-
Notifications
You must be signed in to change notification settings - Fork 11
/
Copy pathStokesVE.jl
403 lines (365 loc) · 15.4 KB
/
StokesVE.jl
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
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
struct BoundaryConditions
# The intent here is that each boundary gets a flag
# 0 = Free-slip
# 1 = No-slip
# other possibilities?
top::Int64
bottom::Int64
left::Int64
right::Int64
end
@inline node_index(i::Int64, j::Int64, ny::Int64) = ny * (j - 1) + i
@inline vxdof(i::Int64, j::Int64, ny::Int64) = 3 * (node_index(i, j, ny) - 1) + 1
@inline vydof(i::Int64, j::Int64, ny::Int64) = 3 * (node_index(i, j, ny) - 1) + 2
@inline pdof(i::Int64, j::Int64, ny::Int64) = 3 * (node_index(i, j, ny) - 1) + 3
function form_stokes(grid::CartesianGrid,eta_s::Matrix,eta_n::Matrix,mu_s::Matrix,mu_n::Matrix,sxx_o::Matrix,sxy_o::Matrix,rhoX::Matrix,rhoY::Matrix,bc::BoundaryConditions,gx::Float64,gy::Float64;dt::Float64=0.0)
# Form the viscou-elastic Stokes system.
# Inputs:
# grid - the cartesian grid
# eta_s - viscosity at the basic nodes
# eta_n - viscosity at the cell centers
# mu_s - shear modulus at the basic nodes
# mu_n - shear modulus at the cell centers
# sxx_o,sxy_o - are the old deviatoric stresses
# rhoX - density at the vx nodes
# rhoY - density at the vy nodes
# bc - a vector describing the boundary conditions along the [left,right,top,bottom]
# gx,gy - gravitational body force in the x and y direction
# dt - the timestep, used in the free surface stabilization terms. dt=0.0 (default)
# disables free surface stabilization.
# Outputs:
# L,R - the left hand side (matrix) and right hand side (vector) of the stokes system
k::Int64 = 1 # index into dof arrays
nx = grid.nx
ny = grid.ny
nn = nx*ny
nnz = 2*11*nn + 5*nn # total number of nonzeros in matrix (not including BCs)
row_index = zeros(Int64,nnz) # up to 5 nonzeros per row
col_index = zeros(Int64,nnz)
value = zeros(Float64, nnz)
dx = grid.W/(grid.nx-1)
dy = grid.H/(grid.ny-1)
kcont = 2*eta_s[1,1]/(dx+dy)# scaling factor for continuity equation
kcont = 1e20/(dx+dy)*2
kbond = 1.# scaling factor for dirichlet bc equations.
R=zeros(3*nn,1)
# Viscoelasticity factor - for convenience
Zn = mu_n.*dt ./ (mu_n.*dt .+ eta_n)
Zs = mu_s.*dt ./ (mu_s.*dt .+ eta_s)
# loop over j
for j in 1:nx
# loop over i
for i in 1:ny
# dxp is the dx in the +x direction, dxm is dx in the -x direction, dxc is the spacing between cell centers
dxp = j<nx ? grid.x[j+1] - grid.x[j] : grid.x[j] - grid.x[j-1]
dxm = j>1 ? grid.x[j] - grid.x[j-1] : grid.x[j+1] - grid.x[j]
dxc = 0.5*(dxp+dxm)
# dyp and dym are spacing between vx nodes in the +y and -y directions
dyp = i<ny ? grid.yc[i+1]- grid.yc[i] : grid.yc[i] -grid.yc[i-1]
dym = i>1 ? grid.yc[i] - grid.yc[i-1] : grid.yc[i+1] -grid.yc[i]
dyc = 0.5*(dyp+dym)
# discretize the x-stokes - note that numbering in comments refers to Gerya Figure 7.18a
# and equation 7.22
this_row = vxdof(i,j,ny)
