3F0288-02, Forwards Addl Info Addressing NRC Concerns Discussed at 880120 Meeting Re Diesel Generator Loading.Supporting Data Re Pump Performance Also Encl

Forwards Addl Info Addressing NRC Concerns Discussed at 880120 Meeting Re Diesel Generator Loading.Supporting Data Re Pump Performance Also Encl

ML20196A666
Person / Time
Site:	Crystal River
Issue date:	02/01/1988
From:	Eric Simpson FLORIDA POWER CORP.
To:	NRC OFFICE OF ADMINISTRATION & RESOURCES MANAGEMENT (ARM)
References
3F0288-02, 3F288-2, NUDOCS 8802050118
Download: ML20196A666 (21)

Text

{{#Wiki_filter:______ l yu L',o ROj$. lO O J,D 09,, g,o m. y-e u:_O; c g

[

Florida Power C 0 R P O R AT t O N February 1, 1988 3F0288-02 Document Control Desk U.S. Nuclear Regulatory Wmion Washington, D.C. 20555

Subject:

Crystal River Unit 3 Docket No. 50-302 Operating License No. DPR-72 Additional Information - Diesel Generator Imding

Dear Sir:

Florida Power Corporation (FIC) met with the NRC staff on January 20, 1988 concerning diesel generator loading. During this meetirg, several concerns were discussed. As a result, additional infomation concerning diesel generator loading is attached. Should you have any further questions, please do not hesitate a contact this office. Sincerely, [ E. C. Sirpson Nuclear Operations Site Support DGG:UU l Attachmnt 1 xc: Dr. J. Nelson Grace Regional Administrator, Region II Mr. T. F. Stetka Senior Resident Inspector 880205011UBBO$b302 PDR ADOCK O PDR P I Post Office Box 219

• Crystal River, Florida 32629
• Telephone (904) 795 3802 0

A Florida Progress Company g

t QUESTICH 1: Explain why punp curves cannot be used. RESIONSE 1: h load (kW) test data showed reasonable correlation with the punp curves except for MJP-1B and BSP-1A. 'Ihe reason for the difference between the test results and the punp curve for MJP-1B is currently under evaluation. 'Ihe use of this punp in its present corrlition aligned to the "A" emergency diesel generator (EDG) will only be done if sufficient margins for EDG loading are justified using existing load (kW) test data or additional testing confirms lower power requirements. 'Ihe test of BSP-1A showed agresTP.:nt with the brake horsepower versus flow curve but the flow to discharge head data was not in agreement with the punp curve. W load (kW) test was paxformed on 12/21/87 (at approximately 1100 hours) to obtain data cr. load (kW) versus flow for BSP-1A. Additional data was also taken during the test which included pucp discharge pressure. 'Ihe test data taken indicates that the discharge pressure was low during the punp run. 'Ihis was not. discovered innediately since no acceptance criteria for punp discharge pressure was defined in the load test procedure. Subsequent punp and valve testing for operability, was performed for BSP-1A, on 12/21/87 (at approxirately 0200 hours) ard on 12/27/87. 'Ihe discharge pressure was within acceptable limits during both of these tests. h applicable data sheets for the load test and the punp arri valve tests are attached. Additional d b - ion with regard to E P-1A discharge pressure is provided in response to Question 12. QUESTICH 2: Why are Gilbert Ocmonwealth calculated values being revised at this time? RESPCNSE 2: h calculations are currently being revised to reflect the load test results, as well as modifications, which were performed to reduce emergency diesel generator (EDG) load. 'Ihese calculations will also formally amment changes in total load that were previously provided to the NRC by letter 3F1287-16 dated Decerter 14, 1987. 'Ihe revised calculations are currently scheduled to be subnitted with the suppleent to IER 87-19 by February 29, 1988. Additionally, docun9entation that supports the flow rate charges for SWP-1A and FWP-2A will also be provided with the calculations. l l QUESirICH 3: Identify all EDG loads. Include those that are auto connected, manually connected, ard those that are tripped and must be reconnected when the EDG load can be reduced to allow adding the desired loads. ) i I

ENCICSURE 4 8, (Page a or 12) SP-340A l DMR SHEEP IV EEiP-1A Initial Sucticri Pressure (BS-9-PI1) 37 caig RBE' CFERATING MtESSGtES Dischartle 199.5 - 219.6 Press 'BS-2-PIl 8 #9 l 1 Suctial 34 - 38 Press BS-9-PIl 3> W ' Differu1tial 165.5 - 181.6 Pressure (calc) (psid) 88 o g% i wg e Ng'" 1 3 o.wo A w t} W Iow paid High paid W d %*M 1

1. N dP Alert 2160.2-<165.5 181.6 -5183.3 La F 6

< 160.2 > 183.3 IN* Ptap Start time: 6/## Pump stop time : /l(70 r. PErf M Bf: MNm Dsta/dJ/-87 Time ad#f - By:ct tal - o.te a.um.e ccm f SP-340A RW. O7 Page 43 3

1 l i f.!Kf/.tiU.! ( r.o. 1 r, MnTOR POWER VER*ES PitMP FLOW DATA F0! Pf!Mf' BSP 1A DATA l FI 'W (CFM) l TCNP (F) PF ES Sl!T<T f r d fi-l POINT l [ CH 10-T! StlCTION . [ _ rein.'JIARM l 9._______'j...___'...___j...__......_. [PS-1 -FII [BS 9, FK1 1 l 80 4 Pil l DO ? I'il i _..q.... 1 I H60 l 1970 1 76 l 145 i .f_.__....q_...__ . j _... _ _ _ _9 _

3. 9..

