Displacement and centrifugal pump in series? 1

[MartinLe]

This one came up a while ago when talking with a customer, and I fear the issue will pop up again.

Here's the situation: We build a biogas plant, the liquid digestate/slurry is to be applied as fertilizer. We use centrifugal pumps that are sitting directly at the storage tanks to fill the slurry spreader wagons.

Now, our customer has wagons with a rotary lobe pump installed to draw digestate from tanks. Depending on the way we connect the wagons both pumps would be in series. I think this is a bad idea and would like you to check my reasoning:

The lobe pump (that is not specified by us and will be provided by a partner of our customer) will have to work at the same flow as the centrifugal pump, if it is over- or undersized something will be damaged.

My second question is - how ould you actually calculate such a system?
Pumps in series means added head, same flow - but the flow of the centrifugal pump will be far more dependent on the head, so I need to know what "share" of the head each pump will bear. Am I correct so far? How do I arrive at this?

 

[Artisi]

Can we assume the centrifugal pump is higher capacity than the lobe pump?

This being the case, the centrifugal pump will just see the lobe pump as a restriction in the discharge line and will move along its performance curve to the flowrate that the lobe pump is able to pump at its operating speed. The lobe pump being "positive displacement" can only pump "x" flow at the speed of operation.
The only real problem that first comes to mind is if the centrifugal is forced to run too far left on its curve you may have shaft deflection, seal, vibration problems, internal re-circulation and accelerated wear - how about supplying the performance curves of the centrifugal and the lobe pump, this will give a better understanding.
There might well be a few other minor problems which crop-up as the discussion develops from those with experience with lobe pumps (not my area of expertise).


It is a capital mistake to theorise before one has data. Insensibly one begins to twist facts to suit theories, instead of theories to suit facts. (Sherlock Holmes - A Scandal in Bohemia.)

 

[bimr]

One would expect the centrifugal and rotary lobe pumps connected in series to pump at the flow rate of the rotary lobe pump since the rotary lobe pump is a positive displacement pump.

Obtain the centrifugal pump performance curve and determine where the centrifugal pump will be operating. The farther you get away from the BEP, the more potential there will be for pump problems.

 

[DubMac]

Is there some reason you can't just bypass the lobe pump and have the centrifugals load the wagons??

 

[psv1990]

I would recommend by passing the the loble pump and filling the tank with just the centrifugal pump. The lobe pump is likely sized to operated a slurry injection system on the wagons. This requires higher pressures and lower volumes. The maximum flow rate will be based on the lobe pump which is likely operated by a hydrolic motor and then pump speed is dependent on the engine speed and will vary with each truck and operator.

The final issue is that the lobe pump is required to reverse for loading. In the event of operator error failing to reverse the to lobe pump operation may cause catistrophic failure due to over pressure.

Due to these couple of issues that i can think of and the many that i have not yet considered, I restate that it is not a good idea to run the pumps in series. If you insist on using the lobe pumps to load the wagons, do not put pumps on the storrage tanks.

 

[MikeHalloran]

I wouldn't ask the customer to modify his wagons, e.g. bypass the lobe pumps, because he's the customer.

I have used a somewhat similar arrangement, a centrifugal pump to supercharge the suction of a very large hydraulic pump, so as to provide the hydraulic pump's NPSHR without pressurizing the reservoir. That was 40+ years ago, and I think the systems are still in service. Those pumps were matched, as in checking the OP of the centrifugal throughout the flow rate range of the big swashplate pump.

In this case, the series connection should be checked for corner conditions, but in general, as Artisi and bimr described, the centrifugal pump will adjust, and the world won't come to an end. If the flows are widely disparate, or the combination will drive the centrifugal into an undesirable flow regime, the solution is as simple as switching the centrifugal off, and letting the lobe pump just draw through it.





Mike Halloran
Pembroke Pines, FL, USA

 

[bimr]

Nice answer, Mike.

 

[Artisi]

Just one point for clarification, you are connecting the centrifugal pump to the inlet of the lobe pump, or is the lobe pump reversible?

It is a capital mistake to theorise before one has data. Insensibly one begins to twist facts to suit theories, instead of theories to suit facts. (Sherlock Holmes - A Scandal in Bohemia.)

 

[MartinLe]

Thank all of you for your input.

I think this will be the route to go:

* specify a range of flow rates close to the OP of the centrifugal pump, if the lobe pumps are within it's ok
* else, we will need to bypass the lobe pump

Just to give you some context:
those wagons have an opening on top, the only issue there could be emissions (smell). I don't see a huge issue there.

Again, thanks for the detailed inputs.

 

[Artisi]

Do you have a curve for the centrifugal pump?

It is a capital mistake to theorise before one has data. Insensibly one begins to twist facts to suit theories, instead of theories to suit facts. (Sherlock Holmes - A Scandal in Bohemia.)