Featured Articles

Check out the latest featured articles.

New Article

Product Viscosity vs. Shear

Featured File

Vertical Tank Selection

New Blog Entry

Low Flow in Pipes- posted in Ankur's blog

4

Converging Hydraulics Modelling In Aft Arrow 9

12 replies to this topic
|

#1 dickymarsa

dickymarsa

Brand New Member

• Members
• 5 posts

Posted 17 September 2023 - 03:50 AM

Hi everyone,

I'm building a model about hydraulics. The general concept of the model is:
3 Flow sources (different flowrate and pressure) from "tank". This tank junction only provide pressure and temperature and does not specify the flowrate.

These 3 flow will go through their separate piping configuration, then will meet in 1 big header. The outlet of the header will then be a 3 separate flow again (these 3 outlet flows are represented by "assigned pressure" junction).

Intent of the design is:
1. Fixed inlet flowrate
2. Fixed outlet pressure and equally divided flowrate for 3 outlet flows.

Currently, I do my Goal seek and control is like this:
1. Goals: flowrate from "tank" junction (to be specified in goals as this was not specified in the junction)
2. Variable: Pressure from "tank" junction (to be the variables as what I needed is the fixed outlet pressure)

The problem is: I cannot get a converged model and the result is always messy (e.g. negative flowrate, different flowrate, etc)

Thank you

#2 breizh

breizh

Gold Member

• 6,075 posts

Posted 17 September 2023 - 06:18 AM

Hi,

Why don't you use excel and build your model.

Solver is the appropriate tool for this type of problem.

I did similar work with many branches.

Get the set of equations right, use the constrains to eliminate wrong solutions.

BTW a sketch will help to better understand your query. What is the driving force?

Breizh

#3 Pilesar

Pilesar

Gold Member

• Members
• 1,251 posts

Posted 17 September 2023 - 07:26 AM

You fixed the sink pressures and therefore the flows to the sinks will be different. The junction is the source pressure for each of the outlet branches. Flow follows pressure drop. The different branches with different piping configurations will not have identical flows as well as pressure drops. Arrow will not force the gas to behave like you want since it tries to calculate unknowns according to the laws of physics. Perhaps add some valves in the branches to adjust the pressure drop in the branches to achieve equal flow.

#4 dickymarsa

dickymarsa

Brand New Member

• Members
• 5 posts

Posted 17 September 2023 - 08:29 AM

Hi,

Why don't you use excel and build your model.

Solver is the appropriate tool for this type of problem.

I did similar work with many branches.

Get the set of equations right, use the constrains to eliminate wrong solutions.

BTW a sketch will help to better understand your query. What is the driving force?

Breizh

Hi Breizh,

Thanks for the response, I was trying to use a better tool to solve the system because it has some branches as well. I've attached the sketch.

Intent of the design is:
1. Fixed inlet flowrate (pressure need to be calculated)
2. Fixed outlet pressure and equally divided flowrate for 3 outlet flows.

Currently, I do my Goal seek and control is like this:
1. Goals: flowrate from "tank" junction (to be specified in goals as this was not specified in the junction)
2. Variable: Pressure from "tank" junction (to be the variables as what I needed is the fixed outlet pressure)

This is my first time using AFT arrow as well, so I'm still trying to figure things out

#5 dickymarsa

dickymarsa

Brand New Member

• Members
• 5 posts

Posted 17 September 2023 - 08:32 AM

You fixed the sink pressures and therefore the flows to the sinks will be different. The junction is the source pressure for each of the outlet branches. Flow follows pressure drop. The different branches with different piping configurations will not have identical flows as well as pressure drops. Arrow will not force the gas to behave like you want since it tries to calculate unknowns according to the laws of physics. Perhaps add some valves in the branches to adjust the pressure drop in the branches to achieve equal flow.

Hi Pilesar, thanks for the response

Yes, I used valves at outlet to get the equal flowrate. I needed to calculate the required inlet pressure that can get the flow to the outlet junction at the specified pressure. I've attached the sketch of how I currently do it.

Intent of the design is:
1. Fixed inlet flowrate (pressure need to be calculated)
2. Fixed outlet pressure and equally divided flowrate for 3 outlet flows.

Currently, I do my Goal seek and control is like this:
1. Goals: flowrate from "tank" junction (to be specified in goals as this was not specified in the junction)
2. Variable: Pressure from "tank" junction (to be the variables as what I needed is the fixed outlet pressure)

This is my first time using AFT arrow as well, so I'm still trying to figure things out

#6 Pilesar

Pilesar

Gold Member

• Members
• 1,251 posts

Posted 17 September 2023 - 01:23 PM

Arrow will not work the way you describe. I think you are making this harder than it needs to be. Just set the sink pressures and adjust the valves to give the flow required. As long as the source pressure is high enough, the problem should converge easily. As you change source pressure, the valve openings will need to change also. Think of this as 'how would I control the parameters in the field to get specific flow rates in downstream branches.' You use a flow control loop!

#7 dickymarsa

dickymarsa

Brand New Member

• Members
• 5 posts

Posted 17 September 2023 - 07:46 PM

Arrow will not work the way you describe. I think you are making this harder than it needs to be. Just set the sink pressures and adjust the valves to give the flow required. As long as the source pressure is high enough, the problem should converge easily. As you change source pressure, the valve openings will need to change also. Think of this as 'how would I control the parameters in the field to get specific flow rates in downstream branches.' You use a flow control loop!

