Pressure Drop in Hot Oil Circulation System 1

[Eric_eng]

Hello everyone,

I'm facing an issue with a hot oil circulation system used to heat the jackets of some mixers. The system operates as a closed circulation loop with a centrifugal pump. Initially, the pressure at the boiler's inlet is 10.5 bar, which aligns with both the system curve from simulations and the pump curve (100 m3/hr). However, after approximately 1 minute, the pressure begins to drop from 10 to around 3 bars in 2 minutes, accompanied by noticeable noise in the pump case, probably indicative of cavitation. The system has an expansion tank and we always have the NPSHr (around 1.3 bar) in the suction.
We're using flow control valves for each consumer, and currently, only one consumer is in operation. when we introduce the second consumer, the pressure drop accelerates further. The pressure at the consumer outlet is 5.5 bar at the beginning of the circulation.
Any suggestions or advice would be greatly appreciated. Please let me know if you need any further information.
Thank you

 

[pierreick]

Hi,
To me the separator is there in case of water ingress in the system (after cleaning the system) meaning the vapor line should be connected to the sky of the expansion tank to release the water through the safety valve. In normal operation the valve on top of the separator should be closed and the separator operates like the continuity of the header to supply oil to the suction of the pump.
Note: I remember using a portable flow meter (non-intrusive) to monitor the flow rate at start-up of the hot oil system to ensure the pump was operating on its curve, at design condition.
note: We don't know the characteristic of the valve, meaning how can you assess the impact of X% increasing on the flow rate? again you need a measurement of the flow rate.

Pierre

 

[Eric_eng]

Thank you all for your advice
Unfortunately, all I have is that catalog curve and my estimation of the flow rate was based on that curve ( at 100m head, flow is 100 m3/hr). I will try to obtain the proper curve from the manufacturer or boiler vendor.
Regarding the pump sizing, we had to adapt ourselves to the pump introduced with the package and perhaps there is a mismatch between pipeline sizing and pump (as you mentioned the suction pipe should be bigger).
Our system is based on the flow control valves to control the temperature of our consumers and we expect a total flow rate of 70 m3/hr totally for 7 identical consumers in case all of them are in the circuit.
According to today's observation ( maintaining the circulation by limiting the flow rate), do you suggest making a bypass line between discharge and suction to control the flow rate of the pump to the consumers? because we have limitations in changing the pipeline sizes.

 

[LittleInch]

Eric,

For me there's too many disconnects here.

If your total flow is only 70m3/hr, why have you got a boiler package apparently copiable of double that flow?

what's your min / max flow range?

I hate to think how your boiler is going to cope with this apparent low flow of 10m3/hr
Also if your single consumer you're using for a test is supposed to be flowing at 10m3/hr, I hate to think what it's doing at >100...

25 to 50cP is starting to get a bit gloopy so you would expect a high outlet pressure and very low flow and maybe that's what is happening that the unit is going too slow, but doesn't explain the sudden fall in boiler pump pressure.

have you checked the temperature of the oil during this period?

You must have something more on the boiler package surely? - I can't believe anyone would buy something like that and not have at least a basic level of data sheet or technical information.

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

 

[georgeverghese]

What is this mysterious "separator" on the pump suction line - what does it do? And why isnt the top line out from this separator connected to the TOP of the expansion drum ? This separator should be as tall as the top of the expansion drum?
When you say oil boiler, you mean hot oil heater?
Can we see a detailed PID of this expansion drum that shows the installation of this level switch? Do you have any other level device to confirm what the "level switch " is telling you ? When you say level switch, you mean level transmitter? What else can you tell us about this level switch ?
Your PFD does not show a pump min flow recycle which should go from oil heater exit back to pump suction. This would also function as a heater min flow line. Without it, oil will coke up in the heater at low consumer demand. What is the stated max possible turndown on the oil heater duty - is it 5:1 or better ?
Describe these consumers - what pressure do these consumers operate at? If you had a hot oil tube leak at any of these consumers, can the cold side of the consumer leak into the hot oil and vaporise ?

