Pump Issues on Discharge of Vacuum Dryer

[jph90]

Hi all,

I am currently half-way through the commissioning of a new vegetable oil neutralisation plant. We are having issues with high surging on a particular progressive cavity pump on the discharge of an oil vacuum dryer. The dryer vacuum pump is sized to hold 100mbar, and evaporates entrained moisture from the oil.

The issue that we are having is that due to the vacuum activity in the vessel, the pump suction side is starved - not flooded. This is causing massive fluctuations in discharge line pressure as per the attached screenshots of a video - it cycles continuously between <0 and 3 bar, and has a loud chattering noise sounding like marbles in the pump (typically dry running noise). If you refer to the attached pump specs, the pump was designed by the supplier to suit the vessel geometry and operating pressure (supposedly). It has an NPSHr of 0.7m. We maintain 50% level in the vacuum dryer, which equates to a static height of approx 2.0m. Even though this meets the NPSHr, the pipe is likely not flooded due to the vacuum disruption and potentially causing this fluction.

My question is am i understanding the problem correctly? How could we stabilise the discharge line pressure and stop the chattering (besides increasing vessel height). I tried adding some back pressure by partially closing a downstream valve on the discharge but this did not help. Could a pulsation dampener help?

Really appreciate any guidance and do advise if i have not given enough information.

 

[jph90]

Attached are tank specs (note operating level)

 

[jph90]

Video of the pressure gauge fluctuations

 

[LittleInch]

Interesting

Can you sketch or describe the inlet piping?
Have you checked for debris or blockage in the inlet piping all the way to inlet flange of pump?
Is pump going the right way round??
Can you reduce speed to minimum then slowly increase speed?)
Have you checked speed of pump?

Is the product at the inlet the viscosity and temperature as stated ( 8.3Cp @ 90C?~)

If the product is supposed to be at 90C I don't understand why the piping is not insulated in the video.

Have you asked what the NPIP is ( PD pumps don't use NPSH.)

Have you asked what the cavitation pressure is?

What is the vapour pressure of this liquid?

Looks initially to me like some sort of fluid starvation of the inlet. If this is vegetable oil then could easily start to wax and plug. Is the inlet line and pump trace heated to prevent waxing?

Look at the process and the fluid rather than the pump for the moment.

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[jph90]

Thanks for responding. Line is clear of debris and swarf, flushed the lines after construction with meshes at pump auctions. Temperature is 90 deg C and properties are accurate. The vacuum is designed to flash water from the oil, but obviously ot drop the pressure low such as to flash oil at 90deg C. Not insulated as time did not permit in construction phase, to be addressed after commissioning.

fairly certain it is fluid starvation as you suggest, and we are getting vapour pockets/slugs into the screw pump. As we increase the Hz on the pump the flow increases but so does the chattering on the gauge and pipe work vibration. Unsure of cavitation pressure. The pump is really operating as intended I guess and we cannot practically flood the inlet without additional height of suction pipe work. it is all soft oil so no chance of clogging or hardening and no need for heat tracing (thought we have it).

Principally the large chattering/swings in pressure cause a lot of vibration and horrible sounds at the pump, looking for an effective solution to dampen this

 

[LittleInch]

What is the piping layout at the suction - size.

Can you sketch / attach relevant piping drawing?

A quick call to the vendor might help. They might have seen something similar.

I think you need to sort out the cause rather than try and deal with the symptom otherwise you will quickly wreck your pump.

Also turn on the trace heating on the inlet pipe. You risk getting a layer of hard wax / oil on the pipe and it won't do any harm - just give it a go.

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[EdStainless]

Does oil enter over a set of baffles or trays? You don't want a pool of wet oil under vacuum or you could be getting wet oil at the pump intake, and as it flashes to steam you are starving the pump inlet.

= = = = = = = = = = = = = = = = = = = =
P.E. Metallurgy, Plymouth Tube

 

[jph90]

Sorted the issue this morning - an air purge on the discharge of the vacuum had accidentally been stroked periodically when a sparky was troubleshooting an actuated valve, causing a pocket of 6bar air in between the screw pump and the vacuum. We drained and bled the line and everything is operating smoothly. The little things right

 

[LittleInch]

jph90,

Thanks for coming back and letting us know that, as in many cases, the issue was nothing to do with the pump.... I confess I didn't understand a single word of your explanation, but glad it's all working now ( was it introducing air pockets?). I'm just surprised it worked at all if you had 6 bar going into a partial vacuum...

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[Artisi]

As normal, most pumps "problems"are not pump problems but application / installation screw-ups.


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.)

 

[jph90]

Hi guys,

The issue was not actually solved as I excitedly indicated. We are still having issues. Attached is a photo of the installation.

As already been discussed in this thread, the issue is not with the pump or NPIPr/NPSHr. I will attach the calcs in a separate post. We are still - however - experiencing large cavitating noises in instances where the flow rate (pump Hz) is increased (current safe flow 5m3/hr, design for 8.5m3/hr). We seem to have partially eased the situation by decreasing the vacuum (now 100mbar abs, was 50mbar), however the issue still prevails at higher flows and low tank levels.

