Feature request: indirect wake integration using test beams
CST, Echo, ABCI and other codes use indirect wake integration methods that are more accurate than the direct method, but are only applicable under certain circumstances.
Indirect testbeams is considered the most accurate method, but applicable only if
- beam is ultrarelativistic
- beam tube cross section and structure entrance and exit are equal
- Structure must be convex, all cross sections must be larger than the pipe cross section (no collimators, striplines etc.)
The test beams are placed on the extruded "shell" of the pipe cross section. With the wake potential known on the shell, it can be computed on any point inside the shell. This requires solving the Poisson boundary value problem because the longitudinal wake is harmonic in transverse plane: $\nabla^2_\perp W_{||}=0$.
References linked below for more details, but even those are rather vague with respect to implementation.
Napoly-Chin-Zotter integrals (indirect interfaces in CST) are applicable for arbitrary cross section but ultrarelativistic beams. This should be another issue.
References
Sketch
Illustration taken from the above reference:
Feature request: indirect wake integration using test beams
CST, Echo, ABCI and other codes use indirect wake integration methods that are more accurate than the direct method, but are only applicable under certain circumstances.
Indirect testbeams is considered the most accurate method, but applicable only if
The test beams are placed on the extruded "shell" of the pipe cross section. With the wake potential known on the shell, it can be computed on any point inside the shell. This requires solving the Poisson boundary value problem because the longitudinal wake is harmonic in transverse plane:$\nabla^2_\perp W_{||}=0$ .
References linked below for more details, but even those are rather vague with respect to implementation.
Napoly-Chin-Zotter integrals (indirect interfaces in CST) are applicable for arbitrary cross section but ultrarelativistic beams. This should be another issue.
References
Sketch
Illustration taken from the above reference:
