Discussion Forum

I'm looking for advice on non-local couplings to transfer the value of variables at boundaries from one component to another.

I have a fairly simple model: 1D components of adsorption beds built by a colleague (which I'd like to keep with slightest modifications) and a new 2D axisymmetric component for ducts to connect between and beyond the beds. My intent is to model gas flows through this system where the reasonable assumption that the air at the boundaries from the "ducts" to the "beds" can be a well-mixed average/integration. My problem is how to properly utilize Non-Local Couplings in my model to go from 2D axi-symmetric down to 1D and from 1D to 2D axi-symmetric. I feel as if this should be very simple, but something fundamental must be miswritten in my model, particularly with coordinates.

These methods have failed and I don't know why.

1. Using the "Integrate" non-local coupling to go from 2D-axi -> 1D produces errors. I integrate the 2D-axi "outlet" boundary and call for the variables that component contained in the 1D model at the "inlet" point. Something like "Bed_P_inlet = duct.integrate1(duct.p)" so that the 1D model has an inlet presssure boundary condition.

2. Using "General Extrusion" for the same, mapping the 2D-axi boundary down to 1D. I cannot get coordinates to behave in either direction. The boundaries are given equivalent values, so z and x are the same value.

3. Using "General Projection" for the same intent as the extrusion, mapping downward in dimensions. I expect the r-coordinate to be integrated and for x in 1D to map to z in 2D.

All of the transfered values are scalars (mass fractions, pressure, temperature, total flowrate), so I see no reason a point-to-boundary mapping should fail.