The motivation for stream connectors is that using potential and flow variables for convective mass transfer leads to systems of nonlinear differential equations (with boolean variables for the direction of the flow). And singularities with zero flow. These systems are unreliable to solve. One way to simplify this is to introduce two directions of the flow and introduce two separate sets of equations for flow in and flow out. That's what the stream connector helps with.
A detailed explanation can be found in these slides by M. Otter, F. Casella, M. Sielemann https://www.modelica.org/education/educational-material/lecture-material/english/ModelicaStream-Overview-Rationale.ppt/view
I tried to read about Stream connectors and couldn't really understand them. Can you tell me when is it better to use them ? What is the main difference between them and normal flow and non-flow connectors ?