Pressure drop calibration using thermal liquid blocks (1D vs CFD)

Hello Dear,

I'm writing this to seek your help and technical advice.

I'm performing pressure drop correlation study 1D TL model with CFD data.

At negative temperature i.e. -40 1D model over estimating for both over-prediction and under prediction for a range of flow rates.

What could be a reason? What factors I can use to calibrate 1D model and their typical values?

Any technical advice for calibration will be really helpful to me.

NOTE: Properties are very well aligned with properties used in CFD.

[William Rose]

@Gaurav,

Does TL stand for transmission line?

Please provide more informaiton about what you are modeling and how temperature plays a role.

Do you have data from a computational fluid dynamics simulation, and you are trying to build a simpler 1D model that reproduces certain aspects of the CFD data, such as pressure drop between two points?

[Gaurav]

Hello @William Rose,

TL here is Simscpe Thermal liquid.

I'm modeling for oil flow through channels of eAxle. Temperature plays role in viscosity of fluid, lower the temperature higher the viscosity, larger will be the pressure drop.

Yes, I've data from CFD for complex sections of system and instead of using it as lookup table I'm building 1D model which will closely replicate CFD results mainly pressure drop as of now.

[William Rose]

@Gaurav, this is not my area of expertise, so I defer to others.

[Gaurav]

@William Rose, Perfect, No worries at all.. I'm greatful you took a look atleast.

[Jakub]

Hi Gaurav,

Can you tell something more about your calibration tools?

I guess that the Simulink Design Optimization is a right tool to help you. You may see the get started tutorial from the documentation page:

https://www.mathworks.com/help/sldo/getting-started-with-simulink-design-optimization.html

It shows the full workflow starting from Data Preprocessing (it plays the important role in calibration, if you have not-cleaned data you cannot expect to get the right parameters set).

If you get stuck in the local minimum, you may choose the global algorithm like pattern search.

First, try to go through the tutorial and, if it is possible cosider to share with us your model.

Best,

Jakub Szyman

[Gaurav]

@Jakub Very well said. Let me try this option.