The file Simulation_parameters.py allows configuring your PyECLOUD-PyHEADTAIL single-bunch simulation. In the following we describe its different parts:
In the beginning of the file you can import python modules or objects required to configure the simulation. In this example we import some constants:
from scipy.constants import c
from scipy.constants import m_p
from scipy.constants import e as qeBy choosing "Synchrotron" as machine class it is possible to define all parameters of the ring (this is the recommended way of configuring the simulation):
machine_class = 'Synchrotron'The description of the accelerator is defined by providing the following parameters that are required to define a PyHEADTAIL synchrotron. They are individually described in the PyHEADTAIL documentation.
optics_mode = 'smooth'
charge = qe
mass = m_p
p0 = 450e9 * qe / c
circumference = 26658.8832
n_segments = 16
name = None
s = None
alpha_x = 0.
beta_x = 92.7
D_x = 0.
alpha_y = 0
beta_y = 93.2
D_y = 0.
accQ_x = 62.27
accQ_y = 60.295
Qp_x = 0.
Qp_y = 0.
app_x = 0.
app_y = 0.
app_xy = 0.
longitudinal_mode = 'non-linear'
Q_s = None
alpha_mom_compaction = 3.225e-04
h_RF = 35640
V_RF = 5e6
dphi_RF = 0.
p_increment = 0.
RF_at = 'end_of_transverse' # needs to be at the end to be compatible with PyPARIS parallelization
wrap_z = False
other_detuners = []The parameter n_non_parallelizable (used in previous versions) is now handled automatically and should not be added to the input file.
An ideal bunch-by-bunch transverse feedback can be enabled. The damping rates are defined as the time constant observed on the envelope of the bunch centroid motion in the presence of the feedback alone:
# Transverse Damper Settings
enable_transverse_damper = False
dampingrate_x = 100.
dampingrate_y = 100.The parameters in this section define the bunch.
Optionally the bunch (particle-by-particle coordinates) can be loaded from file using the following intput variable:
bunch_from_file = NoneAn example on how to save and load a bunch can be found in PyPARIS_sim_class/examples.
Alternatively, the bunch can be defined using the following input
intensity = 1.2e+11 # Number of pearticles in the bunch
epsn_x = 2.5e-6 # Normalized horizontal r.m.s emittance [m]
epsn_y = 2.5e-6 # Normalized vertical r.m.s emittance [m]
sigma_z = 9.181144e-02 # R.m.s bunch length [m]The generated bunch is matched in all planes.
An initial transverse displacement can be defined by the following parameters:
x_kick_in_sigmas = 0.1
y_kick_in_sigmas = 0.1The average position and angle of the bunch longitudinal slices can be shifted to zero (to correct for random fluctuations) by setting:
recenter_all_slices = TrueThe beam is sliced longitudinally to compute its interaction with the e-cloud:
n_slices = 200
z_cut = 2.5e-9/2*c # For slicingz_cut defines the portion of the bunch that is sliced, which is [-z_cut, z_cut].
The number of macroparticles in the bunch is defined by the following parameters:
macroparticles_per_slice = 5000
n_macroparticles = macroparticles_per_slice*n_slices
For very long simulations it is convenient to split the simulations over several shorter jobs. This allows also recovering the simulation from the end of the last success full job in case of problems. This is controlled through the following parameters:
N_turns = 128 # Per job
N_turns_target = 20000 # Entire simulationFor clusters using the LSF management system the job can be resubmitted automatically by setting:
check_for_resubmit = Trueotherwise the user has to take care of the resubmission of the job.
The information about the multijob run is kept in the file simulation_status.sta. This file needs to be removed to start the simulation from scratch.
The simulation can be ended in case a certain fraction of the initial intensity is lost:
sim_stop_frac = 0.9 # Simulation stopped if 10% of the initial intensity is lostTransverse emittance blow-up can also be used as a criterion for ending the simulation:
flag_check_emittance_growth = True
epsn_x_max_growth_fraction = 0.5 # Stop on 50% horizontal emittance blow-up
epsn_y_max_growth_fraction = 0.5 # Stop on 50% vertical emittance blow-upThe ouput data is buffered and dumped to file at regular intervals. The lenght of such intervals can be specified by setting:
write_buffer_every = 3
slice_stats_to_store = [mean_x, mean_z] The simulation mode can be changed to compute the tune footprint in the presence of e-cloud using the footprint_mode flag. In this case the instability simulation is not performed and only the footprint data is produced. A bunch with a very large number of macroparticles is used to compute the e-cloud field maps. The footprint particle tunes are measured using a smaller number of macroparticles. For the computation of the footprint the longitudinal motion is automatically switched off.