I l I 73 1 3M iH4 l 2 11600 __f _j..___f; O O_._______j--p.....__.__ j...__.___..._j 1 7 I ~34 1 i ) i I 550 l i550 _ _p. _. 8.__j _ _ _ _3 4.. _ _.. l j f 4 1 /400 1 13ct O I 7g 1 L /93 I _9.. p-5 I I I = _.___. 9_ ___ l _ _ _. _. _ _ __ __ f_ _._.__.. __ _ j . _ _ ' l I I _ _ _ __. _.. ___. [ _. __..... 3 UA'!A L 5 =. I(.rt G r.a rs M i r. :, l l POINT POTFNTIAL (VOT.TS) CURRENT (AMPS) A-B l B-r l C-A A l B [ C t ..______t_ _ ____4 _.. _ j _q __ __5 I 1 4/50 9290 I _'l15 0 1 26 -f. 6 5 2 27 ._9 - ._ _.. f.... ..... j 2 I 9150 4250 yI 50 1 2 7. 5 1 27.5 ' 27.0.. _ _ _ _ _ ' _ __f -- __t t 3 I HISO 1 Hzs0 ) mSO l 27,5 2 7, 5 27.0 I t --i - ------ -i 1 1 2f' 5 1 26.51 2G.C l a l HI50 l4250 S __ __ _. 9. _4150 f __ _. j.. j .j __ f-I i i i I r __ _p _j l l - -... ] TNITJ ALS/DATE /J/?? I/2 - 2/~ 87 v l l PT m 3' 00 r+

t

• .~

ENCLOSURE 4 (Page 8 or 12) SP-340A DPGA SHEEP IV IEP-1A IN psig suction Pressure (BS-9-PI1) IUW GT3Ud3NG IRESSWES 79} $$Yd " 5 - 2 ' r,,,_2-Pn aoi 9mtion 34 - 38 g Press BS-9-Pn Differential 165.5 - 181.6 l(g Pressure (calc) _ (psid) i Iow psid High paid % Alert 2160.2-<165.5 181.6 -$183.3 < 160.2 > 183.3 s: amp start time: /4 /0 O amp stcp time : /73o j Perfrwmad Bf: /N. Data /2' 27'87h /I 30 af: 1 bk [ DataIA M-8) Tis,19 to naviewed SP-340A' Rev. Q7 % 43 E

RESIDEE 3: A table is attached which identifies the loads di m m % d above. This table is preliminary and subject to verification once the final revised calculations are received. QUESTICH 4: Establish the basis for the acx:uracies assumed in load values (Flow measurement ard electrical loads (W)]. RESIDEE 4: A diession of the treatment of accuracies associated with the flow measurement (ggn) and electrical load measurenent (W) for the tested load values is presented below. This treatment of accuracies will te utilizal in developing the February 1988 subnittal concerning the results of the testing including adjusting for instrument accuracy. W Measurim Instrimasiuxtion Error Trea+= wit The testing provided observed W readings at various flow pointa. Since it was not always possible to attain the exact flow point amnnad in the accident analysis, regression analysis was used to determine what the observed W reading would have been a*.the accident flow. First order regression analysis was performed where three or less than three test points were taken. Socord order regression analysis was done where more than three test points were taken. The next step was to correct this cbserved W value at accident flow for the accuracy of the W measuring instrument (trade name Dranetz). The nultipoint calibration was perfonned prior to and after each punp test, except for Make Up Punp 1A where only a post-test calibration was done. Each calibration point represents W input at a certain voltage, current and pcwer factor. Enough calibratica points were obtained to cover the W range of interest. Second order regression analysis was done to correct the observed W at power factors of unity, 0.86 and 0.5. Finally a first order regression was done to further correct this W value for the power factor observed at or near aa:ident flow. The more conservative corrected W value obtained frun regression analysis performed on pre-test or post-test calibration data was selected as the correct W value. Flow Measurim Instr =witation Error Treatwvit 7he treatment of accuracies associated with the ficw measurirq instruments during the testing is presented here for each of the punps tested since the rationale for treatment varied based on the system beirg tested. 01/27/88 Auto Connected loads on "A" Emergency Diesel Generator. EQUIPMENT FLOW KW BSP-1A 1600 185.7 SWP-1A 8500 485.5 MUP-1A 600 615.4 RWP-2A 15500 537.9 RWP-3A 10500 194.6 DHP-1A 3250 273.8 EFP-1 430 527.9 DECAY HEAT CL CYCLE COOLING WATER PUMP DCP-1A 3400 74 CONTROL COMPLEX LIGHTING 32.6 INVERTERS.............................................. 58.9 MISC AC DISTRIBUTION PANELS............................ 14.7 MISC PUMPS AND SMALL MOTOR LOADS 34.3 REACTOR BLDG FAN AHF-1A 61 DECAY HEAT CLOSED CYCLE COOLING FAN AHF-15A 2.8 FLUSH WATER PUMP DOP-2A 8.9 ES A 4160V/480V TRANSFORMER LOSSES 1.3 TOTAL LOAD ON EDG 3A FOR THE FIRST 10 MINUTES 3109 (The Operator has been provided guid8nce to trip EFP-1 at time T210 minutes, upon determining > 400 GPM LPI Flow.) Manually connected loads applicable to both "A" and "B" Diesel Generators EQUIPMENT KW SPENT FUEL COOLANT PUMP 41 CHILLED WATER SUPPLY PUMP 17 CONTROL COMPLEX WATER CHILLER 193 EFIC CONTROL COMPLEX FAN 13 EDG "A"__ loads that are tripped and must be reconnected by Operator action EQUIPMENT KW HEAT TRACING 41 BATTERY CHARGERS 93.1 The following load is a manually applied swing load which is normally aligned to the "B" side. The load on EDG "A" must be reduced to allow realignment of this load to the "A" side. EQUIPMENT KW ES MCC 3AB 91 NOTES TEST VALUES CORRECTED FOR KW INSTRUMENT ERROR. A FLOW ERROR CORRECTION OF 8 GPM WAS MADE TO EFP-1 FLOW TEST VALUES. ALL OTHER PUMP FLOW TEST VALUES DO NOT INCLUDE ANY FLOW ERROR CORRECTION.