 

[Eric_eng]

Dear LittleInch
Hello again
The total flow for current consumers is considered 70 m3/hr but we will have more consumers in the future and I have to use this system till then. but we always have one consumer running and the others will be in line according to the production schedule.
Unfortunately, I don't have any useful information on the boiler side for the minimum/maximum allowable flow rate, and trying to obtain this data.
My problem is that even without any boiler in the circuit and normal ambient temperature, I am unable to maintain the circulation with fully open valves, and the pressure drops.
the temperature of the oil rises approximately 10 C if I use a minimum flow for circulation (approximately 20% open valve for the consumer) and the pressures are maintained as I reported.
Dear georgeverghese
that separator is there to separate air from the oil and guide it to the expansion tank but according to the boiler manufacturer's plan, we did it however your suggestion makes more sense.
And yes I meant thermal oil heater by boiler )
regarding the level switch, it is only a floating switch that will run the feed pump automatically when it reaches the minimum level. We do not have any transmitters there and only one level switch to protect against an empty expansion tank. but we have checked the function and also the oil level in the expansion tank, and all lines are full.
the consumers are some jacketed mixers and we can have a max of 8 bar as long as we have the circulation.
Regarding the minimum flow line, I am trying to make one between discharge and suction of the pump to control the flow rate into the headers but I don't know if it will work or not. do you have any advice on that?

 

[katmar]

Note that pumps do not create pressure. They impart energy to the liquid and the pressure comes from the friction of the fluid flowing through the system. Although I doubt that you are running off the right hand side of the curve you should verify that. The easiest way to do this is to read the current (amps) being drawn by the pump motor. If the pump is too big and the friction through the delivery piping is too small then you will run off the curve and you will see that from the actual amps being close to the maximum. You don't even need the pump curve to do this check. The motor nameplate will give the maximum current.

At the moment you do not know whether the problem is on the delivery or suction side of the pump. This quick test will give you a good idea very easily and then you can start homing in on the real location of the problem.

Katmar Software - AioFlo Pipe Hydraulics

http://katmarsoftware.com

"An undefined problem has an infinite number of solutions"

 

[pierreick]

Hi,
People are willing to help! the best way is to share all information requested.
Please tell us about your role in this project. Are you plant representative or contractor representative?
Why no vendor assisting you for start-up? You mentioned earlier this is a package (boiler & pump).

A few questions.
About the pump, who is the manufacturer? share with us the information written on the name plate, may be someone can get access to pump curve.
About the separator, you said to separate air from oil? air from where? who sized this separator, 200 l drum for 100 m3/h flowrate? No sense.
Share with us the PID of this project,
why no flow meter on Hot oil (total or individual for users)?
Some resource from Internet:

https://www.sigmathermal.com/resource-library/#brochures

Pierre

 

[georgeverghese]

Min flow line should be located on discharge of oil heater, so that it acts as min flow protection for both pump and heater. The exit from min flow recycle should be to hot oil return header.
Mixer side pressure is higher than hot oil at mixer ? Have you checked for pressure integrity of hot oil tubing within each mixer, if the mixer product can leak into thin wall hot oil tubes through tube ruptures?

Pump suction strainer blocked ?

That top line from the separator should be connected to the top of the expansion drum.

We dont use these float switches these days - use a dp cell which is remote sealed on its impulse lines.

Try operating with the top exit line from this " separator" closed. You may be somehow pulling in gas from the expansion drum into the separator.

 

[Eric_eng]

Dear Katmar,
Thank you for your advice. I have measured the current. It is 72 A (at 10.5 bar) and the maximum current of the motor in a running state stated as 100 A. When the pressure drops from 10.5 bar, the current goes higher to around 85 A (at 6 bar)

Dear Pierr,
Unfortunately, the vendor of the boiler said that we have not faced such kind of problem before.
The pump is a Turkish brand and I have attached the nameplate. As you see there is not a lot of information even on the nameplate.

https://files.engineering.com/getfi...8b5-4a0b-8616-1526af48836a&file=nameplate.png

The separator came along with the package and it has been installed according to the flow diagram.
The Project documents do not have any useful data about the boiler section and all I have is the PFD that I have sent here.

Dear georgeverghese,
Thanks.
we will try to make the bypass line from the boiler outlet to the return header.
We are checking the strainer regularly and it has been cleaned several times. I will try to run the pump with the exit line from the separator closed but we have a sufficient level of oil in the expansion tank and I guess that all the lines are full and it won't draw air from those lines.
Right now, I can only keep the pressure and circulation with one consumer in the circuit and limit the flow rate manually through a globe valve on the consumer side. outlet pressure of 10.5 bar, pump suction pressure of 1.2 bar, and pressure on the consumer outlet is 5.5 bar. we have raised the temperature to 150 C, in order to keep the pressure, we closed the valve a little more but still when we open the valve 100% or add other consumers the pressure will drop to 4 bar and the noise can be heard in the pump case.

