ML20195K116

Suppl to PWR Loop Ser:Use of Large Circulating Pump

ML20195K116
Person / Time
Site:	MIT Nuclear Research Reactor
Issue date:	10/24/1988
From:	Driscoll M, Kohse G MASSACHUSETTS INSTITUTE OF TECHNOLOGY, CAMBRIDGE
To:
Shared Package
ML20195K096	List: ML20195K090 ML20195K116
References
NUDOCS 8812050139
Download: ML20195K116 (11)

Text

.. _ .

'/I; ,'. . .  ;

. 1 j

. t I

r l

t l

r

?

L l SUFFt.DIENT 10 FWR IDOP SER:

USE OF IJutCE CIRCULATING FUMP i

by l 4 l t

i $

M. J. Driscoll ,

j C. E. Kohse '

i i t l  ;

4 i 1 ,I i Revised  !

! October 24, 1980  !

4 i

1 a ,

I s

I i

l i i r

i f i

i l

i 1

I l

i I

i l """

0911050139 DR esgg;.9 p ADOCK 05000o:o PDC

1. INTRODUCTION The purpose of this repwet is to summarize the operational and safety related aspects of the use of a larger, out-of-pile pump (an Autoclave Engineers' Magnepump) as an option for PWR loop operations, should the reference component, the miniature canned-rotor pump manufactured by NREC, fail to meet performance requirements.

Familiarity is assumed with the PWR loop SER (1). Ir. addition, a repo t has been published on project efforts to fix or replace the loop circulating pump (2). It is available as a reference for those inter-ested; however all pertinent information will be recapitulated here.

2. IARCE FUNF OPTION With strong input by the project's sponsors, a decision was made in late July 1988 to purchase the most suitable available, proven, commer-cial pump as an alternative option in cise our modLTind NREC pump f ails its performance tests (now in progress). The unit selected, a 3/4 HP Autoclave Engineers' Magnepump, is described in the appended literature.

It shares one important characteristic with all of the other candidates considered in ene search conducted by the pump task force - a size and weight so lage that it cannot be acconmodated inside the MITR-II core l tank. Hence, as shown in Fig. 1 (taken f rom Ref. 2), an out-of pile locatten is called foe. This is accomplished using the loop's m.akeup/

letdown lines (increased in diameter f rom 1/8" to 5/16" for the purpose) and counting the pump immediately adjacent to the shield block housing these penetrations. Otherwise all major loop features remain unaltered, both inside and outside the core tank.

In the section which follows, changes introdeced through incorpora-tion of this larger, ex-pile pump are reviewed, with a focus on safety-ralated issues.

(1) Safety Evaluuton Peport for the PVk Coolant Cheaistry Loop (PCCL),

MITNRL-020 February 13, 1987.

(2) Fi tal Report of MIT PWR Loop Pump Task Force, August 15, 1988.

.: 4 A. SMALL PUMP (NREC)

MITR Core Tank

(

m C y aam m nn Bridge Inconel Tubing '

~

o -

l B. INTERMEDIATE-SIZE PUMP

[ y Zircaloy

, Connecting Tubing ,

=

9%g -

Pump Pod (in place of second loop)

C. LARGE PUMP (AUTCCLAVE)

Zircoloy Tubing -

ws' insulated +

l

[ '

(_ ) ' ,

r-trace heated i

"o m insnnsnn

" _P ,

MITR Top Platform l -

l l Figure 1. Schematic of PCCL Pump Size / Placement l Options.

l l

L

. 3. MODIFIED th0P OIARACTERISTICS In its revised configuration full loop flow (2 gpm) is circulated outside the MITR-II core tank and its associated shielding. Hence one must shield the pump volute and associated inlet-outlet lines because of the N-16 activity in the water contained therein. A conservative esti-mate indicates an unshielded dose rate of approximately 40 mR/hr at 1 meter from the =200 cc cf contained coolant. This will be shielded using a lead-lined box over the pump, which also serves as a step-over for personnel, and a barrier to inadvertent contact with hot (albeit insulated) surf aces or interference with pump / motor operation.

Since the Autoclave Engineers' (AE) pump is an extremely rugged device (designed for 5000 psi, 650'F), potential fracture or leakage modes are expected to be significantly less than for the NREC unit.

Hence the AE pump is not hermetically sealed in a pod, as is the NREC pump. A leak tape will be used in a drip tray which will collect any water leaking from the system outside the core tank. Note that the insulation required on the t'tbing and pump head will prevent spraying from small leaks and the shielding provides an additional barrier. A ventilation hose will be installed to draw air (plus water vapor, if any) from the vicinity of the pump.

