ML19347F530

Testimony of Cw Dillman Re Doherty Contention 6, Recirculation Pump Overspeed.Prof Qualifications Encl. Related Correspondence

ML19347F530
Person / Time
Site:	Allens Creek File:Houston Lighting and Power Company icon.png
Issue date:	05/11/1981
From:	Dillmann C GENERAL ELECTRIC CO.
To:
Shared Package
ML19347F516	List: ML19347F515 ML19347F520 ML19347F522 ML19347F524 ML19347F527 ML19347F529 ML19347F530 ML19347F531 ML19347F532 ML19347F534 ML19347F535 ML19347F536 ML19347F537 ML19347F539
References
NUDOCS 8105190540
Download: ML19347F530 (10)

Text

i i

[} [tELATED. @NN 5-11-81 p

~

q, 1 UNITED STATES OF AMERICA // Nd#d S NUCLEAR REGULATORY COMMISSION -

2l MAY 13 t00t 8' 24 BEFORE THE ATOMIC SAFETY AND LICENSING BOARD -

3 Omes of the sec. s 4 In the Matter of )

)  %

N 5 HOUSTON LIGHTING & POWER COMPANY) Docket No. 50-466

)

(Allens Creek Nuclear Generating) 6l: Station, Unit No. 1) )

7, )

I 8l t

9, TESTIMONY OF C. W. DILLMANN ON BEHALF OF HOUSTON f LIGHTING & POWER CO. ON DOHERTY CONTENTION 6 -

10 RECIRCULATION PUMP OVERSPEED 11 Q. Please state your name and place of employment.

12 A. My name is C. W. Dillmann and I am employed as Manager, 13 Heat Exchanger & Flow Control Valve Pump Design, by General 14 Electric Company. My business address is 175 Curtner Avenue, 15 San Jose, C&lifornia.

16 Q. Describe your professional qualifications.

17 A. My professional qualifications are set forth in Exhibit 18 CWD-1 to this testimony.

19 Q. What is the purpose of this testimony?

20 A. The purpose of my testimony is to address Mr. Doherty's 21 Contention 6, which states:

22 Applicant has committed itself to provide a decoupler to prevent destructive overspeed of the recirculation 23 pump motor. However, a potential for pump impeller overspeed exists. The Applicant states that impeller 24 missiles will not penetrate the pump case and that ejection or impeller missiles through the open end of 25 the broken pipe will be prevented by additional pipe supports and restraint..7 Petitioner requests that an 26 adequate basis be provided to assure that these measures will be effective.

27 28 .

8105100Sif60

1 2 Q. What is yo'ur understanding of the concern expressed 3 in this contention?

4 A. My understanding of Mr. Doherty's concern is that 5 under certain conditions which Mr. Doherty does not define, 6 overspeed of the recirculation pump could occur, resulting 7 in the destruction of the pump impeller and the creation g of potentially destructive missiles. Mr. Doherty is con-g cerned that these missiles could damage or destroy plant 10 structures and components necessary to safely shut down the 11 reactor, or necessary to mitigate the consequences of a 12 design basis accident.

13 0 As a preliminary matter, would you briefly describe 14 the reactor recirculation system.

15 A. The reactor recirculation system is a system designed 16 to provide forced coolant flow through the reactor core.

17 The system consists of two parallel piping loops, located 18 outside the reactor pressure vessel and inside the drywell.

19 Each loop consists of a recirculation pump and motor, re-20 motely-operc':ed normally open suction and discharge block 21 valves, a flc w control valve, connecting piping to the 22 vessel, internal vessel jet pumps with their connecting 23 piping, and associated system process and control instrumen-24 tation. Each external loop takes suction from the vessel 25 downcomer annulus via the recirculation pumps. The vessel 26 downcomer annulus is the region of the reactor pressure 27 vessel in which water having already passed through the core 28

I l

i 1" 2 is mixed with feedwater. It is located between the core 3 shroud and the vessel wall. Recirculation pump discharge 4 flow is split downstream from the pump to provide flow to 5 multiple nozzles on the vessel shell from which internal 6 piping distributes the flow to the jet pump inlets located 7 in the downcomer annulus. This driving flow will cause the s

8 jet pumps to entrain additional water from the vessel annulus.

9 The combined flow will then pass through the reactor core.

10 Water which is not turned to steam returns to the downcomer 11 annulus, to be mixed with feedwater, completing the recir-12 culation path in the vessel.

