19  Wakefields

View:

The wakefield chapter below documents the legacy OPAL-T wakefield interface. It should not be interpreted as a statement that the same user-facing command set is presently available in OPALX.

OPAL-T provides methods to model coherent synchrotron radiation and short-range geometric wakefields.

19.1 Geometric Wakefields

The legacy wakefield interface supports two main routes:

  1. compute the wakefield of a round metallic beam pipe from an impedance model
  2. import a discretized wake function from file

For a round metallic beam pipe with radius a, OPAL distinguishes DC and AC conductivity models. The classical Drude-style conductivities are

\[ \sigma_{\mathrm{DC}} = \frac{n e^2 \tau}{m} \tag{19.1}\]

and

\[ \sigma_{\mathrm{AC}} = \frac{\sigma_{\mathrm{DC}}}{1 - i \omega \tau}. \tag{19.2}\]

With these ingredients, the longitudinal impedance can be transformed back into the longitudinal wake. For the DC case the impedance model is

\[ Z_{L,\mathrm{DC}}(k) = \frac{1}{c a} \frac{2}{\lambda/k - i k a / 2} \tag{19.3}\]

with

\[ \lambda = \sqrt{\frac{2 \pi \sigma |k|}{c}} \left(i + \operatorname{sign}(k)\right). \tag{19.4}\]

The longitudinal wake is then obtained from the inverse cosine transform of the real part of the impedance,

\[ W_L(s) = 10^{-12} \frac{2 c}{\pi} \Re \left( \int_0^\infty \Re(Z_L(k)) \cos(k s)\, dk \right), \tag{19.5}\]

and the transverse wake follows through the Panofsky-Wenzel relation.

OPAL evaluates these integrals numerically with Simpson integration on an equidistant wave-number mesh. Since this is usually done once during initialization, the computational overhead is typically acceptable.

19.2 CSR Wakefields

The CSR model in OPAL-T is built around one-dimensional steady-state or quasi-steady-state approximations. The derivation starts from the Liénard-Wiechert fields of particles moving on a circular orbit and leads to impedance and wakefield models that act on trailing particles only.

The legacy manual distinguishes:

  • 1D-CSR
  • 1D-CSR-IGF

The 1D-CSR model follows the staged bend-entry, steady-state, and exit picture of Saldin, Schneidmiller, and Yurkov. The 1D-CSR-IGF variant uses an integrated Green function for faster wake evaluation.

Both models neglect transverse structure and several transient corrections, but they capture the dominant longitudinal CSR effect for short bunches in many beamline studies.

19.3 The WAKE Command

WAKE defines the wakefield data attached to an element.

General syntax:

label: WAKE, TYPE=string, NBIN=integer, CONST_LENGTH=bool,
             CONDUCT=string, Z0=real, FORM=string, RADIUS=real,
             SIGMA=real, TAU=real, FNAME=string, FILTERS={...};

For CSR wakes the reduced syntax is:

label: WAKE, TYPE=string, NBIN=integer, FILTERS={...};

19.3.1 Core attributes

Attribute Meaning
TYPE Wake type: 1D-CSR, 1D-CSR-IGF, LONG-SHORT-RANGE, TRANSV-SHORT-RANGE
NBIN Number of bins used to build the bunch line density
CONST_LENGTH Keep bunch length constant during wake evaluation
CONDUCT Beam-pipe conductivity model: AC or DC
Z0 Beam-pipe impedance in ohms
FORM Beam-pipe form; the documented choice is ROUND
RADIUS Beam-pipe radius in meters
SIGMA Material conductivity parameter
TAU Relaxation time entering the AC conductivity model
FNAME SDDS file containing a tabulated wake function
FILTERS Ordered list of filter names applied to the line density

19.3.2 Practical behavior

  • TYPE decides whether the command describes CSR or a short-range geometric wake.
  • CONDUCT, Z0, FORM, RADIUS, SIGMA, and TAU have no effect on CSR wake types.
  • FNAME imports a wake from SDDS data. Once a file is provided, OPAL ignores the analytic round-pipe parameters and uses the imported wake.
  • FILTERS are applied in the given order, and the last filter is also used to compute the derivatives needed by the wake model.

19.4 The FILTER Command

FILTER defines a smoothing or differentiation filter applied to the longitudinal histogram of the bunch.

Syntax:

label: FILTER, TYPE=string, NFREQ=integer, THRESHOLD=real,
               NPOINTS=integer, NLEFT=integer, NRIGHT=integer,
               POLYORDER=integer;

19.4.1 Filter types

Supported filters are:

  • SAVITZKY-GOLAY
  • STENCIL
  • FIXEDFFTLOWPASS
  • RELATIVEFFTLOWPASS

19.4.2 Filter attributes

Attribute Meaning
TYPE Select the filter algorithm
NFREQ Number of Fourier modes to keep for FIXEDFFTLOWPASS
THRESHOLD Relative spectral cutoff for RELATIVEFFTLOWPASS
NPOINTS Moving-window width for SAVITZKY-GOLAY
NLEFT Number of points to the left for SAVITZKY-GOLAY
NRIGHT Number of points to the right for SAVITZKY-GOLAY
POLYORDER Least-squares polynomial order for SAVITZKY-GOLAY

19.4.3 Implemented smoothing logic

SAVITZKY-GOLAY and the FFT-based filters naturally provide derivative information. For the STENCIL filter, OPAL applies two smoothing stencils consecutively:

\[ f_i = \frac{7 f_{i-4} + 24 f_{i-2} + 34 f_i + 24 f_{i+2} + 7 f_{i+4}}{96} \tag{19.6}\]

and

\[ f_i = \frac{7 f_{i-2} + 24 f_{i-1} + 34 f_i + 24 f_{i+1} + 7 f_{i+2}}{96}. \tag{19.7}\]

The derivative is then computed with the second-order stencil

\[ f'_i = \frac{f_{i-2} - 8 f_{i-1} + 8 f_{i+1} - f_{i+2}}{h}. \tag{19.8}\]

For the FFT-based smoothers, the line-density Fourier coefficients are computed, high-frequency content is removed according to the chosen criterion, and the smoothed profile is reconstructed by inverse transform.

  • FIXEDFFTLOWPASS keeps the NFREQ lowest modes.
  • RELATIVEFFTLOWPASS keeps modes above a threshold relative to the dominant coefficient.