• • CALCULATED VALUES. MAY CHANGE AFTER REVISED G/CI CALCS ARE RECEIVED.

?RELP11 NARY

l i WP-2A (Emergency Nuclear Services Sea Water Punp) - The required system flow for its safety ftretion is 13,400 gpn. The system is throttled to 14,600 gpu to account for worst case conditions of tide level, systan fouling, and instrument error. Flow was measured with a portable ANNUBAR with a range of 0-20,000 gpn and accuracy of 200 gpn. The accident flow will be as left and test instrument error does not affect this. WP-3A (Decay Heat Service Sea Water PLmp) - The requiral system flow for its safety function is 9,700 gpn. The systan was tested at 10,500 gpn with no additional throttling. Flow was measured with a portable ANNUBAR with a range of 0-20,000 gpn and accuracy of 200 gpn. The accident flow will be as left and test instrument error does not affect this. SWP-1A (Nuclear Services Closed Cycle Cooling Punp) - The total required systen flow under emergency corriiticns is 8,375 gpn. The systan individual cmponents flow rates were adjusted to their specific design value +3% to account for the inaccuracy of the flow instrument used. The punp load (kW) determinaticn for accident conditions used a flow indication frun an installed flow instrument with a range of 0-16,889 gpn with an accuracy of 200 gpn. The accident flow will be as left and test instrument error does not affect this. BSP-1A (Reactor Building Spray Punp) - The nonnal systan flow is 1500 gps. The systan has autcznatic flow control which is set at 1550 gpu. The flow instrumentation string has a range of 0-1800 gpn with a calibration tolerance of i36 gpn. The punp loads (kW) determination for the accident condition relied upon the same instrument strire as the autanatic control logic to monitor the system flow at 1600 gpn. Accident flow will be controlled to the setpoint with the same error as existed in test, which is within allo m d tolerances. dip-1A (Iow Pressure Injection PLmp) - The mininnn required system flow for core coolirg is 2700 gpn. The systan has autcanatic flow control with a setpoint which is adjustable frcn the Main Control Board (MCB). Procedurally, this control is set at 3000 gpn for the IPI standby mode. This flow control instrumentation striry has a range of 0-5000 gpn with a calibration tolerance of i100 gpn which is also the range and tolerance on the MCB indicator. The punp loads (kW) determination for accident conditions used the autanatic setpoint control on the MCB to set the designed flow and readirgs were taken fmn the MCB irxlicator. Accident flow will not change because the automatic flow control will maintain the rate with the same accuracy as used in the test, which is within the allowed tolerance. ~3- ~