I.e akage into the core tank is also possible with the loop inlet /

outlet tubing running from the thimble top to the core tank wall feed-through. One or two unions will be present on each of the lines. Any leakage greater than the charging rate of 250 cm /h will produce depres-3 surization of the system and hence will be detected by the pressure instrumentation / alarms. The maximum possible leakage in this scenario is 3

the total loop volume of '800 cm3 plus 250 cm /h for the time af ter com-plete depressurization during which c' irging flow is maintained. Allow-ing 9everal hours reaction time gives a conservative total leakage volume of 2 1. Smaller leaks, between the cl.arging rate and =30 n /h,3 would be detected by periodic realing of the discharge flowmeter. For readings every 12 h curing reactor operation, the undetected leakage volume at 250 cm3 /h leak rate would be 3 t. Below 30 cm3 /h, leakage would be detected by imbalance between charged and discharged volume. A discrepancy of 2 L

s

• 4 corresponds to =2.5 in. of charge / discharge tank level and will be readily detectable. Note that the leakage will be added to the primary coolant at the top of the core tank and will be swept out of the tank and diluted before returning to the core inlet. The B 10 content of the i maximum projected leakage volume of 3 i is =48 mg. Diluted into the primary coolant "active" volume of =4000 i and assuming no removal by the reactor ion column gives a B 10 concentration of 12 ppb. This is not projected to have any measurable effect on reactor operation. (Notes  ;

The reactor's ton column has a 20 gpm flowrate which is 1% of the total primary flow. It would eventually remove the boron but the process would be slow.) i l

Unlike the NREC pump which is passively cooled by conduction to the .

cold in-pile thimble wall (bathed in MITR-II coolant), the AE pump requires active cooling to prevent er.ceeding the curie point of its coup-I ling magnets - which if exceeded vourd lead to a gradual loss of magnetic  ;

I field strength. This would probably not cause pump stalling in continu-ous operation, but would prevent restar'.tr.g af ter a shutdown. The pump b will be cooled using MITR-II shield cooling system water and/or stand- [

alone chilled water units. Its rotor can temperature will be monitored  ;

and alarmed using a strategically located thermocouple. Manual switching f

. between redundant cooling systems will be provided.

i The addition of the AE pump and associated tubing runs increases the  !

{

I loop water inventory by about one-third. Hence a review of the loss-of-coolant sequence is appropriate, together with other aspects which also af fect operations "aing the NREC pump.

I

4. 1,0SS OF 000iJutT INCIDENT (LOCI)

! l i

l Sudden release of the entire loop inventory of hot (600'F) pressur-a ized (2250 psia) water will raise the pressure inside its helium-filled (15 psig) aluminum containment thimble if the pressure relief valve and  ;

i burst disk fail to operate. Since the chimb1w contains a massive bed of copper shot (=120 lbs) as a heat transfer medium, the steam water mixture i L

j will virtually instantaneously be brought to quasi-equilibrium with the l shot at very nearly its pre-LOCI average temperature (=350'F), and hence !i a corresponding saturated vapor pressure of 135 psia will be established, (

essentially independent of initial loop water mass (in the 500-800 cc l i i i

- - , , - - - - . - - . .wm.____. _ _ , m y _ y_ _ .,y , . , _ . . _ . .,__...,__.m.-,_...- _ _ _ _ - . , _ _.-.___,m_. - . . _

,e .

range of current interest). Adding the pre LOCI helium pressure gives a

, post-LOCI maximum credible pressure of approximately 150 psig, ignoring steam condensation on the cold (=120'F) aluminum thimble wall.

Thus use of the larger pump does not substantially alter the in-thimble LOCI event. Since publication of the original SER, however, some 4

changes have taken place which substantially alter the analysis of this incidents The mean Cu shot bed temperature, cited above, has been recomputed (using the HEATING-3 program), increasing the confidence level in this value. The pressure increase scenario was also reanalyzed, resulting in the more realistic maximum pressure estimate of 150 psig as described above.

• The thimble burst disc pressure has been reduced from 500 psig to j 65 psig - with a corresponding reduction in its associated PRV l

(now at 20 psig): both changes permitted by installing them out-side the MITR-II core tank (using a larger diameter for the chim-

) ble evacuation / fill line, now also used for p. essure relief serv-ice).

As a result of these changes, the thimble pressure test procedure has been changed from that described in the original SER (See p. 10 and Appendix 1.c). Each thimble used for PCCL experiments in-core will be pressure tested to 150 psig - more than twice the rating of the burst disk (certified by the manufacturer to burst at 25% of the rated pres-sure). The first thimble has been so tested; no leakage or permanent deformation was observed. In addition, a prototype of the in-core sec-tion has been pressure tested to 1000 psig and the change in dimension of the minor axis of the approximately elliptical section measured. The significant results of this telt are as follows:

Permanent deformation oC the section begins at 195 psig.

