def interpolate_fields_to_particle_time():
'''
For each particle timestep, interpolate field values
RECODE THIS TO USE NP.INTERPOLATE()
'''
bx, by, bz, ex, ey, ez, vex, vey, vez, te, jx, jy, jz, qdens = cf.get_array(
get_all=True)
time_particles = cf.time_seconds_particle
time_fields = cf.time_seconds_field
pbx, pby, pbz, pex, pey, pez, pvex, pvey, pvez, pte, pjx, pjy, pjz, pqdens = \
[np.zeros((time_particles.shape[0], cf.NX)) for _ in range(14)]
for ii in range(time_particles.shape[0]):
this_time = time_particles[ii] # Target interpolant
diff = abs(this_time - time_fields) # Difference matrix
nearest_idx = np.where(
diff == diff.min())[0][0] # Index of nearest value
if time_fields[nearest_idx] < this_time:
case = 1
lidx = nearest_idx
uidx = nearest_idx + 1
elif time_fields[nearest_idx] > this_time:
case = 2
uidx = nearest_idx
lidx = nearest_idx - 1
def analyse_helicity(overwrite=False, save=True):
By_raw = cf.get_array('By')
Bz_raw = cf.get_array('Bz')
Bt_pos, Bt_neg = bk.get_helical_components(overwrite)
By_pos = Bt_pos.real
By_neg = Bt_neg.real
Bz_pos = Bt_pos.imag
Bz_neg = Bt_neg.imag
t_idx1 = 200
t_idx2 = 205
if False:
'''
Check that helicity preserves transverse amplitude on transformation : Affirmative
'''
hel_tot = np.sqrt(
np.square(By_pos + By_neg) + np.square(Bz_pos + Bz_neg))
raw_tot = np.sqrt(np.square(By_raw) + np.square(Bz_raw))
plt.figure()
plt.plot(raw_tot[t_idx1, :], label='raw B')
plt.plot(hel_tot[t_idx1, :], label='helicty B')
plt.legend()
if False:
'''
Peak finder I was going to use for velocity
'''
peaks1 = bk.basic_S(By_pos[t_idx1, :], k=100)
peaks2 = bk.basic_S(By_pos[t_idx2, :], k=100)
plt.plot(1e9 * By_pos[t_idx1, :])
plt.scatter(peaks1, 1e9 * By_pos[t_idx1, peaks1])
plt.plot(1e9 * By_pos[t_idx2, :])
plt.scatter(peaks2, 1e9 * By_pos[t_idx2, peaks2])
return
def plot_spatially_averaged_fields(save=True, tmax=None):
'''
Field arrays are shaped like (time, space)
'''
plt.ioff()
fig, [[ax1, ax2], [ax3, ax4], [ax5, ax6]] = plt.subplots(figsize=(18, 10),
nrows=3,
ncols=2)
fig.subplots_adjust(wspace=0, hspace=0)
for ii in range(num_runs):
cf.initialize_simulation_variables(series_dir +
'run_{}/data/'.format(ii))
arr_dir = series_dir + 'run_{}/extracted/'.format(ii)
Bx_raw = 1e9 * (cf.get_array(arr_dir, component='Bx') - cf.B0)
By_raw = 1e9 * cf.get_array(arr_dir, component='By')
Bz_raw = 1e9 * cf.get_array(arr_dir, component='Bz')
lpad = 20
ax1.plot(cf.time_seconds_field,
abs(Bz_raw).mean(axis=1),
label='Run {}'.format(ii),
c=run_colors[ii])
ax3.plot(cf.time_seconds_field,
abs(By_raw).mean(axis=1),
label='Run {}'.format(ii),
c=run_colors[ii])
ax5.plot(cf.time_seconds_field,
abs(Bx_raw).mean(axis=1),
label='Run {}'.format(ii),
c=run_colors[ii])
ax2.plot(cf.time_seconds_field,
abs(Bz_raw).mean(axis=1),
label='Run {}'.format(ii),
c=run_colors[ii])
