timprove plots - sphere - GPU-based 3D discrete element method algorithm with optional fluid coupling
(HTM) git clone git://src.adamsgaard.dk/sphere
(DIR) Log
(DIR) Files
(DIR) Refs
(DIR) LICENSE
---
(DIR) commit 59b4b66f2541d1d53b9cfb08c9b73ccf3bed96e7
(DIR) parent 09657d844261b3e1730f2fb2132d4dd66dba5e5f
(HTM) Author: Anders Damsgaard <anders.damsgaard@geo.au.dk>
Date: Tue, 7 Oct 2014 15:53:25 +0200
improve plots
Diffstat:
M python/shear-results-internals.py | 65 ++++++++++++++++++++-----------
M python/shear-results.py | 2 +-
2 files changed, 44 insertions(+), 23 deletions(-)
---
(DIR) diff --git a/python/shear-results-internals.py b/python/shear-results-internals.py
t@@ -100,9 +100,15 @@ for step_str in steps:
'''
f_pf[s,:] += sim.f_sum[:,2]
- dev_p[s,:] += \
- numpy.average(numpy.average(sim.p_f, axis=0), axis=0)\
- /nsteps_avg
+ dz = sim.L[2]/sim.num[2]
+ wall0_iz = int(sim.w_x[0]/dz)
+ for z in numpy.arange(0, wall0_iz+1):
+ dev_p[s,z] += \
+ (numpy.average(sim.p_f[:,:,z]) -
+ ((wall0_iz*dz - zpos_c[s][z] + 0.5*dz)
+ *sim.rho_f*numpy.abs(sim.g[2])\
+ + sim.p_f[0,0,-1])) \
+ /nsteps_avg
v_z_f[s,:] += sim.v_f[:,:,:,2]/nsteps_avg
t@@ -137,7 +143,8 @@ for step_str in steps:
#fig = plt.figure(figsize=(8,4*(len(steps))+1))
#fig = plt.figure(figsize=(8,5*(len(steps))+1))
-fig = plt.figure(figsize=(16,5*(len(steps))+1))
+#fig = plt.figure(figsize=(16,5*(len(steps))+1))
+fig = plt.figure(figsize=(20,5*(len(steps))+1))
ax = []
for s in numpy.arange(len(steps)):
t@@ -182,17 +189,19 @@ for s in numpy.arange(len(steps)):
#ax[s*4+3].plot(dphi_bar[s,1:], zpos_c[s,1:], '-k', linewidth=3)
#ax[s*4+3].plot(dphi_bar[s,1:], zpos_c[s,1:], '-w', linewidth=2)
- #ax[s*n+3].plot(v_z_p[s]*100.0, zpos_p[s], ',', color = '#888888')
+ ax[s*n+3].plot(v_z_p[s]*100.0, zpos_p[s], ',', color = '#888888')
ax[s*n+3].plot(v_z_p_bar[s]*100.0, zpos_c[s], color = 'k')
#ax[s*n+0].plot([0.0,0.0], [0.0, sim.L[2]], '--', color='k')
# hydrostatic pressure distribution
ax[s*n+4].plot(dev_p[s]/1000.0, zpos_c[s], 'k')
- y_top = sim.w_x[0]
- x_top = sim.p_f[0,0,-1]
- y_bot = 0.0
- x_bot = x_top + (y_top - y_bot)*sim.rho*numpy.abs(sim.g[2])
- ax[s*n+4].plot([x_top/1000.0, x_bot/1000.0], [y_top, y_bot], '--', color='k')
+ #dz = sim.L[2]/sim.num[2]
+ #wall0_iz = int(sim.w_x[0]/dz)
+ #y_top = wall0_iz*dz + 0.5*dz
+ #x_top = sim.p_f[0,0,-1]
+ #y_bot = 0.0
+ #x_bot = x_top + (wall0_iz*dz - zpos_c[s][0] + 0.5*dz)*sim.rho_f*numpy.abs(sim.g[2])
+ #ax[s*n+4].plot([x_top/1000.0, x_bot/1000.0], [y_top, y_bot], '--', color='k')
#ax[s*n+1].set_title(strain_str)
#ax[s*n+1].set_title(' ')
t@@ -204,7 +213,7 @@ for s in numpy.arange(len(steps)):
f_pf_mean_nonzero = f_pf_mean[s][I]
zpos_c_nonzero = zpos_c[s][I]
- #ax[s*n+5].plot(f_pf_nonzero, zpos_p_nonzero, ',', color = '#888888')
+ ax[s*n+5].plot(f_pf_nonzero, zpos_p_nonzero, ',', color = '#888888')
#ax[s*4+1].plot(f_pf_mean[s][1:-2], zpos_c[s][1:-2], color = 'k')
