Question regarding situation involving split case pump and cavitation 6

[PumpDude]

I have 6 single stage split case pumps in a chiller system. Half on the evaporator side and half on the condenser side. All pumps are identical in construction and configuration, i.e. impeller diameter, speed, motor hp, etc. One set pulls water directly from a cooling tower through what has been assessed as a excellent design according to HI standards. (more than 10 pipe diameters from the pump suction to the elbow/side outlet T, eccentric reducer with flat side up, etc.) The other set of 3 has the same plumbing configuration on the suction side and only has two differences from the other. On the discharge plumbing an orifice plate has been added to add head and the water supply is from the discharge plumbing of another system. I am experiencing the same problem with these NEW pumps. All exhibit cavitation and after pulling the top of one of the pumps I am finding circular pitting (appears to be hit with a ball peen hammer....and I am being told by the mfg these pumps are not cavitating!!!!!!!)on the volute on both sides of the split leading up the the eyes of the impeller on the suction side. The discharge side as well as the impeller itself shows no damage. There also appears to be a small amount of "blow by" between the case/volute and casing wear rings (lending to the theory of internal recirculation). I have only pulled one top but all pumps had the same sound signature and vibrations plot. We choked the discharge from wide open to nearly completely closed with little to no change in the sound or vibrations. The pump we inspected the internals on was sent to the mfg. and tested in their lab to no avail. We provided all of the current operating parameters and it appears they were close on most, except the plumbing on the suction side. They have stated they heard none of the noise and supplied vibrations data showing nothing like we found in the field. The differences in the suction plumbing was that we have an isolation (butterfly) valve, then a check valve, then an expansion joint, the reducer up to the suction flange of the pump. They had what appeared to be an approx. 6 foot run of pipe from an elbow into the suction of the pump. Prior to the elbow was their isolation valve. Having not repeated our symptoms in the lab they began offering opinions on why this is happening in the field. We have discussed and dismissed NPSH issues and suction plumbing design faults. We have also discussed disolved oyxgen, entrained air, internal recirculation and flow vortices. We are currently waiting on an analysis on the fluid and I will post those results. I appologize for the length of this post, but I wanted to convey all that we have done and found so far. I am hoping that some one out there has had this happen and can tell me what was done to alleviate it. We are getting nothing from the pump mfg. and the rest of us (rep from the distributor, pump engineers from my home office, myself and personnel here on site) continue to "scratch our heads" in frustration. Thanks in advance for taking the time to provide me with whatever insight you can.

 

[PUMPDESIGNER]

Hey, this is fun stuff since it ain't my headache today.
" We choked the discharge from wide open to nearly completely closed with little to no change in the sound or vibrations."
If, in the above statement you know for a fact that the flow rate varied widely during that test, then cavitation from several sources is eliminated, namely suction recirculation cavitation, suction side cavitation (Low NPSHa), and discharge recirculation cavitation, ALL are eliminated by that test.

But, the flow rate must truly have gone from very low to very high, AND you have to be sure that you were hearing it correctly because that is subjective data and only an experienced "cavitation" listener would be reliable.

You appear knowledgeable and capable of doing the calculations, and this appears to be a top quality installation and engineered well too. Therefore, I ask this, are the calculations for all pump characteristics stacking up as predicted under actual operating conditions? Head and Flow seem on the money, Vacuum readings are close to what is calculated and sufficient for NPSHr with the appropriate margins, etc.?

Also, what is the Specific Speed and Suction Specific Speed of these pumps?

PUMPDESIGNER

 

[25362]

I'll start by repeating I'm not a pump expert, just a user.

I have been told that pump manufacturers sometimes design large eye impellers to reduce NPSHr. And that -in these cases- when flows are at 60% of the BEP hydraulic problems, such as localized cavitation around the impeller eye, are magnified.

The apparent cause of the creation of low pressure areas being that the impeller inlet vane angles do not match inlet fluid angles when at reduced flow rates.

I don't know whether this corresponds to the case in hand; anyway I'd appreciate reading (and learning from) the worthful comments by the experts.

Good luck.

