ML20238D186

Analysis of Feed & Condensate Sys:Peach Bottom,Fluidic Stress-Monitoring

ML20238D186
Person / Time
Site:	Peach Bottom
Issue date:	05/20/1987
From:	Behrens R PUMPCOM, INC.
To:
Shared Package
ML20238D184	List: ML20238D182 ML20238D186
References
NUDOCS 8709110028
Download: ML20238D186 (4)

Text

_

i I

l l

AN AI,Y T;I 5 nF FEFD AND rnNDFNSATE 7YSTEM: PEACHDOTTON May 20, 1987 Pumpcom 1: fluidic stress-monitoring (

l

f. e e d - wa t e r The purpose o f. thid study was to model the system of Peacht>ottom 2, ut ili si ng a Pumpcoin 1 modeling cyatem, and analyze normal and abnor ma l f low phe nomena , and any other generic issues which may become apparent, that could compromise plant uafety. Pumpcom 1 was utilised i because, it is the only system currently available that can i study " secondary flow events" in real-time. Because l Peachbottom 2 is a BWR, not a PWR, the pressures and I temperatures are different, hence, the matter-energy matrix ( I the primary element of a Pumpcom model) was considerably (

different. However; in those areas of the feed-water system )

where the matter-energy matrix approximated that of a PWR, similar flow-patterns were identified. Fortunately, all of these areas are defined by pumps, where they can be phys wally isolated from the system using valves, or monitored with simple instrumentation.

Method:

Pumpcom models are constructed in such a way as to i measure fluid-stress. By this we mean the concentration of {

energy per un!t mass & volume. The proper execution of a l Pumpcom model, requires the entry of the physical / energy characteristics and cross-section of the entire fluid path.

By necessity, the linear interval must be extremely small if l the report is to be useful. For this model, the intervals ranged from one foot for straight pipe, to one thousandth of  ;

an inch, in pump interiors. This model assumes a regular j gecmetric structure to the fluid, developed from principles  !

of synergetic-energetic geometry. The technique requires us j to compare the known physical dimensions of the containment,  ;

to a measured input of energy, either thermal or mechanical.

The comparison is achieved by normalizing the energy-density per unit-volume to a reference source, then normalizing the  ;

mass-density to the same reference source. In the final j step, the two ratios are compared to each other, giving a ratio-of-ratios (RAT). The closer that this ratio (RAT) is to unity, the more stable the system. Ratios greater than 1 normally indicate turbulent c o nd i t i o ns or cavitiation, while ratios less than 1 signify re-circulation or stratifled-flow.

MAJOR SYSTEMS:

• A. DRAIN COOLERS:

870911002807%77 ADOCK g

PDR PDR P

velocitieu through out the drain-coolers were below

10. f eel-per -uecond under all normal operating conditions.

The m.itter-energy ratio was also low. When we compared the mass normallned ( ha rd mode 1) to t. h e energy normalized ( soft model) w. - ,4 c h t e v e d .i r..itto (RAT) of 1.04. This represents a stable system. The maximum RAT occurred at the cooler inlet and is not significant. For this flow to be destructive, we would have to see a RAT of 1.6 or greater. This would occur if flows increased approximately 40% or the temperature rose by 86%, above the maximum design limits. This would also-occur if flows were at 33% of full-load and the temperature was less than 250 degrees F.

B. FEEDWATER HEATERS:

During normal to maximum flow conditions, the feed-water heaters have stable flow, including the inlat and discharge flanges. However, during partial load ( flow = %40 normal) we indicate a RAT of 1.7 to 1. 8 3 a t the first 15 degrees of are of the tube bends, and the last 30 degrees. This is cavitation, although not destructive unless sustained for long periods, such as 72 hours8.333333e-4 days <br />0.02 hours <br />1.190476e-4 weeks <br />2.7396e-5 months <br /> or longer. Because of tube geometry and the local temperature and velocity, this came close to the RAT of 4.66 that occurred at Surrey. If velocity increases 130 above normal, while temperatures remain sub-normal, an exaggerated flow pattern that we call a

wo r m" would result. This is a convoluted, triple vortex tuat is of sufficient strength to remove substantial amounts of pipe-wall, as happened at Surrey [ opinion of author]

This could occur if the system runs at low-load wl.h high-condensate flows, such as occurs during an increase in load, but it is seen as an intermittent phenomenon.