The footprint mode is enabled and controlled using the following parameters:
footprint_mode = False
n_macroparticles_for_footprint_map = 500000
n_macroparticles_for_footprint_track = 5000Two kinds of e-cloud interaction can be installed at the end of each machine segment to model the e-cloud in dipoles and quadrupoles.
Most of the e-cloud parameters are defined by PyECLOUD input files in a specified folder:
pyecl_input_folder = './pyecloud_config'Whenever a parameter in the input files is also mention in the present file, the value specified here takes priority.
The chamber geometry can be redefined by the following parameters (see PyECLOUD reference for more details):
chamb_type = 'polyg'
x_aper = 2.300000e-02
y_aper = 1.800000e-02
filename_chm = 'LHC_chm_ver.mat'The target duration of the electron tracking time sub-steps is defined by:
Dt_ref = 5.000000e-12
The Particle-In-Cell solver used to compute beam and electron fields is configured by the following parameters (more details can be found in the PyECLOUD reference and in the physics models section of the PyECLOUD wiki):
PyPICmode = 'ShortleyWeller_WithTelescopicGrids'
N_min_Dh_main = 10.
Dh_sc_ext = .8e-3
f_telescope = 0.3
N_nodes_discard = 5.
target_size_internal_grid_sigma = 10.
target_Dh_internal_grid_sigma = 0.2
custom_target_grid_arcs = None
# # Uncomment for custom grid
# custom_target_grid_arcs = {
# 'x_min_target': -3e-3,
# 'x_max_target': 3e-3,
# 'y_min_target': -3.1e-3,
# 'y_max_target': 3.1e-3,
# 'Dh_target': 7e-5}By setting:
force_interp_at_substeps_interacting_slices = Truethe electric field acting on the electrons is re-interpolated at each tracking sub-step. If this parameter is set to false the field is evaluated only once for each beam-slice.
The forces of the e-cloud on the beam particle in the transverse planes can be enabled/disabled using the following flags:
enable_kick_x = True
enable_kick_y = FalseThe following parameters configure the e-cloud in the dipoles (which are simulated starting with electrons at rest uniformly distributed in the chamber):
enable_arc_dip = True # Activate interaction with e-cloud in the dipoles
fraction_device_dip = 0.65 # Fraction of the machine circumference with e-cloud in the dipoles
init_unif_edens_flag_dip = 1 # Activate initial uniform distribution for electrons
init_unif_edens_dip = 1.0e+12 # Initial electron density (e-/m^3)
N_MP_ele_init_dip = 500000 # Number of macroparticles
N_mp_max_dip = N_MP_ele_init_dip*4 # Size of arrazys used to store macroparticle coordinates
B_multip_dip = [0.5] #T # Magnetic field (in Tesla)The following parameters configure the e-cloud in the quadrupole magnets (which are simulated starting with an electron distribution loaded from file):
enable_arc_quad = False # Activate interaction with e-cloud in the dipoled
fraction_device_quad = 7.0e-02 # Fraction of the machine circumference with e-cloud in the quadrupoles
N_mp_max_quad = 2000000 # Size of arrazys used to store macroparticle coordinates
B_multip_quad = [0., 12.1] #T # The second element of the list is the magnetic field gradient (in Tesla/m)The following parameters are used to define the file containing the initial e-cloud distribution:
folder_path = '../../LHC_ecloud_distrib_quads/'
N_mp_ele_quad = 500000
sey_load_quad = 1.3
filename_state = 'combined_distribution_sey_%.2f_sigmat_%.3fns_450Gev_N_mp_%d_symm'%(sey_load_quad, sigma_z/c*1e9,N_mp_ele_quad)
filename_init_MP_state_quad = folder_path + filename_stateIndividual e-cloud interactions can be provided at defined kick locations when the used optics is 'non-smooth' (incompatible with the settings above), using the following parameters:
# Dedicated Kick Element Settings
enable_eclouds_at_kick_elements = False
path_buildup_simulations_kick_elements = 'path_to_myfolder/simulations_PyECLOUD/!!!NAME!!!_sey1.35'
name_MP_state_file_kick_elements = 'MP_state_9.mat'
orbit_factor = 6.250000e-01A simple resonator impedance can be included in the simulation by using the following parameters:
enable_impedance = True
resonator_R_shunt = 25e6
resonator_frequency = 2e9
resonator_Q = 1.
n_slices_wake = 500