"""
Radiation and equilibrium emittances
"""
from __future__ import annotations
from math import sin, cos, tan, sqrt, sinh, cosh, pi
import numpy
from typing import Union
from scipy.linalg import inv, det, solve_sylvester
from at.lattice import Lattice, check_radiation, Refpts, All
from at.lattice import Dipole, Wiggler, DConstant
from at.lattice import Quadrupole, Multipole, QuantumDiffusion
from at.lattice import frequency_control, set_value_refpts
from at.tracking import internal_lpass
from at.physics import find_orbit6, find_m66, find_elem_m66, Orbit
from at.physics import find_mpole_raddiff_matrix, get_tunes_damp
from at.physics import ELossMethod
__all__ = ['ohmi_envelope', 'get_radiation_integrals', 'quantdiffmat',
'gen_quantdiff_elem', 'tapering']
_NSTEP = 60 # nb slices in a wiggler period
_submat = [slice(0, 2), slice(2, 4), slice(6, 3, -1)]
# dtype for structured array containing optical parameters
ENVELOPE_DTYPE = [('r66', numpy.float64, (6, 6)),
('r44', numpy.float64, (4, 4)),
('m66', numpy.float64, (6, 6)),
('orbit6', numpy.float64, (6,)),
('emitXY', numpy.float64, (2,)),
('emitXYZ', numpy.float64, (3,))]
def _dmatr(ring: Lattice, orbit: Orbit = None, keep_lattice: bool = False):
"""
compute the cumulative diffusion and orbit
matrices over the ring
"""
def _cumulb(it):
"""accumulate diffusion matrices"""
cumul = numpy.zeros((6, 6))
yield cumul
for el, orbin, b in it:
m = find_elem_m66(el, orbin, energy=energy, particle=ring.particle)
cumul = m.dot(cumul).dot(m.T) + b
yield cumul
energy = ring.energy
if orbit is None:
orbit, _ = find_orbit6(ring, keep_lattice=keep_lattice)
keep_lattice = True
orbs = numpy.squeeze(
internal_lpass(ring, orbit.copy(order='K'), refpts=All,
keep_lattice=keep_lattice), axis=(1, 3)).T
b0 = numpy.zeros((6, 6))
bb = [find_mpole_raddiff_matrix(elem, elemorb, energy)
if elem.PassMethod.endswith('RadPass') else b0
for elem, elemorb in zip(ring, orbs)]
bbcum = numpy.stack(list(_cumulb(zip(ring, orbs, bb))), axis=0)
return bbcum, orbs
def _lmat(dmat):
"""
lmat is Cholesky decomp of dmat unless diffusion is 0 in
vertical. Then do chol on 4x4 hor-long matrix and put 0's
in vertical
"""
lmat = numpy.zeros((6, 6))
try:
lmat = numpy.linalg.cholesky(dmat)
except numpy.linalg.LinAlgError:
nz = numpy.where(dmat != 0)
cmat = numpy.reshape(dmat[nz], (4, 4))
cmat = numpy.linalg.cholesky(cmat)
lmat[nz] = numpy.reshape(cmat, (16,))
return lmat
[docs]
@check_radiation(True)
def ohmi_envelope(ring: Lattice, refpts: Refpts = None, orbit: Orbit = None,
keep_lattice: bool = False):
"""Calculates the equilibrium beam envelope
Computation based on Ohmi's beam envelope formalism [1]_
Parameters:
ring: Lattice description. Radiation must be ON
refpts: Observation points
orbit: Avoids looking for initial the closed orbit if it is
already known ((6,) array).
keep_lattice: Assume no lattice change since the previous tracking.
Default: False
Returns:
emit0 (numpy.recarray): Emittance data at the start/end of the ring
beamdata (numpy.recarray): Beam parameters at the start of the ring
emit (numpy.recarray): Emittance data at the points selected to by
``refpts``
**emit** is a :py:obj:`record array <numpy.recarray>` with the following
fields:
================ ===================================================
**r66** (6, 6) equilibrium envelope matrix R
**r44** (4, 4) betatron emittance matrix (dpp = 0)
**m66** (6, 6) transfer matrix from the start of the ring
**orbit6** (6,) closed orbit
**emitXY** (2,) betatron emittance projected on xxp and yyp
**emitXYZ** (3,) 6x6 emittance projected on xxp, yyp, ldp
================ ===================================================
Values given at the entrance of each element specified in ``refpts``.
