INCREASE THE PRESSURE IN PUMP DESIGN 3

[TechSupport1]

We have a reflux drum from a tower that operates under vacuum. The NPSHr by the pump from this drum is not being satisfied by the NPSHa of the system. We made a decision to install a pump with a slower speed motor. With the slower speed motor, the discharge pressure will be less than the current pump which has a 3600 rpm motor. To boost the pressure from this new pump, we have identified 2 viable options:
1. install a booster pump (i.e. the new pump [with the slower speed motor] would discharge into the suction of another pump that would boost the pressure to the system requirements). A control valve would be installed downstream of the second pump.
2. the pump with the slower speed motor would pump into a vessel with the flow controled by a control valve. From the outlet of this vessel, a second pump would be installed which would increase the pressure to the system requirements and a second control valve would be installed on the discharge of the second pump.

What would be the advantages and disadvantages of each option? From a pump expert's opinion, which is best option considering cost, operation, and any maintenance issues?

 

[sshep]

Presumably your reflux drum is at a low elevation and this is constraining your options. The vessel and associated controls, relief, etc will obviously make option #2 cost significantly more for no significant advantage. If we have to choose between your choices I would go with option #1.

You might check out

http://www.rothpump.com

which we recently used in a low NPSH vacuum situation similar to yours. Their specialty is a regenerative turbine pump requiring only 1ft NPSH which can reach any head you are likely to need. There may be some capacity limits (i.e. 150gpm). The otherway we tackle this problem is to sink a vertical pump into the ground (hot well).

Although two pumps in series to solve this problem seems excessive on the surface, you may well be able to reap an energy benifit from 2 pumps if the head requirements of the reflux and distillate are such that only one requires boosting. An energy efficient design would require two pumps anyway.

best wishes, sshep

 

[Montemayor]

TechSupport1:

You haven't mentioned the following:

1) Raise the reflux drum and use the slower reflux pump; this involves capital, but the maintenance will go down and the operability and instrumentation will be simple and dependable.

2) Utilize a 2-stage centrifugal, running at slower specific speed which will allow for the NPSHa and still produce a higher pressure through the 2nd stage. This eliminates any additional drums and all the related instrumentation and maintenace. The 2-stage model doesn't increase the maintenance, but it is more expensive than one single stager.

I would always steer away from a separate, independent 2nd stage pump; the instrumentation and maintenace is higher. I also would not employ another, intermediate tank with more instrumentation for the same reasons. The reliability and operability of a process starts to decay when one starts to add more, separate equipment.

I hope this experience helps.

 

[rzrbk]

Another option may be to keep the speed at 3600, but install a double suction pump. There should be a significant decrease in NPSHr, unless that is what's there now.
A double suction between bearings pump will be more expensive than a standard overhung pump but, similarly to the double stage pump above, could avoid the need for a second pump and/or intermediate vessel.

 

[TD2K]

I would definitely stay away from adding another vessel, it's going to be far more expensive than using your existing pump (if none of the other suggestions above are feasible) as a booster pump for the new pump.

 

[rmw]

What are your current NPSHa and NPSHr and design flow conditions?

Specific details might help us come up with specific recommendations for some of the ideas already presented.

rmw

 

[TechSupport1]

I appreciate all your comments; however, I was looking for a stronger argument from responders. I would like to know all the reasons why you would choose one option over the others. Thanks

 

[StoneCold]

Techsupport1
You got some awesome answers for as little information as you gave. What about YOU comparing all the options and then making a report back on why you chose one of them?

Goodluck
StoneCold

 

[sshep]

Sooo, you want to know which of the two options you gave us is better. Here is a stronger argument for option #1.

Your first option has the advantage of being more sensible because your second option is silly: You are using two control valves where one would suffice. You are buying and setting a vessel you don't need. You will need a relief valve to protect a vessel you don't need. You will require level controls for a vessel you don't need. You will require vent piping and probably pressure controls as well for a vessel you don't need. You are wasting your company's capital. You are adding extra components which can fail. Every responder that addressed your question basically told you that using an intermediate vessel was the worse of your two ideas. You didn't get any advantages cited for option #2 because there are none in the context in which you asked your question. Your first option is the best choice of those given by default.

best wishes, sshep

 

[TechSupport1]

I thank you all for your responses. For the pump project discussed above, I am working with a few other people (my peers but, unlike me, do not have engineering degrees) in deciding which option works best for the plant. I tried to describe the options in the most unbiased way I could but I feel strongly that the booster pump is the way to go, even before you all responded. A pump vendor (who has done a great job in our plant) has reservations about the booster pump. I have repeatedly asked for definitive reasons why it won't work and have gotten nothing solid - only a subjective negative feeling that it is not a good idea, eventhough he admits, theoretically, it should work. I needed someone other than myself who has pump design knowledge and experience (preferably a pump expert) to make a good argument in my favor without me leading you on.

Raising the reflux drum has been discussed as an option but it is just not practical for our situation. Running 2 pumps in parallel has been done but the suction piping needs to be identical to get a good flow split between the 2 pumps - also not practical due to space limitations and we would like to use one of the pumps as a spare instead of running 2 pumps all the time - also space limitations to add a third pump. I like the Roth pump idea, which I did not know about. I contacted them and got a quote. It looks like this type of pump would work; however, the price is a little on the high side. The new (slower motor speed) pumps are less than 1/2 the cost of the Roth pumps; the booster pump would be our existing (3600 rpm motor) pumps.

