Opened 6 years ago

Closed 6 years ago

#5090 closed defect (fixed)

The NF does not flatten vector equations

Reported by: Francesco Casella Owned by: Per Östlund
Priority: high Milestone: 2.0.0
Component: New Instantiation Version:
Keywords: Cc: Rüdiger Franke

Description

Please check PowerSystems.Examples.AC3ph.Generation.PowerAngle. The NF flattens the equation

  v_s = cat(1, transpose(Rot_dq)*v[1:2], {v[3]});

from here to

  generator.v_s = {transpose(generator.Rot_dq)[1, 1] * generator.v[1] + transpose(generator.Rot_dq)[1, 2] * generator.v[2], transpose(generator.Rot_dq)[2, 1] * generator.v[1] + transpose(generator.Rot_dq)[2, 2] * generator.v[2], generator.v[3]};

while the old FE flattens it to

  generator.v_s[1] = generator.Rot_dq[1,1] * generator.v[1] + generator.Rot_dq[2,1] * generator.v[2];
  generator.v_s[2] = generator.Rot_dq[1,2] * generator.v[1] + generator.Rot_dq[2,2] * generator.v[2];
  generator.v_s[3] = generator.v[3];

which is what I'd expect. I'm not sure whether this is the cause of the back-end failure, but I guess it would be good to get the old FE output anyway in this simple case, as the current output doesn't make much sense.

Change History (2)

comment:1 by Francesco Casella, 6 years ago

Cc: Rüdiger Franke added

comment:2 by Per Östlund, 6 years ago

Resolution: fixed
Status: newclosed

Fixed in 945cb9d, the transpose call is now expanded as it should instead of relying on the generic expansion that just subscripts call.

It doesn't fix the backend failure though, but the model doesn't work in the old frontend either. So it's probably a backend issue.

Note: See TracTickets for help on using tickets.