Field values can be obtained with either
``emit['r66']`` or ``emit.r66``
**beamdata** is a :py:obj:`record array <numpy.recarray>` with the
following fields:
==================== ===================================================
**tunes** tunes of the 3 normal modes
**damping_rates** damping rates of the 3 normal modes
**mode_matrices** R-matrices of the 3 normal modes
**mode_emittances** equilibrium emittances of the 3 normal modes
==================== ===================================================
References:
.. [1] K.Ohmi et al. Phys.Rev.E. Vol.49. (1994)
"""
def process(r66):
# projections on xx', zz', ldp
emit3sq = numpy.array([det(r66[s, s]) for s in _submat])
# Prevent from unrealistic negative values of the determinant
emit3 = numpy.sqrt(numpy.maximum(emit3sq, 0.0))
# Emittance cut for dpp=0
if emit3[0] < 1.E-13: # No equilibrium emittance
r44 = numpy.nan * numpy.ones((4, 4))
elif emit3[1] < 1.E-13: # Uncoupled machine
minv = inv(r66[[0, 1, 4, 5], :][:, [0, 1, 4, 5]])
r44 = numpy.zeros((4, 4))
r44[:2, :2] = inv(minv[:2, :2])
else: # Coupled machine
minv = inv(r66)
r44 = inv(minv[:4, :4])
# betatron emittances (dpp=0)
emit2sq = numpy.array(
[det(r44[s, s], check_finite=False) for s in _submat[:2]])
# Prevent from unrealistic negative values of the determinant
emit2 = numpy.sqrt(numpy.maximum(emit2sq, 0.0))
return r44, emit2, emit3
def propag(m, cumb, orbit6):
"""Propagate the beam matrix to refpts"""
sigmatrix = m.dot(rr).dot(m.T) + cumb
m44, emit2, emit3 = process(sigmatrix)
return sigmatrix, m44, m, orbit6, emit2, emit3
uint32refs = ring.get_uint32_index(refpts)
bbcum, orbs = _dmatr(ring, orbit=orbit, keep_lattice=keep_lattice)
mring, ms = find_m66(ring, uint32refs, orbit=orbs[0], keep_lattice=True)
# ------------------------------------------------------------------------
# Equation for the moment matrix R is
# R = MRING*R*MRING' + BCUM;
# We rewrite it in the form of Lyapunov-Sylvester equation to use scipy's
# solve_sylvester function
# A*R + R*B = Q
# where
# A = inv(MRING)
# B = -MRING'
# Q = inv(MRING)*BCUM
# ------------------------------------------------------------------------
aa = inv(mring)
bb = -mring.T
qq = numpy.dot(aa, bbcum[-1])
rr = solve_sylvester(aa, bb, qq)
rr = 0.5 * (rr + rr.T)
rr4, emitxy, emitxyz = process(rr)
r66data = get_tunes_damp(mring, rr)
data0 = numpy.rec.fromarrays(
(rr, rr4, mring, orbs[0], emitxy, emitxyz),
dtype=ENVELOPE_DTYPE)
if uint32refs.shape == (0,):
data = numpy.recarray((0,), dtype=ENVELOPE_DTYPE)
else:
data = numpy.rec.fromrecords(
list(map(propag, ms, bbcum[uint32refs], orbs[uint32refs, :])),
dtype=ENVELOPE_DTYPE)
return data0, r66data, data
[docs]
@frequency_control
def get_radiation_integrals(ring, dp: float = None, twiss=None, **kwargs)\
-> tuple[float, float, float, float, float]:
r"""Computes the 5 radiation integrals for uncoupled lattices.
Parameters:
ring: Lattice description
twiss: Linear optics at all points (from :py:func:`.linopt6`).
If ``None``, it will be computed.