 

[katmar]

I have attached a typical pump curve for an 80-250 centrifugal pump with a 2 pole 50 Hz motor. Depending on your local voltage and frequency you may have to make slight adjustments. But here you can see that if the delivery pressure drops significantly and the current draw increases (i.e. using more power) it can only mean that you are running off on the right hand side of the pump curve. You will also see from the bottom curve that the NPSH requirement increases rapidly as the flow rate increases.

So, as much as it looks to me that the suction piping is undersized, it now looks like you are just trying to pump too much oil. I would say your next step would be to try to find someone with an external (non-invasive) clamp on flow meter that you can rent or borrow. This would confirm whether you are just pumping too much for the pump.

[URL unfurl="true"]https://res.cloudinary.com/engineering-com/image/upload/v1711705559/tips/80-250_pump_curve_kute6y.pdf[/url]

Katmar Software - AioFlo Pipe Hydraulics

http://katmarsoftware.com

"An undefined problem has an infinite number of solutions"

 

[pierreick]

Hi,
A google search:

http://www.normpompa.com.tr/en/pumps/nky

It should be possible to get the pump curve from the manufacturer in Turkey.
Edit: Another search from Turkish manufacturer

https://saleindustry.com/eng/indust...os-dlya-garyachogo-masla-btkf-k-80-250-3000-/

Pump curves and details on the left side of the document. Similar to Katmar's document
Pierre

 

[LittleInch]

I'm re reading this as it still doesn't make a lot of sense, but a couple of things come out which haven't been explored.

1) your "boiler " or heater. what type of heater matrix is this?
Does the inlet go in at the top as shown or was that just a schematic
2) you say the equipment is a "jacket" Again what type of jacket? Is the inlet at the bottom or the top and where is the exit.
3) You say inlet pressure to boiler is 10.5 bar and pressure at the consumer Outlet is 5.5 barg? Do you have any other pressure gauges.
This looks to m then like you have somewhere between 3 and 4 bar differential pressure for about 100m of 4" pipe, though the 15 lengths of 2" pipe might generate a lot of friction equal to the 4" length. Anyway this sort of pressure drop for that length of pipe is very high.
4) I'm failing to understand how you managed to buy a boiler/heater package at no doubt considerable expense and yet seem to have no idea what exactly you ordered or indeed what you bought.?? There surely must be something more than this (even if it's just written on the order!), though I agree that nameplate on the motor is scandalous in its lack of information that simply wouldn't be allowed in most countries.
5) Just because the pressure is approx. 100m differential doesn't mean you're doing 100m3/hr. The curve is pretty flat and as you don't know what size impellor is in there you could be doing anywhere from 20 to 200 m3/hr and the pressure would be pretty much the same.
6) The key missing data is any sort of flow measurement. I'm not sure why when you add another consumer you don't simply close the first control valve some more to reduce overall flow or at least keep it the same.
7) Without more data you won't fix this and it certainly looks like something is very badly mis matched or there is still a load of air in the system or it keeps getting introduced somehow. Is there a more comprehensive P&ID?

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

 

[katmar]

The question becomes - what could cause the resistance of the system to decrease during the first 2 minutes of operation and result in the pressure dropping by 6 or 7 bar.

It is NOT a simple situation of displacing air in horizontal pipes. Pushing 1m3 of air through a 4" line results in much less pressure drop than 1m3 of oil would. If anything, I would expect the overall pressure drop to increase as the lines fill with oil.

One situation where air could increase the pressure drop is if you are getting 2 phase flow somewhere and this is causing the oil to accelerate. It is hard to predict this without a detailed GA of the plant. The consumer mixers are shown with the oil entering at the bottom of the jacket and exiting at the top. This would be good in terms of air removal but if the actual flow is from top to bottom it could cause a problem. It is hard to displace air downwards and as the oil falls through the oil you could generate a froth which could certainly result in 2 phase flow and high pressure drops until all the air is removed. Please confirm that the flow in the jacket is upwards.

You mentioned that the pressure at the consumer outlet is 5.5 bar at start-up. What happens to this pressure when the pump outlet pressure falls to 3 bar?

Another possible reason for the pressure to fall is if you get some pressure recovery in the return lines as they fill up. But to explain the large drop in pressure over the first 2 minutes it would require the consumers to be mounted +40m high. Can you list the relative elevations of all the pieces of kit.

Do you have Cv values for the globe valves controlling the flow to the consumers. Or even just sizes and hopefully figure numbers.

