Very Very Non Standard Problem in a multistage centrifugal pump

[sashi30102014]

Hi All,

I am here posting a very mysterious problem we are facing in a new multistage centrifugal pump.

Background is, we replaced existing centrifugal pump with new centrifugal pump whose specifications are Flow rate: 250 m3/hr, differential pressure 75 kg/cm2, with normal suction pressure available 3.5 kg/cm2(g), 9 stage pump. Fluid pumped is crude oil. Suction line size being 18" and Discharge line size is 12", pump nozzles are suction 6" and discharge 4", where suitable expanders are used for connecting with pipe line. Pump maximum shut off head is 108 kg/cm2(g). Immediate downstream of the pump after check valve, we have MOV - which starts opening along with pump start & full opening time of MOV is 30 seconds. Fluid SG is 0.83

Problem is, when we start the pump discharge pressure goes abnormally high (beyond 108 kg/cm2g), this pressure lasts for hardly 3 seconds and pressure drops to 92 kg/cm2g. Even, we are not able to measure this shooted pressure exactly due to very less time it exists, because this event occurs for less than my pressure sensor scan time (5 Seconds).

Now, the practical problem is we have PSHH on downstream of pump. This Pressure switch high high is more sensible than my transmitter and this trips the pump due to pressure high high. This switch set point is 96 kg/cm2g.

The pump is brand new one, France made. We have analyzed all process conditions and checked the RPM and pump casing temperature rise etc., and found all are normal. We have not opened the pump and analyzed it, as we are not pump experts and been discussing this issue with the manufacturer - who is also not able to find root cause, as the centrifugal pumps are not supposed to rise the pressure beyond shutoff.

Anyone can help us, what can cause, in a centrifugal pump to rise discharge pressure to abnormal level for a very short period of time, only at the time of starting the pump.

Sashi Kumar. M
India.

 

[LittleInch]

Sounds relatively straight forward to me. You are experiencing some sort of surge event.

Also it is much more normal to refer to pump data in head terms rather than pressure. Can you post the pump curve so we can understand what is happening. Please draw a schematic and add data.

I suspect this is really coming from the inlet conditions. Can you describe the pressures and what is upstream of this pump.

Is the inlet valve fully open before you start the pump?
Is the pressure stable?
what is pump inlet pressure at no flow (just before pump start)?

what type of electric start is it (DOL?)

Having a PSHH set at 96 and a max discharge head of 108 seems like a recipe for disaster. Why is PSHH set below max discharge pressure trip?
what is the design pressure of the discharge pipework up to the MOV?

My motto: Learn something new every day

Also: There's usually a good reason why everyone does it that way

 

[BigInch]

On startup electric motors can develop up to 150% of normal running torque, so brief overpressures are possible .
some things you might try,
1 starting with recycle on and discharge valve closed.
2 add a "cycle-stop" valve.
3 You might also try adding a soft starter.

you must get smarter than the software you're using.

 

[LittleInch]

You might also think about an accumulator / surge vessel.

Starting pumps with flow possible can mean that the motor doesn't get to speed, but if you've got a beast of an electric motor compared to the shaft power required that might not be an issue.

What is the max shaft power and rating of the motor??

We can't do much more until you give us some data.

My motto: Learn something new every day

Also: There's usually a good reason why everyone does it that way

 

[sashi30102014]

Hi all,

Thanks for the replies.
Attached are the Schematics, pump performance curve and FAT report. Please see answers for your queries,

1. Pump suction is flooded. For suction, Oil is coming from the separator vessels, the vessel pressure is constant. At suction nozzle of pump, normal suction pressure noticed is 3 kg/cm2g and max is 4 kg/cm2g.
2. At Inlet of pump, we have an MOV which is fully opened before the pumps starts. We ensured this.
3. Upstream pressure is always stable.
4. Before the pump starts, suction pressure noticed is 3 kg/cm2g and remains same in continuous running of the pump.
5. Type of start is DOL. Rating of the motor is 790 KW, we cannot go for softstarter for this higher rating, also electrical feeder were not in our scope.

We are discussing the problem with the pump bearing Tag No. P -1390. I like to add an another note, there is also a standby pump (P-1380) to this pump (P-1390). Both the pumps are new, procured from same manufacturer in same batch of manufacturing. P 1380 is already commissioned and no such surge pressure problem is faced with this P 1380.

6. PSHH setpoint: Shutoff head noticed in FAT is 1072 m and with available maximum suction 4 kg/cm2g, with SG of 0.83, maximum discharge head can be around 1100 m which corresponds to 91.5 kg/cm2g. Hence we have kept setpoint as 96 kg/cm2g, irrespective of max shutoff head 108 kg/cm2g. Downstream pipeline are designed for 150 kg/cm2g.

