-
Notifications
You must be signed in to change notification settings - Fork 0
/
017_the_loop.py
234 lines (173 loc) · 7.47 KB
/
017_the_loop.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
import socket
import itertools
import numpy as np
from scipy.constants import c
import matplotlib.pyplot as plt
from h5_storage import loadH5Recursive
from EmittanceTool.gaussfit import GaussFit
import wf_model
import elegant_matrix
import myplotstyle as ms
plt.close('all')
# Optimizer around elegant tracking
# 1.
# - Get initial distribution by fitting 4 parameters:
# - heights of left spike, right spike, middle
# - width of spikes
# - Assume known chamber offset
# 2.
# - Fit chamber offset as well (using images from different offsets with known delta)
# 3.
# - Fit 21 slice heights
#
# 4.
# Better approach
# Strategy:
# 1. Wakefield for any simulated or measured profiles can be well approximated by 5-th order polynomial function (≤ 5 parameters)
# 2. Guess the initial wakefield model (single-gaussian wake, etc.)
# 3. Back-propagate the measured profile to the time domain
# 4. Calculate the real wakefield using model
# 5. Forward track to the screen (Model-based code or Elegant)
# 6. Compare with measurements
# 7. Adjust the wakefield model
# 8. Loop steps 2-7 = optimization problem
# files from alex
files_labels = [
('SpectrumAnalysis_2020_07_26_17_56_25_401050.h5', 'no streaking'),
('SpectrumAnalysis_2020_07_26_17_55_28_676729.h5', '-4.04 mm dump'),
('SpectrumAnalysis_2020_07_26_17_46_52_221317.h5', '-4.09 mm dump'),
]
energy_eV = 6.14e9
gap = 10e-3
fit_order = 4
show_4_images = False
zoom_sig = 3
elegant_matrix.set_tmp_dir('/home/philipp/tmp_elegant/')
hostname = socket.gethostname()
if hostname == 'desktop':
save_directory = '/storage/data_2020-07-26/'
other = '/storage/data_2020-10-04/20201004_175405_undulator_wf.h5'
magnet_file = '/storage/Philipp_data_folder/archiver_api_data/2020-07-26.h5'
else:
save_directory = '/sf/data/measurements/2020/07/26/'
other = '/sf/data/measurements/2020/10/04/20201004_175405_undulator_wf.h5'
magnet_file = '/afs/psi.ch/intranet/SF/Beamdynamics/Philipp/data/archiver_api_data/2020-07-26.h5'
simulator = elegant_matrix.get_simulator(magnet_file)
timestamp = elegant_matrix.get_timestamp(2020, 7, 26, 17, 49, 0)
mat_dict, disp_dict = simulator.get_elegant_matrix(0, timestamp)
r12 = mat_dict['SARBD02.DSCR050'][0,1]
other_dict = loadH5Recursive(other)
x_axis = other_dict['x_axis'].astype(np.float64) * 1e-6
y_axis = other_dict['y_axis'].astype(np.float64) * 1e-6
del other, other_dict
### Generate Gaussian beam to calculate initial wake potential
# approximate 40 fs beam
sig_t = 40e-15
time = np.linspace(0., 400e-15, 1000)
curr = np.exp(-(time-np.mean(time))**2/(2*sig_t**2))
curr[curr<0.001*curr.max()] = 0
curr = curr * 200e-12/np.sum(curr)
wf_calc = wf_model.WakeFieldCalculator(time*c, curr, energy_eV, 1.)