An interesting issue is that when the cavitating noises are heard, the low level switch on the pump suction flashes on/off continuously. This strongly indicates to me there are pockets of air being introduced. I want to do a smoke test to see if anything is drawn in around nearby flanges, unions etc. A colleague of mine is also of the belief that the sudden increase in suction pipe size from 2.5" on the vessel discharge to 4" could be the cause for pump noises and discharge pressure fluctuations, though does not explain the level switch triggering on/off continuously as it would suggest the air is introduced in this divergence.

Does either of these two options seem feasible/likely? The suppliers/commissioning party working for us (project managers) are not adequately or hastily addressing the issue so want to investigate independently

 

[jph90]

NPSHa

 

[LittleInch]

You put the vapour pressure down as 0.03m.

I think this in incorrect. Anytime you have a vacuum system like this where something is vapourising or boiling off, in essence the vapour pressure of the liquid is equal to the vacuum pressure, hence all you have for NPSH on the positive side is static head. Therefore you don't have 3.01m NPSH, you have 1.846m. That's a lot closer to 0.7m than before. Also like I said before, there is a big difference between NPSH and cavitation. Also it tends not to be a fixed number. Can you post the pump curve or test figures, not just the data sheet.

I don't know where you get 0.156m from, but as noted above, if you're not trace heating or insulating that line, your viscosity or waxing of the inlet pipe might be causing more friction loss.

I would like to see a P &ID or some sketch which tells me what all those pipes are attached to the 2" and 4" pipe work and those copper tubes.

what sort of instrument is the low level switch. It's rather odd to see one there and not on the vessel itself. I can't see how it works at present.

also how are you measuring / controlling liquid level, especially if you're increasing flow. you don't show a level transmitter, but that might be just missing from your sketch? Are you sure you're not actually getting a low level?

How is this velocity being modified? the pump data sheet shows a variance of 54 to 177 rpm. That's a big ask for a VFD....

Could still be rogue air getting in somewhere - there seems to be a lot of connections and tappings. I don't think the 2" to 4" reducer is doing anything bad - should reduce friction losses.

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[LittleInch]

Any update?

This is an intriguing one so would be good to see if you've got it to work.

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[jph90]

Hey mate,

Sorry for late reply. Still not resolved and still trying solutions. To respond to your questions:

The vapour pressure for soyabean oil at 90degC is a mere 2.15kPa. The purpose of the dryer is not to fully evaporate a solution, its to remove the moisture from the oil. The oil is sprayed into the dryer at 90degC and with 2% moisture (measured). Vapour pressure of water is 70kPa at that temperature, vacuum is 100mbar. Water flashes off, no oil flashes off. This low vapour pressure is the reason my calculation yielded a low result.

The 0.156 for vel and suction head is calculated on a base viscosity of 8.35cPs. These oils are soft oils, and this is their viscosity in ambient conditions. If we were dealing with palm, coconut oil etc then it would be a different situation, but then of course we would have the tracing on. We pulled apart the pipework though and no waxy accumulation or anything. The worst of the gummyness/waxyness of the oil (phosphatides and soaps) have been removed at this point anyway (purpose of the refinery).

Please find attached sketch of how it was installed before this weekend (we modified it). Will attach P&ID in next post. Level switch is a fork design from E&H. We tested it with water etc though and it worked fine - the problem is not with the instrument (first thing we checked). And yes we have an LT to control the level in the tank via the oil discharge pump (where the cavitation is occuring).

Since we last spoke:

We pressure tested the vessel again, as i was concerned around a few air leaks and dodgy mac unions. We found a few, and did the best we could to recitfy but once we resumed production the issue persisted. Following this, over the weekend we modified the pipework layout so the 4" reducer is directly on the pump suction - not half a meter away on the vertical. I am unsure this will help much but worth a shot. We also removed all the mac unions (worst thing ever to assemble a plant with - dont do it) and have replaced with flanges. Will give you an update in a couple of hours to see if it has been rectified

We are in the process of testing this so will let you know if it is solved.​

 

[jph90]

T09 is the vacuum dryer

P09 is the pump

 

[LittleInch]

OK,

Thanks for getting back to us - too many posts just vanish into the ether....

You know you process much better than me so I can believe your vapour pressure.

A couple of things come to mind when looking at the data.

1) Any chance you're getting slugs of water which fall to the bottom of the oil then vapourise suddenly in the inlet?
2) What is that relief valve set at? Does it account for the near vacuum on the top side of it?
3) You have quite a few actuated valves connecting to the inlet. Presume you've cycled all those and know that they're not cycling.
4) Any chance some of the water you referred to for testing is sitting in a low point somewhere and just occasionally mixing and flashing off?
5) While the pump is generally regarded as "near pulseless" it's probably worth trying to find out from the pump tech support whether there is some sort of inrush flow as the rotor turns. You might want to see if there's a drawing showing what the pump inlet actually looks like.
6) does the outlet flow pulse when this is happening?

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[LittleInch]

Any further update?

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.