Recirculation Piping to avoid cavitation in parallel pump operation

[MatAI]

We have three boiler feed pumps operating in parallel that discharge to a common header. The original pumps were installed in 1970, and water demand by the boiler requires two pumps to operate simultaneously, which appears to present a problem. The #2 Pump was replaced within the last year due to pump failure and a broken shaft. The cause was not investigated, but we suspect cavitation. Since the installation of the new pump, #2, the #1 Pump has failed due to outboard bearing demise and general pump excessive wear on the rotating element, totally worn, and the #3 Pump has failed due to the #1 0f 6 impellers being eroded to the point holes existed plus wear. We reviewed the files and found similar problems have existed with the sytem for quite some time. There was a recirc line installed originally, but the line was very small, 3/4," and shut off several years ago. We suspect cavitation and have investigated using pressure differential valves, larger diameter recirc lines for the original system, etc., and we would like to see if there is a routine or standard recric design one could use to prevent cavitation or one pump's flow affecting the other when pumps operate in parallel. I would appreciate any help received, and thank you for your kindness.

 

[TBP]

If you are getting cavitation, you'll hear it.

The only thing that I've ever actually seen break a pump shaft is starting the pump when it's slowly rotating in reverse due to the discharge check valve not holding.

Unless these boilers have on-off level control valves (a cheap and nasty method of control usually only used on small boilers) recirc lines are required for those times when the drum level is made, and the LCV's are closed. Having said that, a very large percentage of recirc lines and valves are grossly oversized.

You should go over the whole feedwater system. Make sure you've got the right pumps for the service. Make sure the suction piping is properly designed and installed correctly. (Botched suction piping is the root of much evil where feedpumps are concerned). Are the pump bases correct? There should be properly sized recirc valves/lines. Check the pump alignment.

At the end of the day, these pumps are 35 years old. It's probably time to send them to the scrapyard and get some new ones.

 

[vanstoja]

This is essentially the same query as thread407-10931 that you put in the Pump Engineering Forum on 12/1/04. See my reply there. Broken shafts and damaged radial/thrust bearings are not typically caused by cavitation. Only the holes in the 1st stage impeller are almost certainly cavitation related. Incidentally, cavitation can occur without audible airborne noise, particularly if air or dissolved gases are in the water which can cushion cavitation implosion damage and related shock noise.

 

[TBP]

If he's got enough dissolved air or gas in his feedwater to cushion cavitation related steam bubbles collapsing, then he's got a LOT more in the way of problems than just pump issues.

 

[MatAI]

In explaining the above, we've more or less concluded the problem is cavitation due to problems evolving from NO recircualtion of any of the three pump discharges. Although the pumps are supposedly the same manufacturer, design, flow, discharge, etc., the strongest pump rules and the other pump operating experiences cavitation My problem is how does one design or what are the devices used in a typical recirculation system needed to circumvent this problem, and what is the system layout. We've developed a design utilizing a "governor" type effluent regulator on each pump, and I think the system will work, but its design is much too sophisticated for a normal recirculation piping system, and the costs to make the device is high. Does anyone have a recirculation system on their pumps operating in parallel? If so, what is it like and what equipment do you use? I'd appreciate any help I can get, and thank you for your concern. Happy Holidays!

 

[KernOily]

In my opinion (since you asked for it ), I think you need to take a step back for a minute. Have you done a formal six-step or "five-whys" root cause failure analysis of the busted pumps yet? Have you done the fundamental calc to see if it's physically possible for cavitation to exist in your system?

Cavitation may exist when:

NPSHr <= NPSHa

NPSHa = (source pressure) + (elevation head of liquid above impeller eye CL) - (friction loss in suction line) - (vapor pressure of liquid) - (losses across any other devices in the suction line)

In my experience doing failure troubleshooting as a member of a multidisciplinary team, the operations guys and the pump repair shop will tend to hijack the failure analysis and usually cite "cavitation" simply based on the appearance of the damaged internal parts. You have to be ready with a correct technical analysis to say whether the conditions are right for cavitation to form.

Sorry if you already know all this, but as someone once said, knowledge of the basic fundamentals is the required foundation for successful professional practice. Don't cave in to the mechanics, pump shop people, operations foremen, and otherwise well-intentioned line supervisors who proffer cavitation when in fact the root cause may be something else; internal recirculation, for example.

I have been involved in a few pump failure analyses where intermittent operating problems, e.g. instrumentation failures, led to conditions that allowed cavitation to occur, when under normal operating circumstances the system is not anywhere near cavtiation.

You say a discharge recycle/recirc setup is needed for your off or part load situation. Usually this is done by one or more pressure letdown valves or restriction orifices, or valves and orifices in series, installed in a smaller line off the discharge header going back to the source tank (or whatever your suction source is). If you look in the Pump Handbook there are diagrams on how to do it.

Alternatively, a variable-speed drive may be in order if the economics will justify it. They are getting cheaper all the time.

Having said that, the devil of course is in the details. Your piping and instruments MUST be sized correctly and the programming for the letdown valves has to be correct or the system will oscillate about its setpoint or it will not work at all. Best to hire a consultant who has expereince doing this sport of thing. The discharge bypass has to be timed properly with the startup-shutdown sequence of the pumps. If there is a main discharge pressure or flow control valve, that has to be considered as well.

Just because your pumps are 30+ years old does not mean they necessarily need to be replaced unless they are over- or undersized for this application. There are plenty of pumps out there this age, and way older, that are happily operating every day, 24/7.

Thanks!
Pete

 

[davefitz]

You need to back up a bit and get some info from teh pump mfr and the site P+ID's or piping diagrams.

Usually the min flow neded to protect a constant speed BFP is about 25% of design flowrate. Each recirc line shuld be sized to pass 25% flow at a velocity of not more than 20 fps if CS pipe is used.

The typical pump config would be, for each pump: inlet FE=> Pump=>recirc tee =>check valve => MOV=> tee for 3 to common outlet pipe. The MOV would be closed for the spare pump to ensure no reverse motion when the motor is not energized.

In this config , and with a typical pump H vs Q curve, then there would always be flow thu an acive pump, even if one pump was stronger than the other. To demonstrate this , look at the Q vs H curves- for a common outlet H , there must be a finite Q thru each pump.

If the recirc valve is a spring loaded poppet, one must confimr the set pressure is cosnsitent with the 25% Q indicated onteh H vs Q curve. If the recirc is a control valev, then its signal must be from the individual inlet FE , set at min req'd min flow.

After 25 yrs service,there is likely to be differential settlement of the foundation and coupling misalignments. Try tsting with a vibration probe or some sort of instrumentation to record the shaft vibrations - soem experts can tell you in 2 minutes what bearing is bad or diagnose the pump , coupling and motor problem.