Why is my Simscape gas fan simulation so incredibly sensitive to temperature and flow resistance?

I am doing a relatively simple simulation in Simscape with a fan, some pipes, a heat exchanger etc. (and a bunch of sensors) simulating helium at T=20K P=20bar.

I have robust fluid properties entered across the entire operating range that I am simulating.

If I initialize at say 20K everything runs smoothly and the temperature may fluctuate between 15K & 25K depending on the heat loads etc. HOWEVER, if I attempt to INITIALIZE at 25K or even 22K, the simulation won't converge to initial conditions!

Additionally, if I change the flow resistance of my circut slightly then the entire simulation may not initialize.

I am commanding the pump with a ramp speed input from an ideal angular velocity source (as below) usually starting at 0 speed (but this doesn't seem to change the initialization success much):

The error is always this:

Does anyone have reccomendations for how to deal with these issues?

Is there some trick to getting the fan to operate more robustly?

Is there a better solver setting I should be using (I have tried starting from steady state and not starting from steady state).

[zhuo chen]

can i ask you some questions about the “fan” block

[Colin O'Shea]

I may be able to answer some now. They actually found a bug that they said they would fix in upcoming releases (I was on 2019b).

[zhuo chen]

I have the same error when using two fans in parallel piping. Have you tried using fans in parallel piping?

[zhuo chen]

and i have another question，The default value of the fan's reference density is 1.2, but in my system, the density of the fluid will change, resulting in a change in the fan's characteristic curve. How can the fan's reference density be set to the actual density of the inlet (instead of constant value)

[Colin O'Shea]

With regards to the density, the reference density is that which you used to calculate the input vectors for fan data (flow The block will interpolate the density vector in your material properties and scale the fan vectors by this new density as tempreature changes. It is equivalent to other blocks requiring a reference temperature and scaling the calculated values by the interpolated material properties.

[zhuo chen]

Thank you very much for your answer. I want to simulate a constant temperature field. How can the density not change with the temperature through the setting, or the temperature of the system does not change, can it be achieved by changing the insight of the material?

Answer 1

[Juan Sagarduy]

Hello Colin,

I think that Starting from Steady State can be a difficult choice to start with.

I would recommend that you start with a simple system: fan, source reservoir, a pipe and a sink reservoir.

 Pressure for both reservoirs is equal and transfer of gas is done through spinning the fan.

BR Juan