i MJP-1A (High Pressure Injection Punp) - The total required flow for the system is 500 gpm delivered to the Reactor Coolant System (RCS) at a pressure of 600 psig. This capability is deas=Lrated by system tests which take into account pressure and flow instrument inaccuracy (+25 psi and actual calibration data of flow strings). The flow is also balanced between the four injection legs and each punp is throttled to maintain flow between 500 arri 545 gpn. The flow instrumentation ranges are O-500 gpn for each of four injection legs and 0-200 gpa for the normal makeup path with calibration tolerances of 10 gpn ard +4 gpo, respectively. The load (kW) test used these same instruments to monitor the flow, therefore, accident flow is not affected by instrument error. The effect of RCS pressure lower than 600 psig is conservatively accounted for by an extrapolation of the punp flow /vs. kW curve out to a runaut flow of 600 gpn. EFP-1 (Emergency Feedwater Punp) - Emergency Feedwater is not required for a large break IDCA because of primary to secon$ary uncoupling. However, the autcanatic ocntrol of EFIC will continue its designated logic of filling both steam generators (SG) to the natural cirullaticri level at a rate of 8 to 2 inches per minute until terminated by the operator. This is dependent on SG pressure which varies this rate frcan 1050 to 800 psig, respectively. The system flow of emergency feedwater will vary frun 330 gpn to 82.5 gpn to each SG arri a==ing shared flow with the turbine drive punp is 330 gpn to 82.5 gpn plus a recirullation flow of 100 gpn for a total of 430 gpa to 182.5 gpn. For conservatism the higher value was used for kW calculations. The load (kW) test flow measurements were corrected for an offset in calibration of +8 gpn at the 25% of span (0-1000 gpn) because the EFIC control will call for a real flow to satisfy its prograntned fill rate. 4 QUESTION 5: Identify design basis ficw rates required to assure function. Also identify flcw rates (actual) for the systems without operator action. In those cases where operator action is required to redu or otherwise control flow, discuss the basis for a==ing this can be accca:plished successfully. Document where the design basis has been changed. RESKNSE 5: The design basis flow rates required to assure functions including revisions to design basis flow requirements for SWP-1A and RNP-2A were provided in letter 3F0188-08 dated January 7, 1988, and letter 3F1187-19 dated November 16, 1987. In the case of the EDG load analysis, no credit was taken for operator action to reduce or control flow. QUESTICN 6: Identify %here in 7bchnical Specifications, etc., the operator will be made aware of the required flows. _- _

4 RESPGEE 6: This informatice will be pla d in the FSAR update cm the EDG analysis. It is currently s&eduled for inclusion in the July, 1988 FSAR sukanittal. Tectinical Specification Bases are expected to include this information after Technical Specification Inprovement is inplanented. QUESTIN 7: With regard to tripping the Class 1E battery charger, what is the battery voltage (terminal and at the load) at its lowest point before the charger (s) are recomected. Miat provisions are there to ensure that they are recomected. Miat is your basis for asstating that the inverter loads will remain functional when c.u==mted to the d.c. bus. RESR N5E 7: Battery chargers 3A, 3C and 3E each have an output capability of 125V, 200 anps. Battery charger 3E is a standby for either 3A or 3C. This is acocaplished via mechanical interlock that allow charger 3E to replace either 3A or 3C. The modification to battery chargers 3A, 3C and 3E provides a control s&eme which trips the battery chargers under a condition of a loss of Offsite Power coincident with ES actuation. This is uv-lished by cu==5cting a contactor in the power circuit of each charger between the charger and MCC [ famim breaker (see the attached drawing). This control schane will allow the operator to reconnect the chargers administrative 1y after approximately 30 minutes. The battery voltage at the end of this time period (30 minutes after tripping the chargers) has been analyzed to be 109 volts. A list of ncn-safety related loads on the "A" battery is also I i attached in. response to an NRC request. However, this list is i preliminary and subject to verification once the final revised EDG calculaticris are received. t i The inverters used at CR-3 have their primary source of power as ES 480V/AC fran Motor Control Centers which are fod fran the EDG. The battery / chargers serve as an alternate source of ( power for the inverters. The primary source of power is AC l converted to DC and is held at a slightly higher DC voltage level than the alternate source DC. Mien primary source is i lost, power will autanatically be supplied frun the alternate source DC and an alarm will be sounded in the Control Room alerting the operator that the inverters are cm alternate I ) source. When primary source AC power is reestablished, the inverters automatically revert back to primary source because of its slightly higher converted DC voltage level. There is no operator action involved in transfer between primary and alternate source. e l -Sa 1 - - - -.. - ~ ~ ._...-.__-_ _,~ ~- _..,-.- --- - J

MI o'3 i

[N f-{il a

i b

f. (!

] .... 3 M -ta < ids]i g

J E

A ~T nl. l j[

j !. ll]

O --+ t)/ o - y-0-E y !:' jl.['j!!,!a[{g4

l g

^ s ~ '1 nl. -os-s i; a1 [ ' yg Mr8,' l 3 [ ~ g p, gw+;i. -p i 0 o ,p ma !l i' ' ill

i l

2 6 l --y E .+ 9 A 4 -w-i i i y = ll l 1 R g .:..e ,W i{ ? 3 I!! ll l !. i i g$!. i ~ a lj l:j,, !op! i i _w 4 a

s i

i i-a .i 1:1 p,

• /Ji y

g .' i - sag gg, s m g l 1' 2 i a g, .-w il s -%-'- ggg! M m: 'l j g,55 le at r g N'O N i11 'i '~ l _.Q; ih-> ,i +w-ll il i 11 =g 2_ ___,___7_____', ['h-)4 yj p-A -w 1 y'l,l in-4 9 j l Eg ~ l A! is l a, --a____a .i 9 4+l I '~ I' ~ g ~.L.-_. -911lll l L j@H %d@' (i>- -l l t _1_ _t'[ V !_++ SL i ~ ~ ~ 1 -9,_g ;;

iq,

,i i.3 = i g.f.9

l l 3

g a 5 e l 5 al. -cEr'l i ~

• ! 1 h

8 Illi "I 3 ! s g c. NI ,l y ) g - -p-!_w ig i g ,' A .l 1 's tIhl ~ i 4' -_g D D' d. -a. i I. e 7 ~~' 4 1. _ - i ~.i.l!!!!M.,.O "I' T if -r+ i