 

[BobPE]

25362:

You talk a lot about the pumps, how about helping us understand your system. The devil is always in the system. As far as the pumps, if they are still spinning, then they are working fine. The system the pumps are in is most likely killing them.

Let us know the system flow and head requirements and what the pumps system head and flows are. From the description of the damage, it sounds like there is too much flow moving through the pumps, causing suction cavitation damage.

Pump reps are good with the brand of pumps they rep, but should not be relied upon to evaluate the system in which the pump is designed. It sounds to me this is a costly event for you, and investing in a good pump engineer could help (in addition to utilizing us wonderful experts here in the ENG-TIPS forums!!!!!!!)

let us know how you make out...

BobPE

 

[25362]

BobPE, are you sure your post was intended for the attention of 25362 ? Or perhaps PumpDude ?

 

[BobPE]

YIKES 25362...yes, it was intended for PUMPDUDE....sorry for the confusion....

BobPE

 

[vanstoja]

From your description, I infer that the three (parallel?) pumps that appear to be cavitating are operating at reduced flows compared to the three non-cavitating pumps due to the orifice in the the discharge of the bad set. What are the flow fractions of the cavitating vs non-cavitating pumps? At flows around 60% of best efficiency flowrate, impeller blade incidence angles may be in the 12 to 18 degree range which is conducive to flow separation (and to cavitation if available NPSH is based on near-bep flowrate where inlet incidence angles are generally low (0-4 degrees)). Another possible contributor to cavitation is intensely swirling flow from the external pumping system feeding your set of cavitating pumps. Even 10 pipe diameters of straight run to all six of your pumps may not be sufficient to provide low incidence conditions to the impeller blades to prevent separation-induced cavitation if swiwling flow exists upstraem of the 10 suction pipe diameters. For example triple back-to-back 90 degree pipe bends need something like 188 pipe diameters to dissipate the swirl induced by the bend combination. If the undescribed pumping system feeding your cavitating pump set has multiple pipebends or poor pump discharge conditions that will induce swirl then your presumably ideal suction piping may be utterly worthless without positive efforts to control the inlet swirl.

 

[PumpDude]

Howdy all and thanks for the beginning of what I am hoping is the end of this issue.....and yes pumpdesigner.....it is a MAJOR headache!!! And remember misery loves company!! :-D

Anyhow....to add more to the discussion and answer repsonses for additional information.

The following are the operating parameters:
Flow = 1350 GPM
TDH = 70 feet
Speed = 1750 RPM
NPSHr = 11.40 feet
S.G. = 1.0 (water)
Motor hp = 30 hp
Specfic Speed = 2,656.82
Suction Specific Speed = 10,370.16

The evaporator pumps have a pressure range on the suction side of 1 psi to 5 psi with the mean pressure of 2.725 psi. The range on the discharge side is 20 psi to 31 psi with a mean of 25.125 psi.

The condenser pumps have a pressure range on the suction side of 11 psi to 12.5 psi with the mean pressure of 11.833 psi. The range on the discharge side is 25 psi to 26 psi with a mean of 25.5 psi.

The water both of these systems are utilizing is "post cooling tower" water and would normally range at this time of year between 50 and 65 degrees Fahrenheit.

And to clear one misconception.....vanstoja, all six pumps are cavitating. This is part of the major "head scratcher"...two COMPLETELY different suction "systems" but identical performance problems. The only similarity on the suction between to the two systems it the layout of the components (reducer, isolation valve, etc.), other than that they are VERY different....one drawing directly from the cooling tower and the other pulling from the discharge of another system.

Ok....have at it my friends....I anxiously await your insight!!! And again thanks in advance for taking the time to help me out.