 

[TD2K]

There are plenty of installations out there with a booster pump in front of the main 'charge' pump. The vendor needs to come up with a better explanation why he doesn't support this idea than just 'because'.

 

[sshep]

A booster pump should operationally work fine as long as you are not planning to put a control valve in suction line of the second pump (obviosuly the control valve should go downstream of both so there is no chance of starving the suction of the second pump). A booster pump may seem like a new idea to your vendor so his objection is probably because of "unforeseen problems with an installation of which he has no experience". I can relate to that, but on the otherhand there are many such booster pump installations in the industry which work fine so this is is an unjustified apprehension.

The downsides of a two pump scheme are qualitatively related to reliability, cost, etc rather than whether or not it will work. You seem to be doing the right thing by checking out as many options as seem to have merit. A retrofit of this nature can seldom achieve the perfection of a new installation (i.e. design for more NPSHa), but success is measured by whether or not the problem is solved. Good luck, sshep

 

[chemELK]

One other option would be to use a pump that has suction lift, such as a progressive cavity or twin screw. I'm not sure how the price compares with Roth, but they're used in the oilfield rather frequently. Since it appears you already have a charge pump, this isn't likely to be your best option, but keep it in mind in the future.

Just one caution on the progressive cavity pumps - do not use them with clean water! They work great on organics and oily waters, but fluids that don't provide at least a little lubrication (read clean water) will allow the rotor and stator to heat up from the friction.

 

[Holly41]

One of the things you asked for was an opinion from a "pump expert". In my experience, if you get 5 "experts" to respond, you'll get 5 different opinions. There is no one right answer. When I work on a system, I try to keep it simple. You state that the Roth pump was an attractive option, but it was too expensive. What do you think the cost of adding another pump to the system will be? Not just initially, but Life Cycle Cost. I'll tell you right now, that cost will be higher.

If you want an expert opinion, you will have to provide more information. What is your NPSHa? What is your required TDH and GPM? If pressure is more important than flow, maybe you ought to look at a positive displacement pump (high pressure/low flow) rather than a centrifugal pump (I assume that is what you have; you didn't say). There are all kinds of options available, but I would stay away from the 3600 RPM driver if possible. The motors are cheaper, but the higher speeds ALWAYS mean higher maintenance costs.

 

[Montemayor]

Holly:

Good thinking; good logic; good knowledge; excellent communication.

Keep up the great comments.

 

[Artisi]

Holly is correct, to get good info from the pump guys you need to be more specific so that we have something to work with.
What is the NPSHa / NPSHr , is the flow constant or variable, if variable it will changes both NPSHa and NPSHr.

you have said that you intend to install a pump with a lower speed -- Why ?? firstly look at what options are avaiable to increase NPSHa - can you change the pipe work configuration - find a pump with a lower NPSHr requirement while using the maximum amount of existing equipment, ie, motor, control gear etc., if not available then look to other options as already suggested by others.

Naresuan University
Phitsanulok
Thailand

 

[djack77494]

TechSupport,
As you've seen said repeatedly, without complete information regarding the problem we cannot supply good, accurate solutions. Having said that, I'll talk a bit in generalities. Unless you have an expensive pumping system that can nearly perform as desired, I would suspect that your best bet would be to replace the existing pump with one pump that can do the required job. Why use two pumps in series, with all the electrical, instrumentation, plot plan requirements, etc. associated with this configuration? This approach is filled with dangers that can be and routinely are mitigated with lots of instrumentation, but what's the point? (One big danger would be protecting the downstream pump against shutdown of the upstream pump.) If a single pump can do the job, use a single pump. If it costs a bit more than you'd like, think of the much greater costs associated with a high degree of complexity. Using a second vessel - Don't even go there.

HTH,
Doug

 

[UmeshMathur]

I take it that this is an existing installation, since you cannot raise the drum.

Two questions need answering immediately:

(a) Has the NPSH problem existed from the very beginning, or have you increased flow through the pump so much that the additional friction losses in the suction line reduce available NPSH to a value below the (now increased) minimum NPSH required? Will increasing suction line size (the cheapest and simplest option) not solve the problem?

(b) Has the existing pump deteriorated mechanically, over time, to where its minimum NPSH requirements have increased beyond what was required originally when it was new? Is a pump replacement all you need?

It seems to me that this whole discussion could have been helped enormously by an explanation as to precisely why you "made a decision to install a pump with a slower speed motor", and thus invited problems with insufficient pump head. If the existing pump requires a higher minimum NPSH than available, and increasing suction line size is not the answer, the sensible answer is to get a new pump that needs a lower NPSH (while delivering the required flow and head) as advised by djack77494. How can that extra cost exceed the installed cost of all the extra equipment you propose with both options?

It seems that 100% of your problems arise from the decision to lower pump speed. Therefore, I question seriously the wisdom of taking this step. In my opinion, it would be sheer folly to add a pump in series, extra piping, control instrumentation, shutdown logic to protect the second pump if the first one goes down, etc., when probably all you need are either a larger suction line, or a low NPSH pump, or both.

Therefore, my vote is to scrap both your options, answer questions (a) and (b) above, and rethink the proper solution.

In this context, I would also state that the opinions of those who are not qualified as engineers, and lack the experience to make up for that deficiency, should not be allowed to overrule the proper engineering solutions.