wf_dict = wf_calc.calc_all(gap/2., r12, beam_offset=(-4.09e-3-0.47e-3), calc_lin_dipole=True, calc_dipole=False, calc_quadrupole=False, calc_long_dipole=False)
wake_zz = wf_dict['input']['charge_xx']
convoluted_wake = wf_dict['lin_dipole']['wake_potential'] * -1
ms.figure('Wakefields')
subplot = ms.subplot_factory(2,2)
sp_ctr = 1
sp_wake = subplot(sp_ctr, title='Initial wake from Gaussian $\sigma_t$=%i fs' % (sig_t*1e15), xlabel='s [$\mu$m]', ylabel='kV/m')
sp_ctr += 1
sp_wake.plot(wake_zz*1e6, convoluted_wake/1e3, label='Wake')
poly_fit = np.poly1d(np.polyfit(wake_zz, convoluted_wake, fit_order))
fitted_wake = poly_fit(wake_zz)
sp_wake.plot(wake_zz*1e6, fitted_wake/1e3, label='Fit')
sp_wake.legend()
def zoom_image(image, x_axis, y_axis, n_sig):
mask = {}
gf_dict = {}
for axis, proj, label in [(x_axis, np.sum(image, 0), 'X'), (y_axis, np.sum(image, 1), 'Y')]:
gf = GaussFit(axis, proj)
min_ax, max_ax = gf.mean - gf.sigma * n_sig, gf.mean + gf.sigma * n_sig
mask[label] = np.logical_and(axis > min_ax, axis < max_ax)
gf_dict[label] = gf
return {
'image': image[mask['Y']][:,mask['X']],
'x_axis': x_axis[mask['X']],
'y_axis': y_axis[mask['Y']],
'gf_dict': gf_dict,
}
if show_4_images:
for file_, file_label in files_labels:
meas_dict = loadH5Recursive(save_directory+file_)
images = meas_dict['scan_1']['data']['SARBD02-DSCR050']['FPICTURE'].astype(np.float64)
ms.figure('Check image %s' % file_label)
subplot = ms.subplot_factory(2,2)
sp_ctr = 1
for n_image in range(4):
image = images[n_image,0]
image_dict = zoom_image(image, x_axis, y_axis, zoom_sig)
image_zoom = image_dict['image']
ax_x = image_dict['x_axis']
ax_y = image_dict['y_axis']
gf_dict = image_dict['gf_dict']
sp = subplot(sp_ctr, title='Image sig=%i $\mu$m' % (gf_dict['X'].sigma*1e6), grid=False)
sp_ctr += 1
sp.imshow(image_zoom, aspect='auto', extent=(ax_x[0], ax_x[-1], ax_y[-1], ax_y[0]))
file_ = files_labels[-1][0]
meas_dict = loadH5Recursive(save_directory+file_)
images = meas_dict['scan_1']['data']['SARBD02-DSCR050']['FPICTURE'].astype(np.float64)
ms.figure('All projX')
subplot = ms.subplot_factory(2,3)
sp_ctr = 1
sp = subplot(sp_ctr, title='All projections', xlabel='Screen X [$\mu$m]', ylabel='Intensity (arb. units)')
sp_ctr += 1
proj_list, axis_list, mean_X_list = [], [], []
for n1, n2 in itertools.product(range(images.shape[0]), range(images.shape[1])):
image = images[n1, n2]
zoom_dict = zoom_image(image, x_axis, y_axis, zoom_sig)
image2 = zoom_dict['image']
x_axis2 = zoom_dict['x_axis'] # - zoom_dict['gf_dict']['X'].mean
projX = np.sum(image2, axis=0)
proj_list.append(projX)
axis_list.append(x_axis2)
mean_X_list.append(zoom_dict['gf_dict']['X'].mean)
sp.plot(x_axis2*1e6, projX)
mean_X = np.mean(mean_X_list)
file0 = files_labels[0][0]
meas_dict0 = loadH5Recursive(save_directory+file0)
images0 = meas_dict0['scan_1']['data']['SARBD02-DSCR050']['FPICTURE'].astype(np.float64)
mean_X0_list = []
for n1, n2 in itertools.product(range(images0.shape[0]), range(images0.shape[1])):
image = images0[n1, n2]
gf = GaussFit(x_axis, np.sum(image, axis=0))
mean_X0_list.append(gf.mean)
mean_X0 = np.mean(mean_X0_list)
min_x = min(np.min(axis) for axis in axis_list)
max_x = max(np.max(axis) for axis in axis_list)
comb_axis = np.linspace(min_x, max_x, int(200))
interp_proj = np.zeros([len(proj_list), len(comb_axis)])
for n_axis, (axis, proj) in enumerate(zip(axis_list, proj_list)):
interp_proj[n_axis] = np.interp(comb_axis, axis[::-1], proj[::-1])
comb_proj = np.mean(interp_proj, axis=0)
comb_proj_std = np.std(interp_proj, axis=0)
x_screen_axis = comb_axis - mean_X0
x_screen = comb_proj
xp_struct_axis = -x_screen_axis / r12
xp_struct = comb_proj
xp_struct_std = comb_proj_std
sp = subplot(sp_ctr, title='Average X Screen', xlabel='Screen X [$\mu$m]', ylabel='Intensity (arb. units)')
sp_ctr += 1
sp.errorbar(comb_axis*1e6, comb_proj, yerr=comb_proj_std)
sp.axvline(mean_X0*1e6, color='red')
sp.axvline(mean_X*1e6, color='blue')
sp = subplot(sp_ctr, title='Average Xp structure', xlabel='Structure Xp [mrad]', ylabel='Intensity (arb. units)')
sp_ctr += 1
sp.errorbar(xp_struct_axis*1e3, xp_struct, yerr=xp_struct_std)
wf_normalized = convoluted_wake / energy_eV
sp = subplot(sp_ctr, title='Normalized wake', xlabel='z [$\mu$m]', ylabel='Xp [mrad]')
sp_ctr += 1
sp.plot(wake_zz*1e6, wf_normalized*1e3)
interp_zz = np.interp(-xp_struct_axis, wf_normalized, wake_zz)
sp = subplot(sp_ctr, title='Interpolated Z', xlabel='z [$\mu$m]', ylabel='Intensity (arb. units)')
sp_ctr += 1
sp.plot(interp_zz*1e6, xp_struct)
plt.show()