 

[vanstoja]

Please clarify the so-called "operating" data. I presume that you haven't measured all of the non-derived values like head, flow and speed and that these are either design point values or manufacturer test parameters from his test facility. NPSHR is presumed to be from vendor tests at the stated 1350 GPM flow with varying suction pressures to determine the 3% head drop sucton pressure. You've apparently measured suction and discharge pressures in psi for all 6 pumps in your facility. My calculations show the evaporator pumps to be developing an average pressure rise that is 73.92% of the presumably rated head of 30.303 psi(75 ft at SG=1). Condenser pumps are averaging only 45.1% of presumably rated head. This tells me that both sets of pumps are actually operating far off the best efficiency flowrate on the high flow side. You'll have to look at the manufacturers head flow curve to determine what flow fractions (more than 100%) are actually being pumped. If your cited NPSHR value is for something near 100% flowrate the actual NPSHR will be much higher at the higher flowrates your pumps are apparently delivering. If the actual flow fractions are approaching or more than 120% then severe cavitation is a distinct possibility. Did the manufacturer run NPSH tests at higher than rated flows? If so, get the results and compare them to the available NPSH for your two pump sets.

 

[PUMPDESIGNER]

With Nss = 10,370 I would look hard at Incipient Cavitation as the problem.

First,
NPSHr values do not have the same meaning for Nss=10,370 pumps as they do for Nss=3,000 pumps. High suction energy pumps take in water so efficiently that by the time the NPSHr value of 3% occurs, significant cavitation is already occurring, sufficient to damage the impeller quickly. On the other hand, on low Nss pumps, the 3% value is directly useful as a predictor of cavitation.

Second,
It is well known that high margins of NPSHa over NPSHr results in incipient cavitation damage, especially when cold water is involved. It is often better to have a small margin of NPSHa over NPSHr to reduce incipient cavitation damage. In those cases where this is true, the rate of incipient cavitation damage increases proportionally with increasing margins of NPSHa over NPSHr until the incipient cavitation point is passed where there is no cavitation at all, which is usually hopelessly high.

Suggestion, reduce NPSH on a pump to see if noise goes away, although going by sound is an iffy method relying heavily on a subjective judgement.

Sources to study would be The Pump Handbook and the HI materials on this subject (Incipient Cavitation).

PUMPDESIGNER

 

[25362]

I once read that to bring severely cavitating centrifugal pumps back to borderline or incipient cavitation one expert advised to add heat (!)

His explanation rested on the fact that the cavitation severity is reduced when the vapor/liquid density ratio comes down, all the while pushing cavitation into the borderline region, consequently diminishing its severity to an acceptable level.

As an example, an increase from 50 F to 70 F for water will decrease the density ratio by 50%. All this, of course, assuming temperature increases must have associated pressure levels to maintain the liquid phase.

The same expert added that a correction for cavitation in centrifugal pumps could be attained by the addition of an inducer, which is a small booster pump that creates a small head in their suction. Pump manufacturers design inducers driven directly from the pump shaft for delivering fluid to pumps' suction to reduce cavitation.

 

[PUMPDESIGNER]

Hey, where did everyone go on this subject?

PUMPDESIGNER

 

[PumpDude]

Nothing further from anyone??

Still no results on my part from the dissolved oxygen testing. I think this is more of a red herring than explanation...but that's just my gut instinct.

Thanks pumpdesigner for the info and I am further investigating a couple of articles on incipient cavitation.

Thanks again.

PumpDude

 

[ApC2Kp]

PumpDude,
Since these are new pumps, and probably new piping, there were probably some temporary startup strainers inserted in the the suction inlet spools. The pressure measurements could be taken upstream of the strainers, and you might not detect blockage of the strainers. Temporary strainers could be one of those items not documented in the drawings.

 

[PUMPDESIGNER]

Ahh, the voice of experience about temporary startup strainers, good idea ApC2Kp.
We put those into all our pump intakes, but we also put on a bright red warning decal to warn people.

However, if would be hard to see strainers on all the pumps clogging up, and also that would be hard to figure if the strainers were on both suction side and discharge side pumps moving the same water.

PUMPDESIGNER

 

[katmar]

Hi PumpDude,

Sorry to jump in so late. I do not believe you will get good answers here until you give all the info. What make and model are the pumps? Where can we download the curves? What are the speed and impeller size? Detailed description of piping and fluid? Have you measured the flows and heads? How do they match the design?

Without this info all you will get is a lot of theory and suggestions, and no real solutions.