Katmar Software - AioFlo Pipe Hydraulics

http://katmarsoftware.com

"An undefined problem has an infinite number of solutions"

 

[georgeverghese]

"I will try to run the pump with the exit line from the separator closed but we have a sufficient level of oil in the expansion tank and I guess that all the lines are full and it won't draw air from those lines."

The PFD shows the 2nd liquid exit leg from the expansion drum has an internal standpipe in the drum. If the normal liquid level maintained by this crude float switch is below say,50-80mm above the top of this standpipe, then it could be drawing in gas from the drum into pump suction. So close both the top exit line from this separator AND the valve on this 2nd liquid leg on the drum ( valve circled blue in this posted PFD ).

 

[LittleInch]

The other missing data is temperature.

Is there a cold slug of oil in the jacket or one of those lengths of pipe which then gets mixed with a larger volume and reduces the viscosity?

What is this oil?

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

 

[pierreick]

Hi,
Probably good to read:

https://www.gces-inc.com/wp-content/uploads/2018/04/14-hot-oil-systems-v3.pdf

This may validate the 2 phase- flow.
In the document attached no separator. statement about cavitation

https://www.therminol.com/sites/therminol/files/documents/Thermal_Expansion_Tank_Design_and_Operation.pdf

the link above is explaining the position of the valves for degassing. Please review your PFD accordingly.
Note: There is also a comment about the pipe size . My understanding is the 2" pipes should be 4", similar to the return header pipe.
Pierre

 

[Eric_eng]

Dear LittleInch,
Thanks for your comments.
The heater has a 3-pass horizontal configuration.
The oil enters the jacket from the bottom and exits from the top.
We have pressure at the pump inlet, pump outlet, heater outlet, and consumer outlet.
As you said, unfortunately, we do not have any flow measure in the pipeline and as Katmar suggests, I am trying to find a portable flow meter to measure the real flow.
I am trying to maintain the pressure and flow by adjusting the flow control valves of each consumer. Till now I could maintain the pressure by adjusting the valve opening for two consumers.
Currently, we do not have the situation of mixing cold and hot in the system. The system is running at the ambient temperature of around 20 C now.

Dear Katmar,
Thank you for the useful diagram and data.
The pressure at the consumer side is around 1.7 bar when the pressure drops at the pump outlet ( 1.5 bar is approximately our static pressure in the system due to expansion tank height)
The header height is approximately 9m and the oil comes down to the consumer jacket inlet (1.5 m) and goes up again from the jacket outlet (4 m) to the return header at 9 m. the elevation of the expansion tank is 13m.
The Cv of control valves is 50.

Dear georgeverghese,
The uprise of the first outlet of the expansion tank is around 15 cm and we have a level of 70 cm in the expansion tank. I thought that it could cause no problem but I will check again the system with that valve closed. Thanks for the advice.

Dear Pierre,
Thank you for your recommendations. I will check the resources.
I am not sure about the 2 phase flow. We regularly check the air bleeding valves on top of the headers separately and try to push the oil from the expansion tank on them. after several times, we have not seen air coming out unless we have not drained enough oil! Also, I do not know, if we have air in the system, can it go to the top points by itself or it stick somewhere in the pipeline and then go to the pump?

 

[katmar]

HI Eric - I think that with this latest info we are zeroing in on the location of the problem. But even if we can work out where the problem is I am at a loss to work out what it could be. The elevations you have and the fact that the flow through the mixer jackets is upwards eliminates the 2 phase flow suspicions I had.

At start up the pressure after the consumer was 5.5 bar and after about 2 minutes it dropped to 1.7 bar. This already sounds dramatic enough, but when we consider that there is 1.5 bar of static head we see that the friction head drops from 4.0 bar (=5.5-1.5) to 0.2 bar (=1.7-1.5).

There has to be something very wrong/strange in the piping between the consumer and the separator that causes the friction loss to decrease by 95% despite an increase in the flow rate. This will require very careful inspection on site and there's not much we can help with while sitting behind our computers.

If there is nothing changing in the friction after the consumer then the drop in pressure can only be caused by a decrease in flow. Is there perhaps an over-pressure protection valve somewhere (that is not shown on the flow diagram) that could be opening slowly and reducing the flow though the circuit while increasing the flow through the pump (and resulting in the increased power draw)?

Katmar Software - AioFlo Pipe Hydraulics

http://katmarsoftware.com

"An undefined problem has an infinite number of solutions"