Non-Safety Dalated Trwha m "A" Batterv IDAD CCHPOND7r VOLTS AMPS WATTS I Turb. Emer. Inbe Oil Bearing Punp TBP-3 Motor 250 213.0 53250.0 'Ibrb. Dner. Inbe Oil Bearing Pung TBP-3 Control 125 1.0 125.0 DIOP-4A: Control Board Section ICSAR 125 3.0 375.0 Turb. Auto Step Trip Sol. Valves 125 1.0 125.0 WDV-1022 WDG Decay Tank Vent Valve 125 0.5 62.5 DPDP-3A: CP3 Main Transformer 125 2.0 250.0 CWV-6, 8 Sec. Svce Ht. Exch. Isol. Vlv/ARV-3, 4 Water Box Rel. Vlv, ARV 52, 54 Cond. Vac. Bkr Vlv 125 3.0 375.0 CWV-1, 3 Sec. Svoe Ht. Exch. Isol. Vlv/ARV-14,17 Water Box Rel. Vlv, ARV 56, 57 Corxi. Vac. Bkr. V1v 125 3.0 375.0 Feedwater Punp 3a 'Iurb. Emer. Oil Punp Motor 250 18.8 4700.0 WICP-1 Cyc. Makeup Demin Regen Cont. Pnl 125 5.0 625.0 CR3 Unit Aux Transformer 125 2.0 250.0 CR3 Start up Transformer 125 2.0 250.0 Turb. 'Ihrust Brg. Wear Detect 125 0.6 75.0 Electro Hydraulic Fluid-Alm., hisc. 125 8.0 1000.0 EEV-2 (Control) Hotwell Iso. Vlv 125 1.0 125.0 EXV-3, 9, 21, 22 Extraction Stm Check Vlv 125 2.0 250.0 i Feedwater ptmp 3a 'Ibrb. Emer. Oil Punp Control 125 1.0 125.0 SCV-3 Sec Svc Cl Cyl Pp 3A Disch Iso Viv 125 0.5 62.5 Main FW PLmp 3A Cont, Pnl. 125 2.0 250.0 ARV-49 (Control) Cond. 3A Vac. Bkr. Relief Vlv 125 1.0 125.0 ARV-49 (Motor) Cond. 3A Vac Bkr. Relief Vlv 250 2.1 525.0 EFV-2 (M: tor) Hotwell Iso. Vlv 250 0.7 175.0 DPDP-2A: Oil Lift Pp for Reac Cbolant Pp 3C (Control) 125 1.0 125.0 Oil Lift Pp for Reac Coolant Pp 3A (Motor) 250 10.7 2675.0 Oil Lift Pp for Reac Coolant Pp 3C (Ibtor) 250 10.7 2675.0 Oil Lift Pp for Reac Coolant Pp 3A (Control) 125 1.0 125.0 DCV-10 Decay Ht C1 Cyc Tk 3A D e Wtr Cbnt Viv 125 0.5 62.5 Hydrogen Purge Filter 125 2.0 250.0 (WATIS) 69,387.5 (kW) 69.4 t t a