# Boundary cases first...
if j==1 || j == nx # left boundary or right boundary
# vx = 0
row_index[k] = this_row
col_index[k] = this_row
value[k] = kbond
k+=1
R[this_row] = 0.0 *kbond
elseif i==1
# dvx/dy = 0 (free slip)
row_index[k] = this_row
col_index[k] = this_row
value[k] = -kbond
k+=1
row_index[k] = this_row
col_index[k] = vxdof(i+1,j,ny)
value[k] = kbond
k+=1
R[this_row] = 0.0*kbond
else
# add free surface stabilization
drhodx = (rhoX[i,j+1]-rhoX[i,j-1])/2/dxc
drhody = (rhoX[i+1,j]-rhoX[i-1,j])/2/dyc
# Z terms
# Z_n = dt*mu_n[i,j] /(dt*mu_n[i,j] +eta_n[i,j])
# Z_nf = dt*mu_n[i,j+1]/(dt*mu_n[i,j+1]+eta_n[i,j+1])
# Z_s = dt*mu_s[i,j] /(dt*mu_s[i,j] +eta_s[i,j])
# Z_sb = dt*mu_s[i-1,j]/(dt*mu_s[i-1,j]+eta_s[i-1,j])
# vx1 term
row_index[k] = this_row
col_index[k] = vxdof(i,j-1,ny)
value[k] = 2*eta_n[i,j]*Zn[i,j]/dxm/dxc
k += 1
# vx2 term
row_index[k] = this_row
col_index[k] = vxdof(i-1,j,ny)
value[k] = eta_s[i-1,j]*Zs[i-1,j]/dym/dyc
k+=1
# vx3 term
row_index[k] = this_row
col_index[k] = this_row
# possible mistake in i+1,j term?
value[k] = -(eta_s[i-1,j]*Zs[i-1,j]/dyp + eta_s[i,j]*Zs[i,j]/dym)/dyc - 2*(eta_n[i,j+1]*Zn[i,j+1]/dxp + eta_n[i,j]*Zn[i,j]/dxm)/dxc - drhodx*gx*dt
# k += 1
if i == ny #vx4
# if i == nx, dvx/dy = 0 -> vx3 == vx4 (see Gerya fig 7.18a)
value[k] += eta_s[i,j]*Zs[i,j]/dyp/dyc
k+=1
else
k+=1
# vx4 term
row_index[k] = this_row
col_index[k] = vxdof(i+1,j,ny)
value[k] = eta_s[i,j]*Zs[i,j]/dyp/dyc
k+=1
end
# vx5 term
row_index[k] = this_row
col_index[k] = vxdof(i,j+1,ny)
value[k] = 2*eta_n[i,j+1]*Zn[i,j+1]/dxp/dxc
k+=1
# vy1 term
row_index[k] = this_row
col_index[k] = vydof(i-1,j,ny)
value[k] = eta_s[i-1,j]*Zs[i-1,j]/dxm/dyc - drhody*gx*dt/4
k+=1
# vy2 term
row_index[k] = this_row
col_index[k] = vydof(i,j,ny)
value[k] = -eta_s[i,j]*Zs[i,j]/dxc/dyc - drhody*gx*dt/4
k+=1
# vy3 term
row_index[k] = this_row
col_index[k] = vydof(i-1,j+1,ny)
value[k] = -eta_s[i-1,j]*Zs[i-1,j]/dxc/dyc - drhody*gx*dt/4
k+=1
# vy4 term
row_index[k] = this_row
col_index[k] = vydof(i,j+1,ny)
value[k] = eta_s[i,j]*Zs[i,j]/dxc/dyc - drhody*gx*dt/4
k+=1
# P1 term
row_index[k] = this_row
col_index[k] = pdof(i,j,ny)
value[k] = kcont/dxc
k+=1
# P2 term
row_index[k] = this_row
col_index[k] = pdof(i,j+1,ny)
value[k] = -kcont/dxc
k+=1
# Right-hand side
Sxy1 = sxy_o[i-1,j]*(1-Zs[i-1,j])#)eta_s[i-1,j]/(mu_s[i-1,j]*dt+eta_s[i-1,j])
Sxy2 = sxy_o[i,j]*(1-Zs[i,j])#)eta_s[i,j]/(mu_s[i,j]*dt+eta_s[i,j])
Sxx1 = sxx_o[i,j]*(1-Zn[i,j])#eta_n[i,j]/(mu_n[i,j]*dt+eta_n[i,j])
Sxx2 = sxx_o[i,j+1]*(1-Zn[i,j+1])#eta_n[i,j+1]/(mu_n[i,j+1]*dt+eta_n[i,j+1])
R[this_row] = -gx*rhoX[i,j] - 2*(Sxx2-Sxx1)/(grid.x[j+1]-grid.x[j-1]) - (Sxy2-Sxy1)/dym
end