. Sufficient deformatior, of the section to bring it into contact with the dummy element (total clearance = 0.150 in.) occurs at 325 psig.

. o

, e No leaking or other failure was observed up to 1000 psig.

Note that contact with the dummy element does not occur until the pres-sure exceeds twice the estimated maximum pressure for a LOCI with pres-sure relief failure. The flow channel would not be blocked during pres-surized deformation of the thimble since increase of the minor axis is accompanied by decrease of the major axis. Elastic springback on depres-surization of =.012 in. per 100 psi is observed, hence total springback l f rom the contact point is = (.012)(3.25) = .039 in., and the thimble is I expected to be removable from the dummy element after a pressurized deformation.

It should also be noted that the bolted flange of the 7. 5 in. ID thimble pod lid is also expected to relieve pressure through its 0-ring seal at a value less than 300 psig - at which a force totalling 13,250 lbs will act to lift the lid. This feature in fact precludes testing the entire thimble to extreme pressure values.

5. NREC PUMP This pump is still the technically preferred option, and will be employed if a high temperature motor can be perfected, and/or bet te r thermal insulation of the stator can be devised. Modifications to achieve these goals are described in Ref. 2. Of principal interest hera is the fact that no changes in pump internals or hydraulic design have been made (other than minor refinements in bearing configuration). How-ever (see Fig. 2), NREC has now changed to a bolted flange seal in place of the eariter snap-ring design. We consider this to be a significant improvcment, with a much larger margin against seal leakage. Note that the same pressure relief system will be used with the NREC pump as with the AE pump.

6. INCREASE IN CONTAINMENT BUILDINC HYDROGEN INVENTORY Since the publication of the original SER, the procurement and control procedures for hydrogen cover gas for the FCCL charging tanks have been revised. A commercial source for small cylinders (Size 4. 0.08 ft3) prelilled with 15 standard cubic feet (SCF) of H2 at the standard fill pressure of =2000 psig has been identified. Use of these cylinders

7 HOLD-DOWN FLANGE METAL C-RING U "e IMPELLER 7 ,

~

- - OUTLET PLENUM a w,,,,,.

s 1% ~

UPPER BEARING c

i i g

@\ ""-

- HEAT SHIELD t y i l -

I f aj, [

l =

STATOR Coll x

=

3 _

IRON ARMATURE t : -

0 I, 1 j s = ;; lNSULATION GAP

@ i h:___ t LOWER BEARING ROTOR STATOR CAN CROSS-SECTIONAL MEW NOT TO SCALE NOTE: 5% of primary flow is bypassed through the bearings on the path marked by 4 .

Figure 2. Schematic of NREC Miniature Canned Rotor Pump.

. O

, is safer, more reliable and easier to control than the originally planned procedure of filling a "transfer cylinder" on-site.

Taking into account the nacinal contents of these cylinders, and making allowance for cylinder to cylinder variation, inventory dissolved in charging and discharge tank cont 6ats, and cylinder change out (i.e.,

bringing a replacement cylinder into the containment before the in-core cylinder is completely empty), the maximum hydrogen inventory in the con-tainment builing is now set at 20 SCF. Note that this does r.ot alter the analysis of any postulated accident scenarios involving hydrogen inside the reactor core tank. The maximum possible energy release due to a hydecgen burn / explosion within the containment building is doubled, but the total energy is still small ( 6800 Beu), and the proposed ventilation and procedural safeguards remain adequate.

7. SUMWLY AND CONCLUSIONS An alternative loop layout has been described, to permit the optional use of a larger pump outside of the MITR-II core tank. No new, previous-ly unidentified ef fects on MITR-II safety have been identified. The most significant difference will be the need for N-16 shielding around the pump head and inlet / outlet lines.

A loop LOCI has been re-analyzed. No significant dif ference bet'reen the two alternativa pumps was found in this analysis. New thimble pres-sure test procedures are described which ' reflect the substantially lower pressure relief valve and burst disk ratings, and the revised estimate of maximum post-LOCI pressure.

M '

, APPENDIX Bulletin 15C Autoclave MagnePump e Both MagnePump styles feature the MagneDrive ll packless drive system. p A

• Rare earth magnets provide high torque drive. -

f,b.,,,,,,,,

s Pump system is completely sealed; there is no packing to wear away and cause leakage, contamination or g.

costly downtime, a No lubrication required, eliminating a common source 1 of contamination. N a Integral e!ectric drive. No belts or pulleys. ,,,,

General  % HP MagnePumps The Autoclave Engineers,MagnePumps ehmi. The % horsepower MagnePumpis a packless nate or reduce many of the problems as. design that provides improved efficiency: high sociated with conventional pumps, such og torque with reduced heat losses. it is designed leakage. contamination and packing heatgen for liquid service and is rated up to 5,000 psi eration. They are ideal for applications where system pressure.

punty of the fluid is a major consideration or Centrifugal whersteakageof materialcouldbehazardous. 5 HP MagnePumps

. powe,ioss is eiiminated eue io no seai tric.