Keyword Args:
dp (float): Momentum deviation. Defaults to :py:obj:`None`
dct (float): Path lengthening. Defaults to :py:obj:`None`
df (float): Deviation of RF frequency. Defaults to
method (Callable): Method for linear optics:
:py:obj:`~.linear.linopt2`: no longitudinal motion, no H/V coupling,
:py:obj:`~.linear.linopt6` (default): with or without longitudinal
motion, normal mode analysis
Returns:
i1 (float): Radiation integrals - :math:`I_1 \quad [m]`
i2 (float): :math:`I_2 \quad [m^{-1}]`
i3 (float): :math:`I_3 \quad [m^{-2}]`
i4 (float): :math:`I_4 \quad [m^{-1}]`
i5 (float): :math:`I_5 \quad [m^{-1}]`
"""
def element_radiation(elem: Union[Dipole, Quadrupole], vini, vend):
"""Analytically compute the radiation integrals in dipoles"""
beta0 = vini.beta[0]
alpha0 = vini.alpha[0]
eta0 = vini.dispersion[0]
etap0 = vini.dispersion[1]
theta = getattr(elem, 'BendingAngle', None)
if theta is None:
xpi = vini.closed_orbit[1]/(1+vini.closed_orbit[4])
xpo = vend.closed_orbit[1]/(1+vend.closed_orbit[4])
theta = xpi-xpo
if abs(theta) < 1.0e-7:
return numpy.zeros(5)
angin = getattr(elem, 'EntranceAngle', 0.0)
angout = getattr(elem, 'ExitAngle', 0.0)
ll = elem.Length
rho = ll / theta
rho2 = rho * rho
k2 = elem.K + 1.0 / rho2
eps1 = tan(angin) / rho
eps2 = tan(angout) / rho
eta3 = vend.dispersion[0]
alpha1 = alpha0 - beta0 * eps1
gamma1 = (1.0 + alpha1 * alpha1) / beta0
etap1 = etap0 + eta0 * eps1
etap2 = vend.dispersion[1] - eta3 * eps2
h0 = gamma1*eta0*eta0 + 2.0*alpha1*eta0*etap1 + beta0*etap1*etap1
if k2 != 0.0:
if k2 > 0.0: # Focusing
kl = ll * sqrt(k2)
ss = sin(kl) / kl
cc = cos(kl)
else: # Defocusing
kl = ll * sqrt(-k2)
ss = sinh(kl) / kl
cc = cosh(kl)
eta_ave = (theta - (etap2 - etap1)) / k2 / ll
bb = 2.0 * (alpha1 * eta0 + beta0 * etap1) * rho
aa = -2.0 * (alpha1 * etap1 + gamma1 * eta0) * rho
h_ave = h0 + (aa * (1.0 - ss) + bb * (1.0 - cc) / ll
+ gamma1 * (3.0 - 4.0 * ss + ss * cc) / 2.0 / k2
- alpha1 * (1.0 - cc) ** 2 / k2 / ll
+ beta0 * (1.0 - ss * cc) / 2.0
) / k2 / rho2
else:
eta_ave = 0.5 * (eta0 + eta3) - ll * ll / 12.0 / rho
hp0 = 2.0 * (alpha1 * eta0 + beta0 * etap1) / rho
h2p0 = 2.0 * (-alpha1 * etap1 + beta0 / rho - gamma1 * eta0) / rho
h_ave = h0 + hp0 * ll / 2.0 + h2p0 * ll * ll / 6.0 \
- alpha1 * ll ** 3 / 4.0 / rho2 \
+ gamma1 * ll ** 4 / 20.0 / rho2
di1 = eta_ave * ll / rho
di2 = ll / rho2
di3 = ll / abs(rho) / rho2
di4 = eta_ave * ll * (2.0 * elem.K + 1.0 / rho2) / rho \
- (eta0 * eps1 + eta3 * eps2) / rho
di5 = h_ave * ll / abs(rho) / rho2
return numpy.array([di1, di2, di3, di4, di5])
def wiggler_radiation(elem: Wiggler, dini):
"""Compute the radiation integrals in wigglers with the following
approximations:
- The wiggler is aligned with the closed orbit
- The self-induced dispersion is neglected in I4 and I5, but is is used
as a lower limit for the I5 contribution
- I1, I2 are integrated analytically
- I3 is integrated analytically for a single harmonic, numerically
otherwise
"""
def b_on_axis(wiggler: Wiggler, s):
"""On-axis wiggler field"""
def harm(coef, h, phi):
return -Bmax * coef * numpy.cos(h*kws + phi)
kw = 2 * pi / wiggler.Lw
Bmax = wiggler.Bmax
kws = kw * s
zz = [numpy.zeros(kws.shape)]
vh = zz + [harm(pb[1], pb[4], pb[5]) for pb in wiggler.By.T]
vv = zz + [-harm(pb[1], pb[4], pb[5]) for pb in wiggler.Bx.T]
bys = numpy.sum(numpy.stack(vh), axis=0)
bxs = numpy.sum(numpy.stack(vv), axis=0)
return bxs, bys
le = elem.Length
alphax0 = dini.alpha[0]
betax0 = dini.beta[0]
gammax0 = (alphax0 * alphax0 + 1) / betax0
eta0 = dini.dispersion[0]
etap0 = dini.dispersion[1]
H0 = gammax0*eta0*eta0 + 2*alphax0*eta0*etap0 + betax0*etap0*etap0
avebetax = betax0 + alphax0*le + gammax0*le*le/3
kw = 2 * pi / elem.Lw
rhoinv = elem.Bmax / Brho
coefh = elem.By[1, :] * rhoinv
coefv = elem.Bx[1, :] * rhoinv
coef2 = numpy.concatenate((coefh, coefv))
