Modelica Forum > closed fluid cycle possible?

Full Version: closed fluid cycle possible?

wendelin

Oct 11 2007, 12:41 PM

Hi I am writing models for fluid systems using the Modelica Fluid Lib (http://www.modelica.org/libraries/Modelica_Fluid). In the examples there is no closed cycle so that for example the outlet of a pump passes other components (easiest example: one pipe with pressure loss) and then enters the inlet of the same pump again.
All the examples use FlowRate or FixedBounderies as Sources/sinks and so don't describe closed cycles. Is this possible in Modelica and especially with the FluidLib?

When trying the easiest example I can think of (see below) I get an error message that the model contains mean circular equalities.
Here my example, written in Dymola:

CODE

model ClosedCycle

  annotation (uses(Modelica_Fluid(version="1.0 Beta 2"), Modelica(version=
            "2.2.1")), Diagram);
  Modelica_Fluid.PressureLosses.WallFrictionAndGravity pipe(
    redeclare package Medium = Modelica.Media.Water.ConstantPropertyLiquidWater,

    length=2,
    diameter=0.02) annotation (extent=[-18,44; 2,64]);
  Modelica.Blocks.Sources.Constant const(k=1200)
    annotation (extent=[-94,28; -74,48]);
  inner Modelica_Fluid.Ambient ambient annotation (extent=[50,64; 70,84]);
  Modelica_Fluid.Pumps.Pump pumps(
    checkValve=true,
    redeclare package Medium =
        Modelica.Media.Water.ConstantPropertyLiquidWater,
    N_nom=1200,
    redeclare function flowCharacteristic =
        Modelica_Fluid.Pumps.BaseClasses.PumpCharacteristics.quadraticFlow (          q_nom={0,
            0.25,0.5}, head_nom={100,60,0}),
    Np_nom=4,
    M=50,
    T_start=Modelica.SIunits.Conversions.from_degC(20))
    annotation (extent=[-40,-18; -14,8]);
equation
  connect(pumps.outlet, pipe.port_b) annotation (points=[-19.2,-0.84; 10,-0.84;
        10,54; 2,54], style(color=69, rgbcolor={0,127,255}));
  connect(pipe.port_a, pumps.inlet) annotation (points=[-18,54; -50,54; -50,
        -7.6; -37.4,-7.6], style(color=69, rgbcolor={0,127,255}));
  connect(pumps.N_in, const.y) annotation (points=[-30.38,0.72; -30,18; -28,36;
        -34,36; -34,38; -73,38], style(color=74, rgbcolor={0,0,127}));
end ClosedCycle;

Is there a way to prevent circular equalities and still modeling a closed cycle? If not, how do you simulate pressure losses in such systems?
Thanks for your advice.

wendelin

Oct 19 2007, 08:13 AM

Nobody has had this problem before? The alternative i see is to cut the cycle. But this means that a static pressure at the "end" of the cyle has to be defined- and so its not truely dynamic modelling of a fluid cycle anymore.

Any ideas or experiences?

Hannes

Mar 13 2008, 03:35 PM

Hi Wendelin!

Currently, I obviously hit the same problem as you did five month ago. Are you still interested in this topic or have you even found a solution?

I cannot make your code running because obviously the Modelica_Fluid library you used is a different one than the one I can download from modelica.org.

As to my problem:

I also need to model pump plus pipe / two-port (representing characteristic curve of a hydraulic plant) in order to have a tool for sizing the pump, electric motor, inverter and PV-module.

My first impression was, that in a closed cycle there are only pressure DIFFERENCES (delta_p) well defined but no absolute pressure. If we were in electrics, a king of GROUND would be missing to define the electric potential. In this case a pressure potential is missing. - So far about my assumption.

I tried to define a hydraulic ground: With the standard connector-concept I implemented with a well-defined parameter p_gnd=1.013e5 (Pascal):

port.p = p_gnd;
port.dotM_flow = 0;
port.dotH_flow = port.dotM_flow * (if port.dotM_flow>=0 then
port.h else
0);
port.dotM_flow_s = port.dotM_flow * (if port.dotM_flow>=0 then
port.X_s else
0);

This is an one-port defining the absolute pressure and setting mass flow rate, partial mass flow rate and enthalpy flow rate to zero. The test-setup from "ModelicaDymola_Screenshot1.png" doesn't work, although number of equation equals number of unknowns.

In "ModelicaDymola_Screenshot2.png" I tried an elastic pressure absorber I programmed once. Same result.

Do you have further ideas?

Regards,

Hannes

wendelin

Mar 13 2008, 07:00 PM

Hi Hannes,
I am still interested in a solution, but didn't work on this problem recently. I will attach the minimal examples for an open cycle (which works) and a closed cycle. I think it's easier to discuss if we have the same mo-files. I used the 1.0-beta2 version of the fluid lib which is still available from modelica.org. So we should be using the same library.

Regards,
Wendelin

Severine

Feb 24 2009, 01:06 AM

Hi
I'm running on the same problem now too.
Did you solve your problem?
Thanks
Severine

QUOTE(wendelin @ Mar 13 2008, 07:00 PM)

ETH

Mar 18 2009, 09:28 PM

QUOTE(Severine @ Feb 24 2009, 02:06 AM)

Hi
I'm running on the same problem now too.
Did you solve your problem?
Thanks
Severine

Try the new Modelica_Fluid downlable from www.Modelica.org (requires Dymola 7.1 mini)

This is a "lo-fi" version of our main content. To view the full version with more information, formatting and images, please click here.