# END X-STOKES
# BEGIN Y-STOKES
dxp = j < nx ? grid.xc[j+1] - grid.xc[j] : grid.xc[j] -grid.xc[j-1]
dxm = j > 1 ? grid.xc[j] - grid.xc[j-1] : grid.xc[j+1]-grid.xc[j]
dxc = j > 1 ? grid.x[j] - grid.x[j-1] : grid.x[j+1] - grid.x[j]
dyp = i < ny ? grid.y[i+1] - grid.y[i] : grid.y[i] - grid.y[i-1]
dym = i > 1 ? grid.y[i] - grid.y[i-1] : grid.y[i+1] - grid.y[i]
dyc = i < ny ? grid.yc[i+1] - grid.yc[i] : grid.yc[i] - grid.yc[i-1]
this_row = vydof(i,j,ny)
if i==1 || i == ny
# top row / bottom row
row_index[k] = this_row
col_index[k] = this_row
value[k] = kbond
k+=1
R[this_row] = 0.0*kbond
elseif j==1
# left boundary - NO slip
row_index[k] = this_row
col_index[k] = this_row
value[k] = kbond
k+=1
row_index[k] = this_row
col_index[k] = vydof(i,j+1,ny)
value[k] = kbond
k+=1
R[this_row] = 0.0*kbond
else
# add free surface stabilization
drhodx = (rhoY[i,j+1]-rhoY[i,j-1])/2/dxc
drhody = (rhoY[i+1,j]-rhoY[i-1,j])/2/dyc
# visco-elastic coefficient
# Z_n1 = mu_n[i,j]*dt/(mu_n[i,j]*dt + eta_n[i,j]) # P1
# Z_n2 = mu_n[i+1,j]*dt/(mu_n[i+1,j]*dt + eta_n[i+1,j]) # P1
# Z_s1 = mu_s[i,j-1]*dt/(mu_s[i,j-1]*dt + eta_s[i,j-1]) # S1
# Z_s2 = mu_s[i,j]*dt/(mu_s[i,j]*dt + eta_s[i,j]) # S2
# used LHS assume 2d incompressible flow
# use the comment out LHS if consistent with book assumption
#vy1
row_index[k] = this_row
col_index[k] = vydof(i,j-1,ny)
value[k] = eta_s[i,j-1]*Zs[i,j-1]/dxm/dxc
k+=1
#vy2
row_index[k] = this_row
col_index[k] = vydof(i-1,j,ny)
value[k] = 2*eta_n[i,j]*Zn[i,j]/dym/dyc
# value[k] = eta_n[i,j]*Z_s2/dym/dyc
k+=1
#vy3
row_index[k] = this_row
col_index[k] = this_row
value[k] = -2*eta_n[i+1,j]*Zn[i+1,j]/dyp/dyc -2*eta_n[i,j]*Zn[i,j]/dym/dyc - eta_s[i,j]*Zs[i,j]/dxp/dxc - eta_s[i,j-1]*Zs[i,j-1]/dxm/dxc - drhody*gy*dt
# value[k] = -eta_n[i+1,j]*Z_n2/dyp/dyc -eta_n[i,j]*Z_n1/dym/dyc - eta_s[i,j]*Z_s2/dxp/dxc - eta_s[i,j-1]*Z_s1/dxm/dxc - drhody*gy*dt
if j == nx
# free slip - vx5 = vx3.
###### CHECK! ######
value[k] += eta_s[i,j]*Zs[i,j]/dxp/dxc
end
k+=1
#vy4
row_index[k] = this_row
col_index[k] = vydof(i+1,j,ny)
value[k] = 2*eta_n[i+1,j]*Zn[i+1,j]/dyp/dyc
# value[k] = eta_n[i+1,j]*Z_n2/dyp/dyc
k+=1
#vy5
if j<nx
row_index[k] = this_row
col_index[k] = vydof(i,j+1,ny)
value[k] = eta_s[i,j]*Zs[i,j]/dxp/dxc
k+=1
end
#vx1
row_index[k] = this_row
col_index[k] = vxdof(i,j-1,ny)
value[k] = eta_s[i,j-1]*Zs[i,j-1]/dxc/dyc - drhodx*gy*dt/4 # - eta_n[i,j]*Z_n1/dxc/dyc
k+=1
#vx2
row_index[k] = this_row
col_index[k] = vxdof(i+1,j-1,ny)
value[k] = -eta_s[i,j-1]*Zs[i,j-1]/dxc/dyc - drhodx*gy*dt/4 # + eta_n[i+1,j]*Z_n2/dxc/dyc
k+=1
#vx3
row_index[k] = this_row
col_index[k] = vxdof(i,j,ny)
value[k] = -eta_s[i,j]*Zs[i,j]/dxc/dyc -drhodx*gy*dt/4 # + eta_n[i,j]*Z_n1/dxc/dyc
k+=1
#vx4
row_index[k] = this_row