if len(coef2) == 1:
di3 = le * coef2[0] ** 3 * 4 / 3 / pi
else:
bx, bz = b_on_axis(elem, numpy.linspace(0, elem.Lw, _NSTEP + 1))
rinv = numpy.sqrt(bx*bx + bz*bz) / Brho
di3 = numpy.trapz(rinv ** 3) * le / _NSTEP
di2 = le * (numpy.sum(coefh * coefh) + numpy.sum(coefv * coefv)) / 2
di1 = -di2 / kw / kw
di4 = 0
if len(coefh) > 0:
d5lim = 4 * avebetax * le * coefh[0] ** 5 / 15 / pi / kw / kw
else:
d5lim = 0
di5 = max(H0 * di3, d5lim)
return numpy.array([di1, di2, di3, di4, di5])
Brho = ring.BRho
integrals = numpy.zeros((5,))
if twiss is None:
_, _, twiss = ring.get_optics(refpts=range(len(ring) + 1), dp=dp,
get_chrom=True, **kwargs)
elif len(twiss) != len(ring) + 1:
raise ValueError('length of Twiss data should be {0}'
.format(len(ring) + 1))
for (el, vini, vend) in zip(ring, twiss[:-1], twiss[1:]):
if isinstance(el, (Dipole, Quadrupole)):
integrals += element_radiation(el, vini, vend)
elif isinstance(el, Wiggler) and el.PassMethod != 'DriftPass':
integrals += wiggler_radiation(el, vini)
return tuple(integrals)
[docs]
@check_radiation(True)
def quantdiffmat(ring: Lattice, orbit: Orbit = None) -> numpy.ndarray:
"""Computes the diffusion matrix of the whole ring
Parameters:
ring: Lattice description. Radiation must be ON
orbit: Avoids looking for initial the closed orbit if it is
already known ((6,) array).
Returns:
diffmat (ndarray): Diffusion matrix (6,6)
"""
bbcum, _ = _dmatr(ring, orbit=orbit)
diffmat = [(bbc + bbc.T) / 2 for bbc in bbcum]
return numpy.round(diffmat[-1], 24)
[docs]
@check_radiation(True)
def gen_quantdiff_elem(ring: Lattice, orbit: Orbit = None) -> QuantumDiffusion:
"""Generates a quantum diffusion element
Parameters:
ring: Lattice description. Radiation must be ON
orbit: Avoids looking for initial the closed orbit if it is
already known ((6,) array).
Returns:
diffElem (QuantumDiffusion): Quantum diffusion element
"""
dmat = quantdiffmat(ring, orbit=orbit)
lmat = numpy.asfortranarray(_lmat(dmat))
diff_elem = QuantumDiffusion('Diffusion', lmat)
return diff_elem
[docs]
@check_radiation(True)
def tapering(ring: Lattice, multipoles: bool = True,
niter: int = 1, **kwargs) -> None:
"""Scales magnet strengths
Scales magnet strengths with local energy to cancel the closed orbit
and optics errors due to synchrotron radiations. The dipole bending angle
is changed by changing the reference momentum. This is the ideal
tapering scheme where magnets and multipoles components (PolynomB and
PolynomA) are scaled individually.
Warning:
This method works only for lattices without errors and
corrections: if not all corrections and field errors will also be
scaled !!!
Parameters:
ring: Lattice description
multipoles: Scales all multipoles
niter: Number of iteration
Keyword Args:
method (ELossMethod): Method for energy loss computation.
See :py:class:`.ELossMethod`
XYStep (float): Step size.
Default: :py:data:`DConstant.XYStep <.DConstant>`
DPStep (float): Momentum step size.
Default: :py:data:`DConstant.DPStep <.DConstant>`
"""
xy_step = kwargs.pop('XYStep', DConstant.XYStep)
dp_step = kwargs.pop('DPStep', DConstant.DPStep)
method = kwargs.pop('method', ELossMethod.TRACKING)
dipin = ring.get_bool_index(Dipole)
dipout = numpy.roll(dipin, 1)
multin = ring.get_bool_index(Multipole) & ~dipin
multout = numpy.roll(multin, 1)
for i in range(niter):
_, o6 = find_orbit6(ring, refpts=range(len(ring)+1),
XYStep=xy_step, DPStep=dp_step, method=method)
dpps = (o6[dipin, 4] + o6[dipout, 4]) / 2.0
set_value_refpts(ring, dipin, 'FieldScaling', 1+dpps)
if multipoles:
_, o6 = find_orbit6(ring, refpts=range(len(ring)+1),
XYStep=xy_step, DPStep=dp_step, method=method)
dppm = (o6[multin, 4] + o6[multout, 4]) / 2
set_value_refpts(ring, multin, 'FieldScaling', 1+dppm)
Lattice.ohmi_envelope = ohmi_envelope
Lattice.get_radiation_integrals = get_radiation_integrals
Lattice.tapering = tapering