7. Rated power of the pump is 671.44 KW, Maximum power @rated impeller is 765 KW and Driver-electric motor rating is 790 KW.

Since it is offshore platform, we cannot think of additional vessel (surge or accumulator)due to space constraints.

Hope the above is clear and provided attachments gives a clear picture.

Sashi.

 

[LittleInch]

No schematic attached.

Is the P1380 identical to 1390?

This is an interesting issue now we have more information, but if the piping is rated for 150 kg/cm2, why don't you just disable the PSHH for 5 seconds when the pump starts?

Also does the MOV start to open at the same second that the motor starts to turn? Maybe delay the MOV opening by a few seconds. If an absolutely identical unit behaves OK, then it is more likely to be some configuration of the piping or the start sequence being slightly different.

My motto: Learn something new every day

Also: There's usually a good reason why everyone does it that way

 

[sashi30102014]

Hi,

Sorry, missed to attach schematic. Attached now.

P1380 and P1390 are identical in all way. Yes, our plan is to provide some temporization of PSHH, may be 5 seconds or more. But, we like to know, what is happening ? why pressure is shooting for such a short time during starting.. is it some mechanical problem with pump ?

When Start command is given from DCS, Motor starts running and gives a feedback to DCS. This 'Motor run feedback' initiates MOV opening. This process of MOV opening is having a slight lag, so there is a delay by default.

We are almost sure, there is no wrong in piping or in process..

More interestingly, piping of both pumps P-1380 and P1390 are identical in all respect except the branching lengths.

 

[LittleInch]

OK, what precisely is different between 1380 and P 1390 in piping configuration ( lengths, elbows etc) can you draw or sketch a simple iso with dimensions?

I'm wondering if the really quite large change from 4" outlet to 12" and the distance from pump to MOV has something to do with this.

My motto: Learn something new every day

Also: There's usually a good reason why everyone does it that way

 

[GPRnD]

I agree that it sounds like a surge or possibly and acoustic resonance that occurs until the check valve or MOV opens enough. In either case it is quite possible to see local pressures higher than the shutoff head of the pump.

In the case of acoustic resonance any hard boundary like a check valve or MOV will act like a wall as will changes in pipe area.

I'd be interested to see the simple sketch of the discharge piping, in particular the physical locations of the check valve, MOV, pipe increasers and the pressure switch+transmitter

 

[gerhardl]

My guess:

If the piping sketch is correct, the flow from the irregular pump will meet the flow from the other pump immidiataly (???) or at some point after the last checkvalve.

If there is a counter pressure already present here, different (higher) than the pressure before the checkvalve, and/or any air or gas present here, this will create a short 'backlash' or pressure peak, as indicated already by others.

The cure could be soft-start, if not already utilized, or re-arrangement of piping,to avoid creating extra pressure at joining of the two streams.

In some (worst) cases you can risk one pump 'dominating' so much that pump two will run at lower capacity, if the joined piping is not optimal.

 

[petrochemical87]

you could try a couple of things
1)bypass the PSHH trip for a run and observe the pressure increase/decrease trend.By the way are you completely sure the PSHH is not faulty. The line is rated for 150kg/cm2g so bypassing pshh shouldnt be a concern
2)Why dont you open the MOV before before starting the pump. If the discharege piping is full then im sure this should not be a concern.

 

[sashi30102014]

Hi All,

Attached is the isometric dwgs of both pumps.. showing suction and discharge piping. There is no difference between the pumps piping...

Please, can you provide more info on acoustic reference..From this iso's any analysis we can do..

 

[JJPellin]

In our refinery, we deal with this challenge differently. We use a primary pump at the tank as a booster. This pump is a low head, high flow vertical turbine pump. The flow from this primary pump is sent to one or more two-stage, split-case pumps that produce the higher head required.

Johnny Pellin

 

[LittleInch]

sashi,

your attachment does not load - can you try again for the isometric

My motto: Learn something new every day

Also: There's usually a good reason why everyone does it that way

 

[GPRnD]

The link below talks about standing waves.

http://phys.strath.ac.uk/12-156/12156PDF/wwol10.pdf

Or you can just google the topic in relation to standing waves excited by pumps. The second link covers the application of standing waves to pumps.

http://turbolab.tamu.edu/proc/turboproc/T5/T5pg55-62.pdf

 

[LittleInch]

sashi,

we need to see the iso - your previous attachment doesn't work, please try attaching it again.

My motto: Learn something new every day

Also: There's usually a good reason why everyone does it that way