QUIsrIGE 8: Present the results of your voltage analysis of the 480 volt systen (Bus and at load terminals). This should also be considered in light of battery discharge and whatever inpact this miWit have on 480/120 volt loads. RESKNSE 8: Upon Ioss of Offsite Ibwer coincident with ES actuation, the autown.6cted loads are applied to the anergency Diesel Generator (EDG) in load blocks at 5 second intervals. At the instant each load block is applied the IDG voltage dips haa of the starting inrush KVA demand of applied load block. The resulting EDG voltage profile on 4000V base was calculated to be-as follows:- Mini== Volh AT Time T= Ioad Block 1 70.8% 0 Seconds Icad Block 2 83.3% 5 Seconds Ioad Block 3 83.2% 10 Seconds load Block 4 84.8% 15 Seconds Iomd Block 5 85.2% 20 Seconds The effect of these voltage dips was analyzed as follows: 1) It was recognized that the Block 1 voltage dip would cause a pick up delay at the Motor control Center contactor level. This delay was conservatively anstaned to be 2 seconds which is the time it takes for the EDG voltage to recover to 100% level fran the 70.8% level. The Second Inval Undervoltage Relaying time of 5 seconds takes this two secerx1 delay into consideration and allows CR-3 to meet 'nKhnical Specification response time requirements. 2) For Blocks 2 through 5, the Block 3 voltage dip is the-worst case. This case was analyzed and it was determined that the cou.w. ling voltage at the Motor control center contactor level is 73.4% of 120 VAC. This is above the contactor drop out level of 65%. Therefore it was concluded that the contactors will not drty out during block loading. Analisis of the testing done to verify voltage dips at the EDG during Is actuation testing has not been ocupleted yet. Them is no,inpact of the battery discharge on the 480/120 volt loads. The Inverters will rely on the battery for their primary source of power for only 50 seconds after a Ioss of offsite Ibwer. The Inverters will autenatically revert back to the primary source of power after EDG block loadirg is ocmplete. The battery discharge load profile for CR-3 assumes that the Inverters are supplied frun the battery for 100 minutes. The load profile also assumes no available AC power. Therefore, battery discharge during this event will be conservative relative to the a-nad load profile. The "A" EDG calculations include 58.9 kW load for the Inverters. QUESTION 9: Is AHF-1A a required load? REERNSE 9: Containnent Cooling Fan (AHF-JA) is a required load. There are three Containment Cooling Fans. Two are required to be operable by Technical Specification 3.6.2.3. The third fan is an installed spare. AHF-1A is currently out-of-service for repair. As a result, the spare third fan has been aligned to the "A" train. QUESTICN 10: Provide the basis for the control ocuplex fans not being an auto-connected load. (Nhere is the analysis to support this charge?) RESKNSE 10: The control ccmplex emergency supply and return fans have always been a manually applied load. As a result, this is not a charge to the plant configuration. A revised Control Rocn Habitability analysis (which continued to assume these fans where a manually applied load) was subnitted by letter 3F0687-16 dated June 30, 1987. This revised analysis was provided in response to NUREG 0737, Item III D.3.4 (Control Rocn Habitability Requirunents). Further, FPC is unaware of any generic guidance, applicable to the CR-3 licensing basis, which delineates the loads that should be auto-connected. QUESTICH 11: The 10P-1B (spare third punp) load is sianificantly laruer than MJP-1A. Will the spare ever be aligned to the EDG as an auto connected load? (What is the AkW of punp?) RESIDEE 11: The spare third makeup punp (!UP-1B) load is larger than assumed in the calculation ard larger than the "A" train punp. As a result of testing, the load for MJP-1B was detennined to be approximately 80 kW greater than MJP-1A. The }UP-1B may be aligned to the "A" EDG in the future. This would only be done if MJP-1A was out-of-service. Under these corditions, if sufficient margin did not exist on the "A" EDG, the Technical Specification Action statement for an inoperable EDG would be entered. In order to assure that the loads on the "A" EDG do not exceed the load at which it was tested (3248 W), an "A" EDG load configuration snagement program will be utilized. 'Ihis program will remin in effect for the remainder of Cycle VII. 'Ibe program consists of the following elements: 1) A matrix of loads and a=mntions used in the EDG calculations will be i=M to system ergineers, maintenance and operations personnel. 'Ihis matrix will be maintained to reflect the current configuration of the plant with respect to the load on the EDG. 2) Any new loads will be evaluated prior to addition to ensure the EDG has available capacity. Design modification procedures will be revised to prwide for this. 3) Operatirg aM surveillance procedures will be revised to provide assurance that positions of valves that were throttled during load testirg will not be charged without engineering concurrence that the effect on EDG load is acceptable. 4) 'Ibe procedure governing work requests will be modified to ensure that no work is performed that could affect the EDG load without engineering concurrence. Guidance will be provided to maintenance regarding the types cf activities that oculd have an impact on ED3 load. Additionally, post-maintenance test procedures will be developed arri utilized to quantify the impact of maintenance activities cn total "A" EDG load (i.e., replacirq an "A" train ptmp or purp rotor). 5) Plant vital bus rea ptacles on the EDG will be identified and controlled to ensure the assur:ptions in the calculation are maintained. 6) Training covering the EDG load configuration management program will be provided to all operations aM maintenance personnel. 'Ihese elments are considered to be sufficient to assure the loads on the "A" EDG do not exceed the tested load value of 3248 kW for the remaiMer of Cycle VII. QUESTICH 12: Is the Reactor Building Spray PLmp (BSP-1A) marginal basM on the Head / Flow curve (actual versus calculated)? RESFCNSE 12: When FPC went back to verify the mininum required head for BSP-1A, it was determined that the previous calculation referenced did not include the suction head available to the systm. When the suction pressure was subtracted from the discharge presenre (surveillance procedure (SP results)), the resulting A P was below the calculated required head. FN subsequently verified - I

