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,
How often do you clean your system and or replace the oil?
Most probably your system is full of product of decomposition of the Hot oil.
Do you perform quality check of the hot oil (analysis)?
Do you have a filtration unit in // to clean the oil?
Do you have N2 blanketing to prevent oxidation of the oil in place?

I may be wrong.
Pierre

 

[Eric_eng]

Hello Pierre and thank you
we are in the commissioning phase and we have not heated the oil yet. we have cleaned the pump strainer several times and now the system seems to be dirt-free but still, we have the pressure drop. we have considered an N2 blanket on the expansion tank and I even put more pressure on the suction (increased to 2.5 bar) to check if it had any effect on avoiding cavitation but it didn't help.

 

[katmar]

Are you able to do a circulation test with the boiler off (or bypassed)? This would help to indicate whether it is a temperature effect or just plain hydraulics.

Edit: Whoops - I only noticed now that you have not turned the boiler on yet. So it has to be just hydraulics. If you have 1.3 bar constantly on the suction it is unlikely to be an NPSHa problem but it could be drawing some air or nitrogen down into the pump and this would give similar symptoms to cavitation. Can you supply a dimensioned drawing of the expansion tank and the piping to the pump. An undersized suction pipe with a too low level in the expansion tank would be my prime suspect.

Katmar Software - AioFlo Pipe Hydraulics

http://katmarsoftware.com

"An undefined problem has an infinite number of solutions"

 

[pierreick]

Hi,
What do you mean "not heat the oil yet"? What is he temperature/viscosity of the oil?
Did you check with hot oil supplier the procedure to start the unit from room temperature to operating temperature?

Pierre

 

[Eric_eng]

We have not run the burner yet and the circulation is at ambient temperature (around 10 to 20 C ) and the viscosity of the oil is around 50-25 mm2/s at these temperature.

 

[LittleInch]

Can you post a diagram and the pump curve to help us a bit.

Initial thoughts are that your expansion tank is running dry or that your operating of the end of the curve somehow

Do you have a flow meter somewhere?

A bypass /spill back loop going wide open?

And 5 bar drop from boiler inlet to consumer sounds very big.

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

 

[pierreick]

Hi, My experience in China, We have installed a HX to preheat the hot oil at start up.Othewise it was impossible to pump the hot oil.
Pierre

 

[Eric_eng]

Hello
I have attached the flow diagram of the circulation system and pump curve. The model is 80-250.
Unfortunately, we do not have any flow meter in the system. We have a level switch on the expansion tank to ensure that we always have a sufficient level of oil in the tank, and we have tested it. This is the reason that has left me confused. We consistently maintain oil in the pump's suction, yet after 2 minutes of circulation, the pressure drops.

https://files.engineering.com/getfi...-42be-8ac1-587e5b6fd159&file=Flow_diagram.png

https://files.engineering.com/getfi...f2-4aab-82aa-4a8255b34198&file=Pump_curve.png

 

[Eric_eng]

Also, we have removed the air in the circulation system as much as we can through valves on the top points of the pipelines and I guess that air is not causing this.

 

[katmar]

Can you mark the pipe dimensions (ID, approx length) on the flowsheet, and also indicate for each valve whether it is open or shut during the test. Also the dimensions of the separator.

Does the oil drain from the consumers between runs? What I am getting at is whether there is air/nitrogen that is displaced back to the separator at start-up or if the system remains oil-filled.

Katmar Software - AioFlo Pipe Hydraulics

http://katmarsoftware.com

"An undefined problem has an infinite number of solutions"

 

[Eric_eng]

Please check the updated flow diagram with size and approximate length. All the valves indicated in the flow diagram were open. The separators should have oil all the time as our expansion tank always has a minimum oil level.

https://files.engineering.com/getfi...-48bb-8d0c-916cdc068b49&file=Flow_diagram.png

 

[katmar]

Thanks for the extra info. It appears to me that your suction piping is just too small. If you put 100 m[sup]3[/sup]/h through a 4" Sch40 pipe the velocity would be 3.4 m/s (11.1 ft/s). This is way above what would normally be used for suction piping. And as soon as the separator has been drained and you need to draw via the 2" line from the expansion tank the velocity would increase to 12.8 m/s (42 ft/s) and the pressure drop over a 30 m length would be 12 bar.

It is surprising that the pressure gauge on the suction does not indicate a problem and your first step should be to check the gauge and its connection.