ax4.plot(cf.time_seconds_field,
abs(By_raw).mean(axis=1),
label='Run {}'.format(ii),
c=run_colors[ii])
ax6.plot(cf.time_seconds_field,
abs(Bx_raw).mean(axis=1),
label='Run {}'.format(ii),
c=run_colors[ii])
ax1.set_ylabel(r'$\overline{|\delta B_z|}$ (nT)',
rotation=0,
labelpad=lpad)
ax3.set_ylabel(r'$\overline{|\delta B_y|}$ (nT)',
rotation=0,
labelpad=lpad)
ax5.set_ylabel(r'$\overline{|\delta B_x|}$ (nT)',
rotation=0,
labelpad=lpad)
ax1.legend()
for ax in [ax1, ax2, ax3, ax4]:
ax.set_xticklabels([])
for ax in [ax2, ax4, ax6]:
ax.set_xlim(0, cf.time_seconds_field[-1])
ax.set_ylim(0, None)
ax.set_yticklabels([])
for ax in [ax1, ax3, ax5]:
if tmax is None:
ax.set_xlim(0, cf.time_seconds_field[-1] / 5)
else:
ax.set_xlim(0, tmax)
ax.set_ylim(0, None)
for ax in [ax5, ax6]:
ax.set_xlabel(r'Time (s)')
ax1.set_title('Spatially averaged fields'.format(cf.method_type))
if save == True and ii == (num_runs - 1):
fig.savefig(series_dir + 'sp_av_fields.png',
facecolor=fig.get_facecolor(),
edgecolor='none')
print('Spatially averaged B-field plot saved')
return
def single_point_field_timeseries(cells=None,
overwrite=False,
save=True,
tmax=None):
'''
Plot timeseries for raw fields at specified cells
maxtime=time in seconds for endpoint (defaults to total runtime)
'''
print('Plotting single-point fields...')
if cells is None:
cells = np.arange(cf.NX)
ts_folder_B = cf.anal_dir + '//single_point_fields//magnetic//'
ts_folder_E = cf.anal_dir + '//single_point_fields//electric//'
if os.path.exists(ts_folder_B) == False:
os.makedirs(ts_folder_B)
if os.path.exists(ts_folder_E) == False:
os.makedirs(ts_folder_E)
bx, by, bz, ex, ey, ez, vex, vey, vez, te, jx, jy, jz, qdens = cf.get_array(
get_all=True)
plt.ioff()
for x_idx in cells:
print('Cell {}...'.format(x_idx))
figB = plt.figure(figsize=(18, 10))
######################
### MAGNETIC FIELD ### Could loop this but I'm lazy
######################
figB = plt.figure(figsize=(18, 10))
## FIELDS: One period ##
axbx = plt.subplot2grid((3, 2), (0, 0))
axby = plt.subplot2grid((3, 2), (1, 0))
axbz = plt.subplot2grid((3, 2), (2, 0))
axbx.plot(cf.time_seconds_field, 1e9 * bx[:, x_idx])
axbx.set_ylabel('$B_x (nT)$')
axby.plot(cf.time_seconds_field, 1e9 * by[:, x_idx])
axby.set_ylabel('$B_y (nT)$')
axbz.plot(cf.time_seconds_field, 1e9 * bz[:, x_idx])
axbz.set_ylabel('$B_z (nT)$')
axbz.set_xlabel('Time (s)')
## FIELDS: Full time ##
axbx_full = plt.subplot2grid((3, 2), (0, 1))
axby_full = plt.subplot2grid((3, 2), (1, 1))
axbz_full = plt.subplot2grid((3, 2), (2, 1))
axbx_full.set_title('B-field at cell {}: Total time'.format(x_idx))
axbx_full.plot(cf.time_seconds_field, 1e9 * bx[:, x_idx])
axby_full.plot(cf.time_seconds_field, 1e9 * by[:, x_idx])
axbz_full.plot(cf.time_seconds_field, 1e9 * bz[:, x_idx])
axbz_full.set_xlabel('Time (s)')