col_index[k] = vxdof(i+1,j,ny)
value[k] = eta_s[i,j]*Zs[i,j]/dxc/dyc - drhodx*gy*dt/4 # - eta_n[i+1,j]*Z_n2/dxc/dyc
k+=1
#P1
row_index[k] = this_row
col_index[k] = pdof(i,j,ny)
value[k] = kcont/dyc
k+=1
#P2
row_index[k] = this_row
col_index[k] = pdof(i+1,j,ny)
value[k] = -kcont/dyc
k+=1
# get old stress
# syy_o = -sxx_o
# this should be -rho*gy + xelvis*dsyy0/dy + xelvis*dsxy/dx
R[this_row] = -gy*rhoY[i,j] + (sxx_o[i+1,j]*(1-Zn[i+1,j])-sxx_o[i,j]*(1-Zn[i,j]))/dyc - (sxy_o[i,j]*(1-Zs[i,j])-sxy_o[i,j-1]*(1-Zs[i,j-1]))/dxc
end
# END Y-STOKES
# discretize the continuity equation
# dvx/dx + dvy/dy = 0
this_row = pdof(i,j,ny)
if i==1 || j == 1 || (i==2 && j == 2)
row_index[k] = this_row
col_index[k] = this_row
value[k] = kbond
k+=1
R[this_row] = 0.0
else
dxm = grid.x[j] - grid.x[j-1]
dym = grid.y[i] - grid.y[i-1]
row_index[k] = this_row
col_index[k] = vxdof(i,j,ny)
value[k] = kcont/dxm
k+=1
row_index[k] = this_row
col_index[k] = vxdof(i,j-1,ny)
value[k] = -kcont/dxm
k+=1
row_index[k] = this_row
col_index[k] = vydof(i,j,ny)
value[k] = kcont/dym
k+=1
row_index[k] = this_row
col_index[k] = vydof(i-1,j,ny)
value[k] = -kcont/dym
k+=1
row_index[k] = this_row
col_index[k] = this_row
value[k] = 0.0
k+=1
R[this_row] = 0.0
end
# END CONTINUITY
end
end
@views row_index = row_index[1:(k-1)]
@views col_index = col_index[1:(k-1)]
@views value = value[1:(k-1)]
L = sparse(row_index,col_index,value)
return L,R
end
function unpack(solution, grid::CartesianGrid; ghost::Bool=false)
if ghost
nx1 = grid.nx+1
ny1 = grid.ny+1
P = zeros(Float64,(ny1,nx1))
vx = zeros(Float64,(ny1,nx1))
vy = zeros(Float64,(ny1,nx1))
ny = grid.ny
for j in 1:grid.nx
for i in 1:grid.ny
vx[i,j] = solution[vxdof(i,j,grid.ny)]
vy[i,j] = solution[vydof(i,j,grid.ny)]
P[i,j] = solution[pdof(i,j,grid.ny)]
end
end
# right boundary
j=nx1
for i in 1:grid.ny
vx[i,j] = 0.0
vy[i,j] = vy[i,j-1];# free slip
end
i=ny1
for j in 1:grid.nx
vx[i,j] = vx[i-1,j];# free-slip along bottom
vy[i,j] = 0.0
end
else
P = zeros(Float64,(grid.ny,grid.nx))
vx = zeros(Float64,(grid.ny,grid.nx))
vy = zeros(Float64,(grid.ny,grid.nx))
ny = grid.ny
for j in 1:grid.nx
for i in 1:grid.ny
vx[i,j] = solution[vxdof(i,j,grid.ny)]
vy[i,j] = solution[vydof(i,j,grid.ny)]
P[i,j] = solution[pdof(i,j,grid.ny)]
end
end
end
return vx,vy,P
end
function compute_timestep(grid::CartesianGrid,vxc::Matrix,vyc::Matrix;dtmax::Float64=Inf,cfl::Float64=0.5)
# compute the maximum timestep based on cell-centered velocities in vxc and vyc and the cfl number.
for i in 2:grid.ny
for j in 2:grid.nx
dx = grid.x[j]-grid.x[j-1]
dy = grid.y[i]-grid.y[i-1]
dtmax = min( dtmax , cfl*dx/abs(vxc[i,j]) , cfl*dy/abs(vyc[i,j]) )
end
end
return dtmax
end