13 Q. Under what conditions can potentially damaging recir-14 culation pump overspeed occur?

15 A. Damaging recirculation pump overspeed can occur only as 16 a result of a break in the recirculation loop piping.

17 Q. Will you explain how such a break could result in 18 recirculation pump overspeed?

19 A. Should a pipe break occur in a recirculation loop, 20 reactor coolant will escape from the vessel to the drywell.

21 Under normal operating conditions, reactor pressure is 22 about 1020 psia, and drywell pressure is atmospheric (about 23 15 psia). Reactor coolant will flow from the vessel to the 24 break by each of the two available paths. The recirculation 25 pump will be in one of these paths depending on break 26 location. If the break is on the pump discharge line, 27 reactor coolant will rush through the pump much faster 28 than it will normally be pumped, due to the large

1 2 pressure dif ferent.ial between the reactor and the drywell.

3 ( 1000 psi). Th recirculation pump normally operates 4 against a differential head of about 780 ft. This greater 5 flow will exert a force on the impeller as it passes, 6 causing the ir.peller to accelerate. If the break is on the 7 pump suction line, the large pressure-differential between 8 reactor and drywell will reverse the direction of flow 9 through the pump, causing it to spin in the direction opposite 10 to its normal rotation. In both cases, the pump is now acting 11 as a hydraulic turbine and is subject to overspeed. Flow 12 in the reverse direction will exert a greater force on the 13 impeller than would the equivalent flow in the normal direction.

14 This is due to the hydraulic characteristics of the impeller 15 design. In an overspeed candition the impeller will 16 possibly be accelerated to the point where stresses created 17 within the impeller exceed the tensile strength of the 18 impeller material. The impeller would then fail, with pieces 19 of the failed impeller flying off the spinning pump and be-20 coming the missiles of concern in this contention.

21 Q. Mr. Doherty's contention states that " applicant has 22 committed itself to provide a decoupler to prevent destructive 23 overspeed of the recirculation pump motor." Would you 24 discuss this commitment?

25 A. The original design of the recirculation pump included 26 a decoupler as described by Mr. Doherty. The design has 27 since been modified, eliminating this feature. In the 28 current design, the motor will be coupled to the pump impeller.

r j

1 2 Thus, it will be subject to overspeed as well. Therefore, 3 Applicant will address the consequences of motor missile 4 generation as well as pump impeller missile generation.

5 Q. Why was the design changed to remove the decoupler?

6 A. In the period of time since the use of a decoupler 7 was identified, two steps were taken. One was continued 8 development and testing of the decoupler. The second was 9 added analysis of the probability of missile generation.

10 As a result of the decoupler testing, it was concluded 11 that while the decoupler was acceptable and would serve its 12 intended purpose, it would add complexity and servicing 13 requirements that were undesirable. The parallel study 14 of missile generation, using proven methods, concluded 15 that motor missiles would not be a problem. Therefore, 16 the use of a decoupler was deleted.

17 Q. What structures or components would a pump impeller 18 missile cr a motor missile strike?

19 A. In the case of the motor, missiles generated by 20 failure of the rotor will impact the stator and stator 21 frame. Missiles generated by failure of the rotor end 22 structure (consisting of the retaining ring, the end ring 23 and the fan) will strike the overhanging ends of the stator 24 coils, the stator coil bracing, support structures, and a 25 wall of one-half inch thick steel plate.

26 In the case of the pump impeller, missiles will either 27 impact the pump casing which is 2.9 inches thick, or be 22 injected into the recirculation loop piping. The missiles

1 2 which are injected into the pipe have the potential for 3 lea 1;ing the broken end of the pipe and striking various 4 drywell structures and equipment.

5 Q. What is your conclusion concerning the damage that 6 pump and motor missiles could cause and what is the basis for 7 this conclusion?

8 A. Motor missiles will in no case penetrate the motor 9 housing. Pump impeller missiles will not penetrate the 10 pump casing, and those impeller missiles which escape from 11 the broken end of the pipe will not damage the drywell 12 walls, any piping system, or the main steam isolation 13 ' valves. The basis for these conclusions is a GE report 14 transmitted to the NRC by letter of March 30, 1979, entitled 15 " Analysis of Recirculation Pump Under Accident Conditions."

16 Q. Will you briefly out ine the method used by GE in 17 this analysis.

18 A. Since it is impossible to exactly predict the size 19 and shape of the pieces of a failed pump impeller or motor 20 rotor, the configuration of the piece which would possess 21 the maximum translational kinetic energy was determined.