if tmax is None:
# Set it at 20% full runtime, just to get a bit better resolution
tmax = cf.time_seconds_field[-1] / 5
axbx.set_title('B-field at cell {}: 1/5 total time'.format(x_idx))
else:
axbx.set_title('B-field at cell {}: One period'.format(x_idx))
for ax in [axbx, axby, axbz]:
ax.set_xlim(0, tmax)
for ax in [axbx_full, axby_full, axbz_full]:
ax.set_xlim(0, cf.time_seconds_field[-1])
ax.set_yticklabels([])
for ax in [axbx, axby, axbx_full, axby_full]:
ax.set_xticklabels([])
axbx.set_ylim(axbx_full.get_ylim())
axby.set_ylim(axby_full.get_ylim())
axbz.set_ylim(axbz_full.get_ylim())
figB.tight_layout()
figB.subplots_adjust(hspace=0, wspace=0.02)
if save == True:
figB.savefig(ts_folder_B +
'single_point_Bfield_{}.png'.format(x_idx),
edgecolor='none')
######################
### ELECTRIC FIELD ###
######################
figE = plt.figure(figsize=(18, 10))
## FIELDS: One period ##
axex = plt.subplot2grid((3, 2), (0, 0))
axey = plt.subplot2grid((3, 2), (1, 0))
axez = plt.subplot2grid((3, 2), (2, 0))
axex.plot(cf.time_seconds_field, 1e3 * ex[:, x_idx])
axex.set_ylabel('$E_x (mV/m)$')
axey.plot(cf.time_seconds_field, 1e3 * ey[:, x_idx])
axey.set_ylabel('$E_y (mV/m)$')
axez.plot(cf.time_seconds_field, 1e3 * ez[:, x_idx])
axez.set_ylabel('$E_z (mV/m)$')
axez.set_xlabel('Time (s)')
## FIELDS: Full time ##
axex_full = plt.subplot2grid((3, 2), (0, 1))
axey_full = plt.subplot2grid((3, 2), (1, 1))
axez_full = plt.subplot2grid((3, 2), (2, 1))
axex_full.set_title('E-field at cell {}: Total time'.format(x_idx))
axex_full.plot(cf.time_seconds_field, 1e3 * ex[:, x_idx])
axey_full.plot(cf.time_seconds_field, 1e3 * ey[:, x_idx])
axez_full.plot(cf.time_seconds_field, 1e3 * ez[:, x_idx])
axez_full.set_xlabel('Time (s)')
if tmax is None:
# Set it at 20% full runtime, just to get a bit better resolution
tmax = cf.time_seconds_field[-1] / 5
axbx.set_title('E-field at cell {}: 1/5 total time'.format(x_idx))
else:
axbx.set_title('E-field at cell {}: One period'.format(x_idx))
for ax in [axex, axey, axez]:
ax.set_xlim(0, tmax)
for ax in [axex_full, axey_full, axez_full]:
ax.set_xlim(0, cf.time_seconds_field[-1])
ax.set_yticklabels([])
for ax in [axex, axey, axex_full, axey_full]:
ax.set_xticklabels([])
axex.set_ylim(axex_full.get_ylim())
axey.set_ylim(axey_full.get_ylim())
axez.set_ylim(axez_full.get_ylim())
figE.tight_layout()
figE.subplots_adjust(hspace=0, wspace=0.02)
if save == True:
figE.savefig(ts_folder_E +
'single_point_Efield_{}.png'.format(x_idx),
edgecolor='none')
plt.close('all')
return
def single_point_helicity_timeseries(cells=None, overwrite=False, save=True):
'''
Plot timeseries for raw, +ve, -ve helicities at single point
Maybe do phases here too? (Although is that trivial for the helical components
since they're separated based on a phase relationship between By,z ?)