 

[ApC2Kp]

To Pumpdesigner,and Pumpdude,
It is puzzling that all six pumps have cavitiation, and I was trying to pick at anything that would be common for the six pumps. If there are startup strainers in suction of each pump, then hopefully strainers are conical type and not perforated flat plate. The perforated plate temporary strainers are cheap, but have only 40% open area, so they could be significant restriction to flow even without the cooling tower packing, trash, gravel, 2x4 lumber, bolts, nuts, hardhats, or lunchboxes that plug up startup strainers. Startup strainers should have tab / decal to identify locations for later removal, as pointed out by pumpdesigner.
Few other items that could be common for the pumps. Is the driver rotation correct? There was a split case pump which had capability for driving from either end (Gould 3415). After we had to re-arrange piping and flip pump + driver to avoid horizontal elbow on the inlet, the vendor went round and round with us about the direction of rotation for pump. Could that also explain the discharge pressures?
Are there any valves near pump suction? Reduced port of a valve could cause restriction or turbulence.
Finally, a stretch to consider if high altitude might cause NPSH margin problems for the cooling tower pumps.

 

[PUMPDESIGNER]

ApC2Kp - All good points. Especially the suction strainer, people would be tempted to put in a cheap perforated metal plate and you are correct, that would reduce open area far too much, but I would consider that in very poor tasted because the pumps could never work correctly even when the strainers are new. We use Type 316 Wire #10 mesh usually, that allows full flow but stops the big stuff from getting through which is all you are trying to do.

I hesitate about rotation becuase it aggravates people, but you are correct because simple things are overlooked, and in this case there would be good reason to overlook it except that the pump casing should have the correct direction arrow on it. Besides, only inordinately proud people resent someone double checking their work.

Anyway, I think you did a good job. If none of your stuff stirs up a problem, then I am sticking with Incipient Cavitation, very bad boy to deal with.



PUMPDESIGNER

 

[ApC2Kp]

PumpDude,
Next time you have a casing open on one of the pumps, take look at the impeller. Please have the vendor verify that the impeller vanes are oriented correctly (backward-incline?). If the split case design allows set up for a driver from either end, then the impeller probably flips around to accomodate the different rotation. The rotation check against the arrow on the casing, as pointed out by pumpdesigner, needs to be followed. After thinking about the split case design, it seems an external rotation check still might have a possible error. It would be horrible to find incorrect assembly of a pump impeller to shaft, but with some of the pump companies being bought up, merged, and moved, it may be resulting in assembly errors and low quality.

 

[PumpDude]

Hi folks!!

Still no feedback on our dissolved oxygen study...but as I stated before...red herring I think.

Ok to address some of the questions/suggestions (which again I can't tell you how much I appreciate!! )

I wish I could say that strainers were the issue because I would be at "beer:30" right about now!!! Unfortunately, I observed all installs and there are no strainers....temporary or permenant.

I have had the "top" off of one pump...the one that went back to the mfg. The impeller was orientated correctly and driven in the proper direction. And see my earlier posts for description of what we saw internally as to the damage.

Katmar....I will post replies to all your questions today or tomorrow....

I did find an interesting article dealing with incipient cavitation (thanks to pumpdesigner!!) detailing how NPSHR and NPSHA ratios and values effect pump performance and how design costs for "necessary NPSHA" for "no" cavitation might out weigh repair costs in time, etc. Was horribly under-informed on this topic.

Yes ApC2Kp, there is a valve VERY close to the suction flange of the pump. Let me explain...

Let me begin from the pump discharge flange and work my way towards the cooling tower. (source) The very first piece is the expansion joint. The next is the eccentric reducer. The flat section is on top. Then I have a butterfly isolation valve. This all happens within 4 feet of the suction of the pump. I wish I could post a picture here....I don't have anywhere to make a pic available to ya'll and I have some jpegs of these pumps and the systems.
But I understand the storage space reqs to have pics as well as all the discusion threads. Oh well c'est la vie...

I will follow up witht he additional data soon...

Thanks again,
PumpDude