolled cooldown. Ue suspect the integrated control system does not have the required redundancy and independence for a protection system and we have reservations on relying on an unreviewed, ncn-protection system to provide a safety function. Also the auxiliary FW system would be supplying water to both stecm generators, but the applicant on page 14-22 appears to allow the flow to go to zero at 16 seconds af ter the break. 8) The affected steam generator blows dry about 50 seconds af ter the rupture. As the flashing water lesel decreases below the 2-foot level, the ICS would re-open the FW valve to this steam generator. The applicant assumes the operator has switched the ICS from automatic to manual to keep the feed water valve closed. We are concerned that switching to manual before the ICS has - closed the valve will leave the valve at a partially opened position and he will have to take. additional action to close the valve. He will also have to close the auxiliary FW isolation valve to the affected steam generator. We doubt that the operator has sufficient redundancy and independence to take any safety action within 50 seconds of the time of the rupture. 9) Af ter closing of the main ' feed water valves by the ICS, the unaffected steam generator is isolated and will build up in pressure until-the steam turbine bypass. control valve opens to dump steam to the main condenser. We think the. bypass control valve operates to maintain a fixed turbine header pressure, but we do not have a. complete. description of this system.
- 10) - With proper actuation of the ICS and operator isolation of 'the affected steam generator,.the emergency-FW system supplies ficw to the unaffected steam generator _and the steam bypass' dump to the_ condenser provides heat : transfer for decay heat remova_
f rom the primary 'sys tem. 4 ~ y m. m
6 11) The cooldown of the primary system af ter reactor trip adds reactivity to the reactor, but the reactor only returns to 0.4% ik/k suberitical about 50 seconds af ter the break. We understand ) that the primary sys tem has not depressurized 5 below 1500 and the HP injection sys tem has not actuated. 12) With the break inside containment, the applicant es timates thata the reactor building pressure veuld go to 13 psi, b'ut he has not been conservative in his mass and energy calcula-tions. The building cooling and isolatien system actuates at 4 psi and the building spray system actuates at 10 psi.1) 13) The applicant estimates the environmental effects by assuming a 1 gpm steam generator tube leak for 20 minutes, and 1% detective fuel reds. Most of the activity is iodines in 180 gallons of primary coolant released directly to the atmosphere resulting in an integrated thyroid dose of 0.12 rem at the exclusion area boundary. IV. Results of Other Analyses In a single paragraph (FSAR 14-20) the applicant mentions three other caces ass"-iag this time, no operator action. Few detailc are given. With nominal or minimum shutdown margins assuming a stuck rod the applicant predicts returns to criticality and 25% to 35% rated power. Because of the high peaking factors associated with the stuck-out rod, this may lead to fuel pin or cladding damage. The applicant has not addressed himself to these problems. Similarly, the applicant has not addressed himself to the problem of f ailure of the integrated control system to close the feed water valves. V. Mode [andCode We have a brief description of.the physical model centained in the proprietary, nameless code used by B&W.2),4) It is a hybrid, analog-digital-model of the secondary 'and. primary sys tems and the models, but not the code,.are the basis of B&W's Link training simulator being installed in Lynchburg.4) The model simulates the-steam generator in five regions-conserving mass, energy, and momen-tum and ~ simulates the action of the main feed water and startup. valves, emergency feed water valves, steam bypass and relief valves, and turbine stop valves. The reactor coolant loop simulator includes mass and energy conservation in the pressurizer, a lumped parameter reactor model with - a _ fuel, gap, and cladding region, point reactor-kinetics with Doppler and moderator feedback and two groups of delayed neutrons. Trip logic for high flux and high or low. pressure and transport delays are included. We have no reason to believe the model is inadequate, but we have not been able to evaluate this - because the' code is proprietary. _ B&W has not submitted a descrip-tion giving equations, principal' assumptions,' input parameters, etc.. t t}}