Issue1575 add pv model

[LauraMaier]

This closes #1575

We added a new PV model to the IBPSA library.
The model has the following characteristics:

• The model is based on two different approaches for the I-V- characteristic: a single- and two-diode approach.
• We included three modeling approaches to calculate the cell temperature depending on the PV mounting: open rack, close-to-ground, and in-contact-with-ground
• The model was validated using measurement data from a rooftop PV system in Berlin, Germany. Further information can be found here: http://www.solar-rooftop.de

[LauraMaier]

@mwetter: I do not understand why the CI fails for IBPSA.Electrical.DC.Sources.Validation.PVSingleDiodeRooftopBuildingValidation since the "Simulate and Plot" script works fine on my desktop.

[JayHuLBL]

@LauraMaier Running the model IBPSA.Electrical.DC.Sources.Examples.PVSingleDiode with Dymola (2025x) GUI, I see following error:

[mwetter]

@LauraMaier
IBPSA.Electrical.BaseClasses.PV.PVElectricalSingleDiodeMPP
lines 71 ff needs no regularization because k is a constant and TCel is always larger than zero.

In the same file, the expressions

  RSh/RSh0 = 1/max(absRadRat, Modelica.Constants.eps);

  // Simplified power correlations at MPP using Lambert W function (Batzelis et al., 2016)

  IMP = max(0,IPh*(1 - 1/w) - a*(w - 1)/max(Modelica.Constants.eps, RSh));

can be rewritten without regularization. The problem in the original formulation is that RSh is computed with guard against division by zero, and then IMP uses this in the denominator. Eliminating RSh avoid direct use without this guard. If RSh is indeed needed as an output, then we can add it back in, but for now I commented it.

I added these changes in 8421baa80 together with some other changes further down in that file, but now I am stuck: If IPh0 could be a parameter, then I can easily proceed, otherwise it gets more elaborated. Given the names and comments of IPh0, IS0, ... w0, I think the can all be parameter expressions, but I don't understand the model use sufficiently. Can you make them parameters and if so, do a commit to make the change?

Also, please correct the unit error that @JayHuLBL pointed out so the regression tests run and we can make sure that we don't introduce an error when making these changes.

[LauraMaier]

@LauraMaier Running the model IBPSA.Electrical.DC.Sources.Examples.PVSingleDiode with Dymola (2025x) GUI, I see following error:

Thanks for comment. Adressed @JayHuLBL's comment in commit aee2cfc

[LauraMaier]

@LauraMaier IBPSA.Electrical.BaseClasses.PV.PVElectricalSingleDiodeMPP lines 71 ff needs no regularization because k is a constant and TCel is always larger than zero.

In the same file, the expressions

  RSh/RSh0 = 1/max(absRadRat, Modelica.Constants.eps);

  // Simplified power correlations at MPP using Lambert W function (Batzelis et al., 2016)

  IMP = max(0,IPh*(1 - 1/w) - a*(w - 1)/max(Modelica.Constants.eps, RSh));

can be rewritten without regularization. The problem in the original formulation is that RSh is computed with guard against division by zero, and then IMP uses this in the denominator. Eliminating RSh avoid direct use without this guard. If RSh is indeed needed as an output, then we can add it back in, but for now I commented it.

I added these changes in 8421baa80 together with some other changes further down in that file, but now I am stuck: If IPh0 could be a parameter, then I can easily proceed, otherwise it gets more elaborated. Given the names and comments of IPh0, IS0, ... w0, I think the can all be parameter expressions, but I don't understand the model use sufficiently. Can you make them parameters and if so, do a commit to make the change?

Also, please correct the unit error that @JayHuLBL pointed out so the regression tests run and we can make sure that we don't introduce an error when making these changes.

Thanks for your support, @mwetter.
You are right. I redeclared the former variables "_0" to parameters in e89ebf9

[LauraMaier]

@mwetter and @JayHuLBL : Again, models look fine with Dymola 2023. I cannot update to the latest Dymola version since I am no longer employed at the university. Can you maybe share the error log since I cannot find it in the CI and cannot reproduce it locally since everything runs smoothly here. Thanks for the support!

[JayHuLBL]

@mwetter and @JayHuLBL : Again, models look fine with Dymola 2023. I cannot update to the latest Dymola version since I am no longer employed at the university. Can you maybe share the error log since I cannot find it in the CI and cannot reproduce it locally since everything runs smoothly here. Thanks for the support!

I will test it and let you know if there is any problem.

[JayHuLBL]

@LauraMaier All the models can run successfully with Dymola 2025x. I have run the unit test and pushed the reference data.

[LauraMaier]

@LauraMaier All the models can run successfully with Dymola 2025x. I have run the unit test and pushed the reference data.

@JayHuLBL : Many thanks for your valuable support. I added some minor adjustment for OM compatibility.
@mwetter : From my side, the models are ready. Please feel free to review whenever suitable. I don't know when I will be available since I am an maternity leave and the EDD is in 9 days from now ;-) but I will try to get back to it as soon as possible.

[JayHuLBL]

@LauraMaier I did some changes to modify the format and appearance of the models. Also:

• IBPSA.Electrical.BaseClasses.PV.PVThermalEmpMountContactToGround is not used by any class. Do we need it?
• IBPSA.Electrical.BaseClasses.PV.PVThermalEmpMountOpenRack is not used by any class. Do we need it?
• In IBPSA.Electrical.BaseClasses.PV.BaseClasses.PartialPVOptical, how is the input tilSet being used? Although the base class PV.BaseClasses.PartialPVOptical is used by IBPSA.Electrical.BaseClasses.PV.PVOpticalAbsRat and IBPSA.Electrical.BaseClasses.PV.PartialPVSystem, I don’t see tilSet being needed.
• Has the imageResources/Images/Electrical/AC/OnePhase/Storage/Examples/SOCs.png been used anywhere? Otherwise, it should be deleted.

[LauraMaier]

Thank you so much, @JayHuLBL !

• Regarding the unused models IBPSA.Electrical.BaseClasses.PV.PVThermalEmpMountContactToGround and IBPSA.Electrical.BaseClasses.PV.PVThermalEmpMountOpenRack: these are empirical models based on literature which are documented in the main model. Depending on the mounting situation, you have to select a different thermal model. The selection is based on the tilt and described accordingly. In my opinion, we should leave them in the model so that other users can adapt the PV model to their mounting situation.

• The situation with tilSet is a comparable one. It is an option that if users want to have an adaptable tilt (like a motorized PV cell following the sun), the models offers the opportunity. I extended the documentation to explain the use of it.

• Good catch, thanks. I deleted the image.