FEEDWATER PUMP:

These pumps are Byron-Jackson 14 x 14 x 17 DVSR, double-suction single stage pumps. The PUMPCOM model divided these pumps into 150 parts. As we said earlier, as a SYSTEM, Peachbottom 2 has a lower matter-energy density than a PWR, and consequently, exhibits different micro-flow patterns.

However, the pumps, internally, have the same matter-energy density as a PWR and exhibit the same micro-flow patterns. ,

In our opinion, an analysis of this regime, in general, is in order.

The pumps were stressed to three modes of operation.

These are:

1. low flow- defined as %30 of Best-Efficiency-Point: 2700 GPM

2. Normal flow- defined as 195 of BEP: 9000 GPM

1 J

l i

I T. high flow- defined a6 't.12 0 O f DEP : 10000 GPM 1.OW FLOW:

At Lh1s f1aw, the pump experienced blgh leve.u of inlet-eye recirculation, with a RAT of 3.88 to S.70. The RAT's. oscillated at a frequency of 3 hert and an amplitude of 13 mils. This is a sufficiently strong axial motion.to 3 cause premature bearing failure. The pump is equipped with fluid-film bearings which are.more tolerant of this condition .

than rolling-element bearings.

As a consequence of the inlet-eye recirculation, there is )

cavitation. This is concentrated along the vane / shroud l I

Junction on the top 4 inches of the vane tip.

NORMAL FLOW:

This pump was well-designed for this application. It shows no abnormal or destructive flow patterns at all. There is however; a vortex which forms in the inlet cavity during flow or temperature transients. This is not harmful by ,

itself. But, if the bearings are already damaged, this can cause seizure.

1 1

HIGH FLOW:

This pump ebould never be run at high flow. Cavitation occurs over 165 of the vane area. More importantly, there is an alternating area of cavitation in the area defined by the outer surface of the shroud and the casing wall. This In addition, a imposes axial pulse-loads on the bearings. I precessing vortex forms near the vane-tips and imposes a radial load. .i l

I CONDENSATE PUMP:

This pump is an Ingersoll-Rand Model APKD, seven-stage vertical can pump. These pumps have mixed-flow impellers, stacked and hung vertically inside a scaled can. This is done to insure adequate submergence and NPSH. These pumps operate in a specific speed range best served by radial-flow impellers. Because their impellers have a substantial axial-flow component, they are more prone to damage at low flows. This-pump was also tested in low, normal and high

• flow conditions,and the results were similar in all cases.

This pump requires an increasc in NPSHA of 15 feet.

Swirling of the water entering the first stage distorts the entrance angle 15 degrees, resulting in a pre-mature flow u__.___._m_ ____

1 i

1 ueparation. The dlhedral of the impeller causes high i velocities along the impeller outer edge and deceleration I j

along the inner edge This causes a flow reversal in the first stage The consequence of this is a small (about 112) I teduction in flow that causes cavitation in the following I stages At low flow, this pump cavitates in stages 2,3, and I

4. At normal flow there is inild, intermittent cavitation in stage 3 or 4. At high flow there is cavitation in stages 1, 6 and 7. Since a vertical can pump is not hydraulicly balanced, cavitation always results in radial loads and intermittent up-thrusts. The up-thrusts occur during flow or temperature changes, and always cause bearing wear / damage / seizure. Periodic visual inspection of the top bearing would be in order. f I

MAJOR VALVES A!iD PIPING:

No significant flow phenomena were produced at any power level. The maximum RAT recorded was 2.1 at high flow in a check valve on the feed and condensate discharge lines. This is not abnormal, nor damaging. No "Surry-type" worm was evident in any of the elbows. Particular attention was given to elbows following a "T" junction. RAT't did not exceed 1.35.

CONCLUSION: Iopinion1 The Peachbottom 2 condensate and feed system does not show any flow patterns that would be expected to lead to catastrophic-type failures, in general, nor any characteristics that could cause a surry type of failure.

While such micro-flow patterns were observed, they were inside pumps, where they manifest themselves as excessive bearing wear. This is readily observable and easily monitored with vibration or temperature probes.

Pampcom Inc. is grateful for this opportunity, and we ,

hope we will be informed of the results of any field testing of this plant.

RUDOLPH BEHRENS, president

.