MassFlowSource_T and MassFlowSource_h only work if there is at least one port

[HansOlsson]

The two models Modelica.Fluid.Sources.MassFlowSource_T and Modelica.Fluid.Sources.MassFlowSource_h only work if there is at least one port - since they set a mass-flow and you cannot have a mass-flow without ports.

Due to the design of connectorSizing we should clearly not change the default value, but I instead propose to add a min-value.

The reason for this change is two-fold:

• Avoid false positives when checking the library; as the default configuration shouldn't work.
• When (mis)using these components the error can be detected early.

[casella]

LGTM in principle, but we already have

ModelicaStandardLibrary/Modelica/Fluid/Sources.mo

Line 792 in 35ad28d

assert(nPorts > 0, "At least one port needs to be present (nPorts > 0), otherwise the model is singular");

Isn't that redundant?

[HansOlsson]

LGTM in principle, but we already have

ModelicaStandardLibrary/Modelica/Fluid/Sources.mo

Line 792 in 35ad28d

assert(nPorts > 0, "At least one port needs to be present (nPorts > 0), otherwise the model is singular");

Isn't that redundant?

Well, I would say there's overlap - not that they are redundant.

• The disadvantage with asserts is that it depends on when they are checked - clearly such a model cannot simulate, but exactly when it fails isn't clear. In contrast a min-value for an evaluable parameter can clearly be checked when we have a value.
• The disadvantage with a min-value in general is that it isn't clear what the problem is. However, for this specific case the combination of connectorSizing and min-value clearly indicates the issue tools can easily detect that; whereas the assert as currently written isn't very clear - it discusses nPorts, but a user should ideally not see that parameter.

[casella]

LGTM

[arunkumar-narasimhan]

Looks good to me

[Esther-Devakirubai]

@casella Can this be backported to Maint 4.1.x? Also there is no milestone mentioned.

[casella]

@Esther-Devakirubai, this was not strictly necessary, but since we are still waiting for the regression testing, please do.

[Esther-Devakirubai]

Backported to maint/4.1.x by #4469