ax[s*n+5].plot(f_pf_mean_nonzero, zpos_c_nonzero, color = 'k')
#ax[s*4+1].plot([0.0, 0.0], [0.0, sim.L[2]], '--', color='k')
t@@ -236,7 +245,7 @@ for s in numpy.arange(len(steps)):
#ax[s*n+0].set_xlim([-0.01,0.01])
#ax[s*n+0].set_xlim([-0.005,0.005])
#ax[s*n+0].set_xlim([-0.25,0.75])
- ax[s*n+4].set_xlim([595,625]) # p_f
+ #ax[s*n+4].set_xlim([595,625]) # p_f
#ax[s*n+2].set_xlim([-0.0005,0.0005])
#ax[s*n+2].set_xlim([-0.08,0.08])
t@@ -256,10 +265,21 @@ for s in numpy.arange(len(steps)):
ax[s*n+0].set_xlabel('$\\boldsymbol{x}^x_\\text{p}$ [m]')
ax[s*n+1].set_xlabel('$\\bar{\\phi}$ [-] (solid)')
ax[s*n+2].set_xlabel('$\\delta \\bar{\\phi}/\\delta t$ [-] (dashed)')
- ax[s*n+3].set_xlabel('$\\boldsymbol{v}^z_\\text{p}$ [cms$^-1$]')
- ax[s*n+4].set_xlabel('$\\bar{p_\\text{f}}$ [kPa]')
+ ax[s*n+3].set_xlabel('$\\boldsymbol{v}^z_\\text{p}$ [cms$^{-1}$]')
+ #ax[s*n+4].set_xlabel('$\\bar{p_\\text{f}}$ [kPa]')
+ ax[s*n+4].set_xlabel('$\\bar{p_\\text{f}} - p_\\text{hyd}$ [kPa]')
ax[s*n+5].set_xlabel('$\\boldsymbol{f}^z_\\text{pf}$ [N]')
- ax[s*n+6].set_xlabel('$\\bar{\\boldsymbol{v}}^z_\\text{f}$ [cms$^-1$]')
+ ax[s*n+6].set_xlabel('$\\bar{\\boldsymbol{v}}^z_\\text{f}$ [cms$^{-1}$]')
+
+ # align x labels
+ labely = -0.3
+ ax[s*n+0].xaxis.set_label_coords(0.5, labely)
+ ax[s*n+1].xaxis.set_label_coords(0.5, labely)
+ #ax[s*n+2].xaxis.set_label_coords(0.5, labely)
+ ax[s*n+3].xaxis.set_label_coords(0.5, labely)
+ ax[s*n+4].xaxis.set_label_coords(0.5, labely)
+ ax[s*n+5].xaxis.set_label_coords(0.5, labely)
+ ax[s*n+6].xaxis.set_label_coords(0.5, labely)
plt.setp(ax[s*n+1].get_yticklabels(), visible=False)
plt.setp(ax[s*n+2].get_yticklabels(), visible=False)
t@@ -297,12 +317,13 @@ for s in numpy.arange(len(steps)):
#plt.text(0.05, 1.06, strain_str, horizontalalignment='left', fontsize=22,
#transform=ax[s*n+0].transAxes)
#ax[s*4+0].set_title(strain_str)
- ax[s*n+0].set_title('a')
- ax[s*n+1].set_title('b')
- ax[s*n+3].set_title('c')
- ax[s*n+4].set_title('d')
- ax[s*n+5].set_title('e')
- ax[s*n+6].set_title('f')
+
+ #ax[s*n+0].set_title('a')
+ #ax[s*n+1].set_title('b')
+ #ax[s*n+3].set_title('c')
+ #ax[s*n+4].set_title('d')
+ #ax[s*n+5].set_title('e')
+ #ax[s*n+6].set_title('f')
ax[s*n+0].grid()
ax[s*n+1].grid()
t@@ -318,7 +339,7 @@ for s in numpy.arange(len(steps)):
#fig.text(0.1, y, strain_str, horizontalalignment='left', fontsize=22)
#ax[s*4+0].annotate(strain_str, xytext=(0,1.1), textcoords='figure fraction',
#horizontalalignment='left', fontsize=22)
- plt.text(0.05, 1.06, strain_str, horizontalalignment='left', fontsize=22,
+ plt.text(-0.38, 1.15, strain_str, horizontalalignment='left', fontsize=22,
transform=ax[s*n+0].transAxes)
#plt.title(' ')
(DIR) diff --git a/python/shear-results.py b/python/shear-results.py
t@@ -87,7 +87,7 @@ def smooth(x, window_len=10, window='hanning'):
return y[window_len-1:-window_len+1]
-smooth_window = 11
+smooth_window = 40
shear_strain = [[], [], []]
friction = [[], [], []]