Hi Pilesar,

Thanks for the response,

but how do I control the flow from the source pressure? I put flow control valves at the source as well but it doesnt seem to work as the result is not how I wanted it to be

Second, I actually dont need the control valve at the end of the system as it increases pressure drop, but I just put it there in the model to make sure I have an equal flow

Is there anything specific that I need to change in my model? maybe my junction or my GSC variable and goals?

Thank you!

#8 Pilesar

Pilesar

Gold Member

• Members
• 1,251 posts

Posted 17 September 2023 - 09:17 PM

"Intent of the design is: 1. Fixed inlet flowrate (pressure need to be calculated) 2. Fixed outlet pressure and equally divided flowrate for 3 outlet flows."

Your first intent is too much. You cannot control every parameter that affects flow and pressure in the model. There must be enough 'unknowns' to allow the software to solve the system equations. Your sketch did not come through so I am guessing you have one source, one junction, three sinks, with pipe connecting all. Once you control the flow to the three sinks, you cannot then control the flow from the source independently. The software will add the three sink flows and calculate the flow from the source. I don't see the need for 'goal seek'. As long as the source pressure is high enough, the problem seems done. Your solution would then have fixed sink pressures and equal flows. If you change your source pressure, the valve positions must change to maintain the flowrate set points. There are still infeasible solutions caused by asking for more flow than the system can handle or allowing too little pressure drop.

When you build hydraulic models, lay out the model in stages. You can first put the source, junction and one sink into a system. Converge that model. Review the results to see if they match your expectations. Then add another branch and converge the new model. Then add the third branch. This way, you can see where problems arise. It is difficult to troubleshoot non-converged simulation models. It is so much easier to troubleshoot converged models that have unexpected answers. Get your simulation to converge by simplifying it. Then add complexity one at a time to see where the model breaks.

#9 shvet1

shvet1

Gold Member

• Members
• 209 posts

Posted 17 September 2023 - 11:31 PM

I'm building a model about hydraulics.

Why? What reason does this world need one more hydraulic calculator? What is wrong with other 100? Why programming should be discussed at chemical processing forum? I is a student term paper?

#10 breizh

breizh

Gold Member

• 6,075 posts

Posted 17 September 2023 - 11:47 PM

Hi,

No sketch!

To Schvet1,

He is not building a calculator, just using AFTArrow 9. Nothing wrong with that.

Breizh

#11 shvet1

shvet1

Gold Member

• Members
• 209 posts

Posted 18 September 2023 - 12:32 AM

Thanks, sorry

#12 dickymarsa

dickymarsa

Brand New Member

• Members
• 5 posts

Posted 18 September 2023 - 11:29 PM

Hi,

No sketch!

To Schvet1,

He is not building a calculator, just using AFTArrow 9. Nothing wrong with that.

Breizh

"Intent of the design is: 1. Fixed inlet flowrate (pressure need to be calculated) 2. Fixed outlet pressure and equally divided flowrate for 3 outlet flows."

Your first intent is too much. You cannot control every parameter that affects flow and pressure in the model. There must be enough 'unknowns' to allow the software to solve the system equations. Your sketch did not come through so I am guessing you have one source, one junction, three sinks, with pipe connecting all. Once you control the flow to the three sinks, you cannot then control the flow from the source independently. The software will add the three sink flows and calculate the flow from the source. I don't see the need for 'goal seek'. As long as the source pressure is high enough, the problem seems done. Your solution would then have fixed sink pressures and equal flows. If you change your source pressure, the valve positions must change to maintain the flowrate set points. There are still infeasible solutions caused by asking for more flow than the system can handle or allowing too little pressure drop.

When you build hydraulic models, lay out the model in stages. You can first put the source, junction and one sink into a system. Converge that model. Review the results to see if they match your expectations. Then add another branch and converge the new model. Then add the third branch. This way, you can see where problems arise. It is difficult to troubleshoot non-converged simulation models. It is so much easier to troubleshoot converged models that have unexpected answers. Get your simulation to converge by simplifying it. Then add complexity one at a time to see where the model breaks.

Hi Pilesar and breizh,

Thanks a lot for getting back to me again, sorry I forgot to click the "attach" button previously.

FYI, I've done the simulation and get the converged solution with GSC, but just in case you have a better and more efficient model.

I still use GSC to reach the solution, currently, I do my Goal seek and control is like this (successful):
1. Goals: flowrate from 2 of the  "tank" junctions (to be specified in goals as this was not specified in the junction) and pressure upstream of the first control valve "assigned pressure" < this line has the biggest pressure drop, so I use it as the benchmark.
2. Variable: Pressure from "tank" junction (to be the variables as what I needed is the fixed outlet pressure)

Intent of the design is:
1. Fixed inlet flowrate (but this is not specified in the "tank" junction yet)

2. Outlet pressure to meet the minimum required pressure and equally divided flowrate for 3 outlet flows.

Anyway, thank you so much for your help so far

Cheers,

#13 breizh

breizh

Gold Member

• 6,075 posts

Posted 19 September 2023 - 01:14 AM

Hi,

Good to know you was able to solve your problem.

I've downloaded a simple AFT free software (PSIM2) where the number of pipes is limited to 12.

Without data it's difficult to perform calculation....

Best of luck

Breizh