~ that the calculations being used were the proper ones and also re-ran the SP at various flow rates to generate a new punp curve. 'Ihe SP results confirmed previous test results. A detailed evaluation of the test data and instrument location / orientation revealed that tne pressure gauges may not be giving a true indication of the total head developed across the punp for the following reasons: Both the suction and discharge pressure gauges are not reading directly from the process piping, but rather frun a seal water tap off of the discharge pipe. It is felt that the seal water flow through this 1/2" and 3/4" piping (in addition to a cyclone s@arator) is affecting the pressure readings. FPC is presently putting together a modification package which will install measuranant equipnent directly on the suction and discharge process piping in order to get a true '191 measurunent across BSP-1A. 'Ihis is urderway, with another test presently scheduled to be run early durirg the week of February 1,1988. If this test indicates that BSP-1A is actually perfoming below the minimum system requirements, the applicable Technical Specification Action statement will be entered. FIC is also pursuing the following actions: 1) QRgulaticms - A) A system curve will be plotted along with the punp curve (plotted fran actual test results) to determine how BSP-1A will perform in the present configuration (both flow ard pressure). 'Ihis will be ccx: pared against the systan design requirements. B) FIC has an analysis that confirms that a spray flow as low as 1200 gpm is acceptable frun the stardpoint of iodine remwal,111 ard equipment qualification. It is being confirmed that the lower iodine removal (at lower flows) is acceptable for Control Rocan Habitability. 2) Maintenance - A) FPC is researching past testing data to detemine if there was a large decrease in BSP-1A perfomance. If so, maintenance records will be evaluated to detemine if there were any major overhauls (e.g., inpeller replacements / modifications) that occurred durirg this pericd of tire. B) Required spare parts for a punp rebuild are being identified ard located, should purp maintenance be required. Additionally, if maintenance is performed which would affect the load kW, another load test will be performed for BSP-1A. In conclusion, if pucp performance has degraded, actions are t underway to insure that it renains capable of perfomirg its safety function. I QUESTIN 13: naar ribe assunptione used to support system line-up for each function (IDSP, IDCA, Failure of d.c. (one train of ESF's)] since the Emergency Feedwater flw of 430 gpa does not appear to be acceptable. RESIDEE 13: 'Ihe large break IDCA Flw of 430 gpn for the emergency feedwater i punp (EFP-1) is a result of shared flow with the turbine driven punp and EFIC control logic as follows: [ 1) EFIC Iogic whl require filling of both steam generators at i a rate of 8 inches per minuta until the natural circulation l setpoint is reached or the operator terminates EFW. 'Ihis is-a conservative value because EFIC varies the level rate froen 8 to 2 in/ min hamad upon steam generator prammire fran 1050 to 800 psig, and during a large break IDCA this pressure drops very rapidly. In addition, when the l turbine driven pump is unavailable due to depletion of the steam supply the lewil rato would be 2 in/ min which requires a flow rate of 365 gpa including recirculation. 2) Each steam generator requires fran 330 gpn for an 8 in/ min fill rate down to 82.5 gpn for a 2 in/ min fill rate. 3) 'Ihe ocznmon emergency feedwater punp recirculation line allows a ocabined flow of 200 gpn for punp protection (100 gpn each punp). 1 4) 'Ihe turbine-driven emergency feedwater punp is actuated by t either A or B channels of EFIC through opening of steam admission valves ASV-5 frun "B" train and ASV-204 from "A" train. 'Ihese valves are DC motor operated with concling DC power separation. 5) 'Ihe EFIC flow control valves fail open on loss of pcuer. 'Ihis assures flow will be fran the turbine-driven punp even in the event the failure is the "B" battery. Overfill protection in this event will be provided by vector valves which are frun the "A" train side of the systen. 6) 'Ihe turbine-driven punp requires no support systems for operation other than opening of either steam admission valves. QUESTI N 14: Identify training, prmartires, and instrumentation that has huma essential to ensuring anmaaaful "load maw; A." on EDG "A". Name each action. Also h=arit which loade would be considered candidates to be shut down in order to permit i restarting loads required (e.g., chargers, turbine bearing oil pung, control couplex fans) say 30 minutes after a DBA. RESFWSE 14: 'Ihis information was originally provided by letters 3F1187-26 dated Novenber 25,1987 and 3F1287-16 dated nar=har 14, 1987. i It has been revised to address additional concerns diamaaad during the January 20, 1988 meeting and is presented below.

i i l In order to aid the operators in managing the diesel generator loads, alarms have been provided in the control rom. These alarms indicate when the diesel generator is beiry cperated in the 30 minute rating. The alaIns (which deal with diesel generator loading) have been placed in close proximity of the diesel generator load indication (kW meter). This allows the operator to remain near the location of the annunciator alarm and be able to determine the magnitude of the load on the diesel generator. Ackiitionally, an elapsed time indicator (which records the cunulative time the diesel generator has operated above 3000 kW) is located below the alarms. i Following ccmpletion of in=ailate actions in mergency and abnormal procedures, a control roczn operator will be available and assigned the task of EDG load management until the EDG load is within the 2000 hour rating. In response to the 30 mirute alarm, this operator is instructed to refer to the abnormal prccedure concerning diesel generator actuation. This procedure indicates that if the diesel generator load is greater than 3000 kW, the load should be ruhvwi by stopping diesel generator supplied equipnent that is not required. A revision was made to provide guidan regarding possible loads to shed. This guidan acktresses loads of large ocmponents that have 100% redundant ocmponents (i.e., SWP-1A or SWP-1B) available. Additionally, synptom based procedures currently exist which indicate the conditions when high pressure injection (HPI) and building spray (BS) pums are not required. Guidance is also i provided to indicate the corx11tions when emergency feedwater (EFW) punps are not required. A description of the guidance provided for EDG load management (including instrumentation utilized) is attached. Training for this procedure revision on diesel generator actuation was provided to each of the licensed operators at Crystal River Unit 3. This was corducted as part of the abnormal pro dure revision process and included a discussion of the rationale for the charges. Durirg operator training it was

stressed, 1) cxxponents which are required will not be prematurely removed frcra service, and 2) where specific guidance indicates the conditions when cxxponents may be runoved from service (i.e., HPI, BS, and EfW), that guidance will be followed.

