1

Inputs in my multibody model are road unevennes as function of distance the model is working fine for time simulation but when I linearize the system, then i get error saying no inputs or outputs defined.As In the picture:

https://drive.google.com/file/d/0B8AmxzbhEQ_FczFuaDNhQjlqVUE/view?usp=sharing

I defined combitable as a input and parameter then I get the error: Translation of TRyViertalMbsa: Component type specifier CombiTimeTable is a block, this cannot be combined with the prefix input . Component declared as CombiTimeTable combiTimeTable in Modelica.Blocks.Sources Conflicting causality for combiTimeTable (input ) and its element y(output )

How can I define inputs and outputs for the state spac system?

Yusuf çatı
  • 21
  • 2

1 Answers1

2

I think the problem is that you are asking Dymola to linearize the model including the boundary values (the input time series). What you really want is to linearize the model only and let Dymola perturbate the inputs in order to determine the linear model. See example model:

https://drive.google.com/file/d/0B8ojPn4YxnI9a1FFa0ozY2tlT1U/view?usp=sharing

So, you must replace the two CombiTimeTable with two input connectors (Modelica.Blocks.Interfaces.RealInput). Also, you would probably want to add one or more output connectors (Modelica.Blocks.Interfaces.RealInput) via mechanical sensors - angle, position or whatever.

Operating points

If you want to linearize your model around a specific operating point different from zero, you should add these (constant) values to the inputs and subtract them from your outputs. I don't know if there is a smarter way to handle the operating points in Dymola's linearization function.

Best regards Rene Just Nielsen

Rene Just Nielsen
  • 3,023
  • 12
  • 15
  • Thank you very much! I defined outputs with the sensor applying this way but when I give combi time table as an input it still does not recognize the table as an input when I want to have the bode plot of the system. https://drive.google.com/file/d/0B8AmxzbhEQ_FUFB2c2JFemx1X0k/view?usp=sharing – Yusuf çatı Feb 18 '15 at 13:26
  • Yusuf, if you just delete the two `CombiTimeTables`, Dymola can perform the linearization. The `CombiTimeTables` contain time series of input data (boundary conditions) which are _not_ required to linearize your model. You can always re-apply them to your linearized model afterwards if you want to compare the **time responses** of the linear and nonlinear models. Also, there are a number of methods to compute the **frequency response** of the linearized system but for that purpose you won't need the `CombiTimeTables`. – Rene Just Nielsen Feb 18 '15 at 21:31