'''
if cells is None:
cells = np.arange(cf.NX)
ts_folder = cf.anal_dir + '//single_point_helicity//'
if os.path.exists(ts_folder) == False:
os.makedirs(ts_folder)
By_raw = cf.get_array('By')
Bz_raw = cf.get_array('Bz')
Bt_pos, Bt_neg = bk.get_helical_components(overwrite)
By_pos = Bt_pos.real
By_neg = Bt_neg.real
Bz_pos = Bt_pos.imag
Bz_neg = Bt_neg.imag
plt.ioff()
for raw, pos, neg, component in zip([By_raw, Bz_raw], [By_pos, Bz_pos],
[By_neg, Bz_neg], ['y', 'z']):
for x_idx in cells:
fig = plt.figure(figsize=(18, 10))
ax1 = plt.subplot2grid((2, 2), (0, 0), colspan=2)
ax2 = plt.subplot2grid((2, 2), (1, 0), colspan=2)
ax1.plot(cf.time_seconds_field,
1e9 * raw[:, x_idx],
label='Raw B{}'.format(component),
c='blue')
ax2.plot(cf.time_seconds_field,
1e9 * pos[:, x_idx],
label='B{}+'.format(component),
c='green')
ax2.plot(cf.time_seconds_field,
1e9 * neg[:, x_idx],
label='B{}-'.format(component),
c='orange')
ax1.set_title('Time-series at cell {}'.format(x_idx))
ax2.set_xlabel('Time (s)')
for ax in [ax1, ax2]:
ax.set_ylabel('B{} (nT)'.format(component))
ax.set_xlim(0, cf.time_seconds_field[-1])
ax.legend()
fig.tight_layout()
fig.subplots_adjust(hspace=0)
ax1.set_xticklabels([])
if save == True:
fig.savefig(
ts_folder +
'single_point_field_B{}_{}.png'.format(component, x_idx),
edgecolor='none')
plt.close('all')
return
def AGU_plot_spatially_averaged_fields(save=True, tmax=None):
'''
Field arrays are shaped like (time, space)
'''
tick_label_size = 14
mpl.rcParams['xtick.labelsize'] = tick_label_size
fontsize = 18
plt.ioff()
fig, [ax1, ax2] = plt.subplots(2, figsize=(13, 6))
fig.subplots_adjust(wspace=0, hspace=0)
for ii, lbl in zip(range(num_runs),
['High Growth Case', 'Low Growth Case']):
cf.initialize_simulation_variables(series_dir +
'run_{}/data/'.format(ii))
arr_dir = series_dir + 'run_{}/extracted/'.format(ii)
By_raw = 1e9 * cf.get_array(arr_dir, component='By')
Bz_raw = 1e9 * cf.get_array(arr_dir, component='Bz')
lpad = 24
ax1.plot(cf.time_seconds_field,
abs(By_raw).mean(axis=1),
label=lbl,
c=run_colors[ii])
ax2.plot(cf.time_seconds_field,
abs(Bz_raw).mean(axis=1),
label=lbl,
c=run_colors[ii])
ax1.set_ylabel('$\overline{|\delta B_y|}$\n (nT)',
rotation=0,
labelpad=lpad,
fontsize=fontsize)
ax2.set_ylabel('$\overline{|\delta B_z|}$\n (nT)',
rotation=0,
labelpad=lpad,
fontsize=fontsize)
ax1.legend(loc='lower right', prop={'size': fontsize})
for ax in [ax1]:
ax.set_xticklabels([])
for ax in [ax1, ax2]:
ax.set_xlim(0, cf.time_seconds_field[-1])
ax.set_ylim(0, None)
ax1.set_title(
'Spatially Averaged Fields :: Low/High Growth Parameters',
fontsize=fontsize + 4)
ax2.set_xlabel(r'Time (s)', fontsize=fontsize)
if save == True and ii == (num_runs - 1):
fig.savefig(series_dir + 'AGU_sp_av_fields.png',
facecolor=fig.get_facecolor(),
edgecolor='none',
bbox_inches='tight')
print('Spatially averaged B-field plot saved')
return