-n-

EDG Icad 9tmagmuert addanan Action JJ;Wlg) Instrumerit Itsgg ED3 load is >3000 W, Megawatt meter 0-4 Mr Reduce load <3000 kW MVAR meter (-4)-(+4) ) CAR Stcp EDG supplied EFP-1 EFP-2 flw 0-820 gpm equipment that is not SWP-1A or 1B disdarge pressure 0-200 psig required (Any 100% capacity equipnent may RWP-2A or 2B breaker status lights be stopped if its current meters 0-150% redundant equipnent is operating) AHF-1A, 1B, or 1C breaker status lights current meters 0-150% a n LPI flw >400 gpn, EFP-1 IPI flw O-5000 gpn EFP-1 may be secured LPI flw >1000 gpm per HPI punps IPI f1w 0-5000 gpn injection line for ikkeup Punps) 220 min, HPI may be secured RB pressure is <10 psig BSP-1A RB pressure (-10)-(+70) psig and not rising, RB spray may be secured i

For ocuponents where specific guidance is not provided for remwal frm service, the decision making pr-a of the Shift Supervisor will be relied upcm. However in this case, it was straawi during training that u.mpants for which 100% redurviant ocmponents are available should be considered as primary carxiidates for runoval frm service. W e training for this procedure revision was ocmpleted prior to reaching Mode 1 (Ibwer Operation) after Refuel VI, and provides assurance that each of the operators is able to manage the diesel generator loads. QUETIm 15: What will you do if you inadvertently use up (load >3000 kW) the reaining portion of the EDG 30 minute rating (24-25 minutes currently exist)? RESFW SE 15: Should the "A" EDG operate at a load greater than 3000 kW (for any reasco)'during the remainder of Cycle VII, the marnfacturer required inspection and surveillance 4.8.1.1.2.d.4 as provided in 'ISON 157, Revision 1 (dated 11/16/87) would be performed. m is will provide assurance that the "A" EDG will remain capable of supplying its required engineered safeguards loads. In the event the ED3 is required to operate in the 30 minute rating, this provides the operators with a minimm time of 23 minutes to reduce EDG load. QUETIm 16: Discuss your comitment to your long tem fix. Is it apparent to you at this time that you need more EDG capacity in operation following the refueling of the present core? RESIDEE 16: he omnitment for the long tem corrective action to the EDG load conoem is currently scheduled to be provided by the end of March, 1988. QUETION 17: Dianm the EDG power derating that would be required if the ombustion air teperature exceeds 105 F (95 F cutside anblent). What percent per degree F would you expect? What is your fix (tentative)? RESIOEE 17: FIC has ccrpleted a cxrbustion air tertperature transient analysis for the "A" Diesel Generator rectn. De results indicate that with the diesel operating at rated load, the l naxinum room terperature will be 106.6 F (on the hottest projected day of the year, 95 F). m is is 1.6 F above the maxinum room taperature recuu.arded by the ven$or, Fairbanks }brse. 'Ihis analysis assumed that both HVAC fans were running. FPC has otrpleted the design of an HVAC modification that will provide outside air directly to the inlet of the EDG turbocharger. 'Ihis will assure that emhmtion air will remain 1 i ~

below 305 F. 'Ihe decision to inplement this modification will be made as socn as the total EDG projected loading is finalized (and before outside air temperatures eynaad 93 F). If the modification is not implemented, m-3 would be required to de-rate the EDG fran the standpoint of maximum load capability. '1his would be a@roximately 11 kW for each degw of tenparature ' above 105 F. (e.g., for a room temperature of 106.6 F, the de-rating would be approximately 18 kW). If this INAC modification is inplanented, a portion of the total supply air from outside will be diverted directly to the turbocharger. As a result, the rocan air will increase above 106 F. 'Ihe follcwing ailrasmaa this concern: Both fans running provide 47,000 CFM of air. With 15,000 CEM diverted for ocatustion, only 32,000 CFM is available for room ventilatierVoooling. A GC/I analysis showed that with only one fan running (31,800 CEM without the INAC modification) the taperature of the rom could reach 120 F when the outside air tenperature was 95 F. 'Ihe equipnent in the roca was qualified to 120 F, taking into - consideration that the equipnent would only see this high taperature un$er the Equipnent Qualification design conditions of 1) outside air temperature = 95 F and, 2) EDG running at rated load (this would nmm only during surveillance testing (X # of times per year) plus an adiitional 7 days running, post-accident). As a result, with two fans runnirg, the rocan temperature is not expected to eyrmd 120 F. QUESTICE 18: Omfirm that modifications to annunciators have been reviewed for human factors (e.g., verified that information provided is clear aM that no new HEDs were intrrdM). RESIONSE 18: 'Ibe additional annun:iator Windows that were added were reviewal for human factors before they were installed. 'Ihey were prioritized, the legenis were properly worded, aM the appropriate location selected in accordance with the Human Factors Design Conventions n rwnt for the Main Control Ibard for Crystal River Unit 3., .}}