22 This is the worst possible missile. The conficuration was 23 found to be any 90 sector. The kinetic energy of the 24 worst possible impeller missile was calculated, based upon 25 impeller speed at failure. The Stanford missile penetra-26 tion formula (see Reference 6 of GE report) was used to 27 conservatively estimate the energy required to penetrate 28

1 2 the impeller casing. A comparison of these values showed 3 that the worst possible impeller missile would be contained 4 by the recirculation pump casing. A similar calculation was 5 performed for the motor rotor. Similarly, the results 6 showed that the worst possible motor missile would not 7 penetrate the motor housing.

8 Further, the possibility that a pump impeller missile 9 could be ejected through the recirculation loop piping and 10 out of the pipe break opening into the drywell was investigated.

11 The study, which considered possible pipe break locations, 12 planned pipe support location, and postulated missile 13 trajectories, concluded that with the application of break 14 criteria in accordance with Regulatory Guide 1.46, no 15 damage would occur to the primary containment, any major 16 piping system, or to an inboard main steam isolation valve.

17 Absence of damage is due to the fact that trajectories of 18 postulated missiles do not intersect these systems.

19 The analysis discussed above assumed that the pump impeller 20 would be accelerated to the fracture point by the postulated 21 recirculation loop pipe break. Additional analysis was 22 performed to determine whether the pump could actually 23 be accelerated to the overspeed failure point by such 24 a break. Pump speed is related directly to volumetric 25 flow through the pump. Volumetric flow rates through the 26 pump for different breaks were determined, based upon the 27 very conservative assumption that the reactor presstre does 28 not drop during the time required to accelerate the

ir

=

1 2 pump to its ultimate windmilling speed. Results of this 3 conservative analysis indicate that overspeed conditions 4 could exist for certain types of recirculation line breaks, b such as a double-ended guillotine break on the pump suction 6 line. The analysis demonstrated that for the complete 7 spectrum of discharge line breaks, no overspeed conditions 0 will exist.

O Q. Please discuss the conservatism of the GE analysis.

10' A. The calculation of whether an impeller or motor missile, 11 once created, would penetrate the pump casing or motor IE housing, respectively, is conservative in the following 13 ways. The Stanford missile penetration formula conserva-14 tively predicts the total energy absorption of a complete Ib ' penetration, since it is designed to compute merely the 16 energy required for partial penetration and significant 17 additional energy would be required for complete penetration.

18 In addition the calculation of ultimate windmilling 19 pump speed is conservative because the reactor is assumed 20 to remain at full pressure until ultimate pump windmilling 21 speed is reached. Also, flow through the pump for each 22 postulated break was calculated based on both a homogeneous 23 flow model (all water) and a two-phase flow model (water 24 and steam). Whichever result yielded the highest rate 25 was the one used. Another conservative assumption was 26 that the curves used to determine maximum pump speed 27 based on the previously calculated volumetric flow rate 28 are based on homogeneous flow thereby giving the highest

e 1 2 momentum. The expected two-phase flow would produce lower 3 pump speeds than the assumed single phase flow.

4 Q. What are your conclusions?

l A. A recirculation loop suction pipe break accident may 5

I result in overspeed to failure of the recirculation pump impeller and recirculation pump motor. This conclusion is based on extremely conservative assumptions. However, 8l 9 even in this unlikely event, the overspeed condition and 10 its consequences will not impair the operation of any

/

11 systems necessary for the safe shutdown of the plant.

12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28

'1 Exhibit CWD-1 2 EDUCATION AND PROFESSIONAL QUALIFICATION 3 Charles W. Dillman 4 Mr. Dillmann is Manager of the Flow Control Valve, 5 Heat Exchanger and Pump Design Subsection of General Electric's 6 Nuclear Energy Business Group. In this capacit1, he is in 7 charge of mechanical design, analysis and structural evaluation 8 of Nuclear Steam Supply System mechanical equipment.

9 Mr. Dillmann has 21 years of engineering experience, 10 15 of which are in the nuclear business. He has been a 11 manager of design engineering and manufacturing groups for 12 the last 10 years. This work has all been on light water and 13 breeder reactors.

14 Mr. Dillmann received his BSME from Purdue University 15 in 1959. He has taken 40 credit hours of graduate engineering 16 courses. He holds a professional engineer's license in the 17 State of California.

18 19 20 21 22 23 24 25 26 27 28