Ejector (2P)

How can I design a two phase flow ejector in simulink / simscape because this components does not available in simulink?

[Yifeng Tang]

Does the fluid change phase from the primary & secondary inlets to the exit? If not, there can be a simple solution. If it does, please describe the actual physical system a bit more and let's see if it indeed requires an ejector.

[Hussein]

Dear Tang,

THe process is, the fluid discharge from condenser as a subcooled saturated liquied then flow as primary fluid to motive nozzle then entrain the second liquied from lower pressure zone as a saturated vapour to expand in the secondary nozzle after that these two fluid (same refrigerant but different in state) mix together at mixing section to become mixture at vapor quality let's assume X=0.6, then this mixture goes to diffuser to incrrase pressure. Thanks

Answer 1

[Yifeng Tang]

Hi Hussein,

Sounds complicated :p This reminds me of certain design of flash tank or economizer design I saw in some multi-stage refrigeration system. Some thoughts:

(1) let's see if we can model this device using an empirical approach. For example, if you have flow rate and entrainment data available at a few different conditions, you can try to build a few lookup tables or regression models in MATLAB. You can then use ideal sources, PS lookup tables and sensors to implement the LUT or regression you've built. I have a hunch that you'll only need the flow rate correlations, because an ejector is a passive device and no total energy is added to the fluid (so we don't need to worry about adding or removing additional heat).

(2) a full physics approach will require a set of equations that describe/predict the flow rate and entrainment ratio, as a function of states (pressure and enthalpy) at each port and the geometry parameters. If such information is available, you can write your own equations in Simscape language, but there will be learning curve if you haven't done much of component customization before.

[Paulina]

Hi, looks like the original author has not continued with the topic, but I encounter similar problems. I have some experimental points, and I have a full working map achieved from high quality CFD simulations. However, I am not very fluent in coding. Can you give me a hand with that? :)