Katmar Software - AioFlo Pipe Hydraulics

http://katmarsoftware.com

"An undefined problem has an infinite number of solutions"

 

[LittleInch]

Eric,

That is not a pump curve but a selection diagram from a vendors catalogue.

Please post your actual pump curve for your pump

What does that separator separate?
Does the top connection really go into the bottom of the expansion tank connected to that pipe?
Have you tried turning the valve off ( blue circle) that connects back into the pump suction? Something strange may be going on there.
Why is that valve even there?

Also what is your planned range of viscosities when the boiler does start working. Normally if you've got a large range of viscosities, you need some sort of speed or flow control to be able to pump it properly or a pre heater like Pierre says to keep the oil low enough viscosity to pump.

At cold I would expect you flow to be about 10% of what you need in operation...
Agree with my fellow posters that you have too high a velocity.

You might be finding that the oil is steadily shearing at a faster rate and acting like a non newtonian fluid. The viscosity can vary at the same temperature with different rates of shear. It's all a bit weird tbh, but does happen.

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

 

[Eric_eng]

Hello LittleInch and katmar and thank you for your replies.
Unfortunately, We don't have anything more than that curve. we have bought the oil boiler and the pump as a whole package and these are the only data that we have.
Dear katmar, the pressure gauge on the suction line was okay, we also checked the oil level during the process and it remains almost constant in the expansion tank. the pressure change in suction was very small and it almost always stayed at 1.3 bar.
Those valves on the suction line from the expansion tank were designed there by the boiler manufacturer for air bleeding during the oil charging process ( the valve between should be closed and the oil comes from one line and the air pushed to the expansion tank until line is full). and I think after charging the oil it has no other function. they should be open during the circulation process.

Today, I could maintain the circulation and pressure by limiting the flow rate before the consumer by opening the flow control valve at about 20%. we had 10 bar at the pump outlet, 5 bar at the consumer outlet, and 1.3 bar at the pump suction, and the circulation remained okay but when we opened the valve more than 50% or added the second consumer with 20% of the flow, the pressure will start to fall again.

Do you think that we may have a very high flow rate for our pipeline? or we may have an undersized pump?

 

[pierreick]

Hi,
Not knowing your thermal fluid, I share with you and others the TDS of a very standard fluid. In operation the viscosity of the fluid is 20 to 30 times less than at room temperature. But with the oxidation of the product over the time it will be a different story.
Why don't you start the boiler and operate the system with hot oil?
Breizh

 

[pierreick]

Hi,
I've attached a handy calculator to support your work.
Good luck.
Pierre

 

[Eric_eng]

Thank you Pierre
We are using the analog of this thermal oil that you have sent. As I checked, the viscosity ranges are almost the same. at 40 C, it is 20 mm2/s. We will check again at a higher temperature.
Earlier Today, we could maintain the circulation by limiting the flow rate but the system is very sensitive to flow range and valve openings and seems to be unstable during time.

 

[katmar]

I do not think that you have an undersized pump. With what we have seen so far it is a suction piping throttling problem. In any pump circuit it would be unusual for the pump delivery, the circulation header and the suction piping all to be the same size. I would suggest that you need a minimum of 6" and possibly even 8" to cope with cold start-up conditions for the suction.

You should try to get hold of the actual pump curve. You can get the manufacturer's name from the plate on the pump and most likely you will find the curve online. This will give you curve for the power draw vs flow rate and then you can monitor the current actually going to the pump motor to see what is happening as you open the delivery valve. If the delivery pressure drops because the pump is running off the curve on the right hand side (i.e. too high flow rate) then the current will be very high, but what you will most likely find is that the pressure has dropped because the flow through the system is low and in this case you will see a much lower current to the motor.

Katmar Software - AioFlo Pipe Hydraulics

http://katmarsoftware.com

"An undefined problem has an infinite number of solutions"

 

[LittleInch]

Eric,

Then you need to ask them for a proper data book where this information should be provided. /what you have is next to useless I'm afraid.

Without the design data there's very little anyone can do to see where the error lies.

Do you have the data sheet you sent to the supplier or his back?

So e.g. where did 100m3/hr come from?
How are you supposed to control this system? Temperature or flow?
There's something not right with your flow data for each consumer and the pump appears at the moment to be oversized. If 100m3/hr is >3m/sec in a 4" line then I don't understand why the pump needs to be so big.
How many consumers is it supposed to feed?
What's the flow to each consumer supposed to be?
I think based on the control valve information that you're going off the end of the pump curve by a large amount as the pressure drop even with high viscosity oil is not enough to restrict flow through the pump.


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