Tne 3 no,,epowe, MagnePump is s packle,,

n ti n, theref te delivenng full m t r h rse- design which requires no supplemental coohng i P8CMOSS VnUmV power to the pumping unit. source. Built in centnfugal air circulator fOr CONTINUOUS

• When adverse conditions exist, the magnet achieves iower magnet temperature.it uses a flow Systems drive functions as a clutch, ei.m:nating over- standard NEMA motor and high energy rare earth magnets for high torque dnve. its outlet load and mo'c bumout.

and inlet flanges are in the same plane for easy p' ping. The 5 HP MagnePump is designed for hquid service and is rated to an allowable working pressure of 2,500 psi.

Speelfications

% HP 5 HP y*,,[' ,*,[' 5000ps:(344 5 bars) 2500 ps) (172 bars)

Maumum temperature 650 F(343-C) 650-F (3&C)

Operating 3d 50 'P*

• 60 "2 50 'pm a 60 Hz Conditions Q*r'W 'P'ed 2s75 rpm @ 50 Hz 3's75 2 rpm e 50 Hz MagneDnve stat < torque 15 n LDs 150 m. es.

Magfielcookag ,d,$ 83 No coolog reqwred Pump casing housang 316 35 M86 316 55 MS6 6rnoeaer 316 SS 316 SS Pump 1 Snai 316 ss 31s Ss Construction a Fasteners 300 ser.es sta oess 300 senes sta.rtess seuings fc%[y$ j3 Carbon gWe A /'

Sea W',

316 32155 su er o'ateo sr.1000 CA Nc e4 ano s.tver M Deeg r 0 3ea

<i o tue ng.i a 66s-)

1 W 2500 e. fiange conn.

1.5 i.o.

Mt SF 1000 CX conn P 2500 c. flange conn (10*teng. t O ESS*) 95 i D.

60 HZ: Exp-proot Type class 1 group B 50 60 H2 TEFC

& So HZ: TEFC 3 Pc a er 4. H P 5 HP j

11523o v ac, 't) H2 sangte pase 230460 i ac Vonage or 50 60 H.*

110MO vac.50 HZ 3 phaa s<ng's phase

, Autoclave MagnePump -l ~

General information

% HP 5 HP Moynbog pos4toi Hontontal Hontental Weight (appoxeats) 170ts. 380 es.

Masevm fbw 19 GPM 40 GPM Maneum head 40 feet (appres.17ps ) 194 feet (appros. 84 psi)

Note: Consum factory for byer head and hcw rates.

% HP 5 HP 70 200 --

00

) ,

100 5 N to Cree

{60 0 100 g(3454 rpm)-

d~

40 [ rpm) - t40 30 l l l '% ~

120 Tx \

20 l

a.f&

so 100 l \,

so I

I is 00 a.Tr' 0 l l ls l s to is 20 10 30 30 40 60 FLOW (GPM) FLOW (GPM)

Dimensions (in inches)

% HP 5 HP * * "

e.. _ ~. ~e ,

E i, -b l s I

'c  :- o - F A fp #r ,

7 [ )'>lW

• v L.,s M i

3 ra

, nr - ,i r --.

% ,.;; =r3 g i L g ffj .

-im --

u e,,== w ,ew.,ce .* m.v . me C^ " * ***'v a'"'" Ordering information e arve e prowoe seestmory se w e 84. Oce W a w shoe e4 mey 6wy evey we sammg N , - _. Cetalog ordet no.

sewe ecm ture, pres *See of Waa. temp,' VoMete ,%,,g 8

"v en orce'*e#'

a ** * *

• cre Maoutwr %HP MP2040 05 65 60 60 HZ neS9tCn000 SNOULD H niAM 343*C gracMe O W av pe=Nore m e,o o arye repecoc es rosared  % HP MP2040 05 SS 60 60 HZ Rubn-A 288* C

% HP MP2040 05 65 50 SO H2

v. HP pygp MP m o 0, ,, ,0 _

.0 HZ R-

. H, MPc., 02 .. 0  :::; em p,;,

29N Hf 2kw_bl SGRG1 Ve M %St2 U$a itM)S.%2C?t AMAVO$$7GV as os ww twee aM esaemeae of see e e2mvsa Wwsweas leNtemene of ew W9 ese'1 t n en we eM emmeen ee4