New, staged pump records

[Mathadon]

I have some records for the Wilo TOP-S series (http://productfinder.wilo.com/en/COM/productrange/000000090002c94500020023/fc_range_description?lang=en&iwp%5B%5D=powertree-R-0).

These are staged pumps and it seems we have not foreseen a proper way of handling this in Annex60.Fluid.Movers.Data.SpeedControlled_Nrpm. In my opinion it should contain:

  parameter Real[:] speeds(each final unit="1/min") = {0} 
    "Vector of speed set points when using stages";

since it's a property of the pump.

I propose to add this so I can add the records for these pumps correctly.

I'm not sure if we should also add

  parameter Real[:] normalized_speeds(each final unit="1") = {0} 
    "Vector of normalized speed set points when using stages"
    annotation(Dialog(enable=inputType == Annex60.Fluid.Types.InputType.Stages));

to Annex60.Fluid.Movers.Data.SpeedControlled_y.

I'll already make a pull request without this.

[mwetter]

I have the following comments in addition to the comments in the pull request (https://github.com/iea-annex60/modelica-annex60/pull/397/files)

• In SpeedControlled_Nrpm, the parameter N_nominal and speeds (which should be called N) should be assigned from the record per as they are a property of the pump.
• The same should be done for SpeedControlled_y with normalized_speed and normalized_speeds, which need to be added to the record definition and given default values. Should we rather call this n_nominal and n to indicate that it is dimensionless? normalized_speed does not indicate anything about being for the nominal conditions and hence it looks like something different from N_nominal. Since Data.SpeedControlled_Nrpm extends Data.SpeedControlled_y, it needs to assign n with the final keyword so that these entries are consistent.

[Mathadon]

Regarding your second suggestion: if we add parameter n_nominal then I think we should change the implementation of speedControlled_y since it currently assumes that the speed/efficiency/power curve is defined for n_nominal=1. The user may/should interpret the value of n_nominal as the normalised speed for which he/she provides the pump curves and not changing it may lead to wrong model use. Do you agree we should add this too?

[mwetter]

I suggest to

• make a list in the user guide that shows the different parameters such as speed and speeds. This will also help us to verify that a consistent naming is used for the four models.

Revised: this is in my view not needed as the selected parameters are clear.

[mwetter]

@Mathadon I refactored the models, in particular the choice of parameter names which are now speed or speed_rpm, and speeds and speeds_rpm. Can you please have a look at it?

I also removed {y,N,m_flow,dp}_filtered as I don't see there is any use for these outputs. One can get their values from the output signal {y,N,m_flow,dp}_actual.

For SpeedControlled_{y,Nrpm}, I moved all performance data, also for the stages, to the performance data record.
For FlowControlled_{dp,m_flow}, I left the constant head (or flow) and the values for the stages in the model as these are artificial values which don't really belong into any (product-specific) performance record.

I also update the performance data for the Wilo pumps due to the new definition of the record, e.g., the nominal speed is automatically taken from the highest stage.

I get different results for the regression tests for MoverParameters, MoverContinuous and MoverStages. After my revision 0e65a37, these regression tests now have the same mass flow rate for SpeedControlled_y and SpeedControlled_Nrpm, while they had different values before. I think the new results are correct and the old are wrong. Can you please verify. I have not yet committed the new results.

Please let me know if you agree to these changes, in which case I can make a pull request.

[Mathadon]

@mwetter I agree with

• the new parameter names,
• the removal of _filtered outputs,
• moving performance data into the records,
• leaving the artificial parameter values in the mover models.

The different regression test results are probably caused by the fact that you assign speed_rpm_nominal= speeds_rpm[size(speeds_rpm, 1)]. Since in MoverParameter you redeclare redeclare Annex60.Fluid.Movers.Data.Pumps.Wilo.Stratos25slash1to4 per(speeds_rpm={0,1000,2000}, this changes the pump speed for which the pressure() record is defined, which leads to wrong results.
My suggestion would be to require speed_rpm_nominal to be set in the performance records and to set speeds_rpm={speed_rpm_nominal} by default. I'll verify if that solves the problem and then make a pull request!

Note also that defining motorEfficiency and hydraulicEfficiency inFlowControlled may be misleading since a proper implementation of similarity laws requires the pressure curve and nominal speed to be incorporated in the efficiency calculation. But I think we established before that we accept this error.

[mwetter]

In 7f331dd I added speed_nominal and speed_rpm_nominal back to the data record, and use them in gaiSpe to normalized the input signal.
I also use them for the default values for speeds and constant_speed.

All unit tests now produce the same results as before the change.

[Mathadon]

Fixed in #397

Thanks!