1 | !> @file pres.f90 |
---|
2 | !------------------------------------------------------------------------------! |
---|
3 | ! This file is part of the PALM model system. |
---|
4 | ! |
---|
5 | ! PALM is free software: you can redistribute it and/or modify it under the |
---|
6 | ! terms of the GNU General Public License as published by the Free Software |
---|
7 | ! Foundation, either version 3 of the License, or (at your option) any later |
---|
8 | ! version. |
---|
9 | ! |
---|
10 | ! PALM is distributed in the hope that it will be useful, but WITHOUT ANY |
---|
11 | ! WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR |
---|
12 | ! A PARTICULAR PURPOSE. See the GNU General Public License for more details. |
---|
13 | ! |
---|
14 | ! You should have received a copy of the GNU General Public License along with |
---|
15 | ! PALM. If not, see <http://www.gnu.org/licenses/>. |
---|
16 | ! |
---|
17 | ! Copyright 1997-2019 Leibniz Universitaet Hannover |
---|
18 | !------------------------------------------------------------------------------! |
---|
19 | ! |
---|
20 | ! Current revisions: |
---|
21 | ! ------------------ |
---|
22 | ! |
---|
23 | ! |
---|
24 | ! Former revisions: |
---|
25 | ! ----------------- |
---|
26 | ! $Id: pres.f90 4180 2019-08-21 14:37:54Z scharf $ |
---|
27 | ! variable child_domain_nvn eliminated |
---|
28 | ! |
---|
29 | ! 3849 2019-04-01 16:35:16Z knoop |
---|
30 | ! OpenACC port for SPEC |
---|
31 | ! |
---|
32 | ! |
---|
33 | ! Description: |
---|
34 | ! ------------ |
---|
35 | !> Compute the divergence of the provisional velocity field. Solve the Poisson |
---|
36 | !> equation for the perturbation pressure. Compute the final velocities using |
---|
37 | !> this perturbation pressure. Compute the remaining divergence. |
---|
38 | !------------------------------------------------------------------------------! |
---|
39 | SUBROUTINE pres |
---|
40 | |
---|
41 | |
---|
42 | USE arrays_3d, & |
---|
43 | ONLY: d, ddzu, ddzu_pres, ddzw, dzw, p, p_loc, rho_air, rho_air_zw, & |
---|
44 | tend, u, v, w |
---|
45 | |
---|
46 | USE control_parameters, & |
---|
47 | ONLY: bc_lr_cyc, bc_ns_cyc, bc_radiation_l, bc_radiation_n, & |
---|
48 | bc_radiation_r, bc_radiation_s, child_domain, & |
---|
49 | conserve_volume_flow, coupling_mode, & |
---|
50 | dt_3d, gathered_size, ibc_p_b, ibc_p_t, & |
---|
51 | intermediate_timestep_count, intermediate_timestep_count_max, & |
---|
52 | mg_switch_to_pe0_level, nesting_offline, & |
---|
53 | psolver, subdomain_size, & |
---|
54 | topography, volume_flow, volume_flow_area, volume_flow_initial |
---|
55 | |
---|
56 | USE cpulog, & |
---|
57 | ONLY: cpu_log, log_point, log_point_s |
---|
58 | |
---|
59 | USE grid_variables, & |
---|
60 | ONLY: ddx, ddy |
---|
61 | |
---|
62 | USE indices, & |
---|
63 | ONLY: nbgp, ngp_2dh_outer, nx, nxl, nxlg, nxl_mg, nxr, nxrg, nxr_mg, & |
---|
64 | ny, nys, nysg, nys_mg, nyn, nyng, nyn_mg, nzb, nzt, nzt_mg, & |
---|
65 | wall_flags_0 |
---|
66 | |
---|
67 | USE kinds |
---|
68 | |
---|
69 | USE pegrid |
---|
70 | |
---|
71 | USE pmc_interface, & |
---|
72 | ONLY: nesting_mode |
---|
73 | |
---|
74 | USE poisfft_mod, & |
---|
75 | ONLY: poisfft |
---|
76 | |
---|
77 | USE poismg_mod |
---|
78 | |
---|
79 | USE poismg_noopt_mod |
---|
80 | |
---|
81 | USE statistics, & |
---|
82 | ONLY: statistic_regions, sums_divnew_l, sums_divold_l, weight_pres, & |
---|
83 | weight_substep |
---|
84 | |
---|
85 | USE surface_mod, & |
---|
86 | ONLY : bc_h |
---|
87 | |
---|
88 | IMPLICIT NONE |
---|
89 | |
---|
90 | INTEGER(iwp) :: i !< |
---|
91 | INTEGER(iwp) :: j !< |
---|
92 | INTEGER(iwp) :: k !< |
---|
93 | INTEGER(iwp) :: m !< |
---|
94 | |
---|
95 | REAL(wp) :: ddt_3d !< |
---|
96 | REAL(wp) :: d_weight_pres !< |
---|
97 | REAL(wp) :: localsum !< |
---|
98 | REAL(wp) :: threadsum !< |
---|
99 | REAL(wp) :: weight_pres_l !< |
---|
100 | REAL(wp) :: weight_substep_l !< |
---|
101 | |
---|
102 | REAL(wp), DIMENSION(1:3) :: volume_flow_l !< |
---|
103 | REAL(wp), DIMENSION(1:3) :: volume_flow_offset !< |
---|
104 | REAL(wp), DIMENSION(1:nzt) :: w_l !< |
---|
105 | REAL(wp), DIMENSION(1:nzt) :: w_l_l !< |
---|
106 | |
---|
107 | |
---|
108 | CALL cpu_log( log_point(8), 'pres', 'start' ) |
---|
109 | |
---|
110 | ! |
---|
111 | !-- Calculate quantities to be used locally |
---|
112 | ddt_3d = 1.0_wp / dt_3d |
---|
113 | IF ( intermediate_timestep_count == 0 ) THEN |
---|
114 | ! |
---|
115 | !-- If pres is called before initial time step |
---|
116 | weight_pres_l = 1.0_wp |
---|
117 | d_weight_pres = 1.0_wp |
---|
118 | weight_substep_l = 1.0_wp |
---|
119 | ELSE |
---|
120 | weight_pres_l = weight_pres(intermediate_timestep_count) |
---|
121 | d_weight_pres = 1.0_wp / weight_pres(intermediate_timestep_count) |
---|
122 | weight_substep_l = weight_substep(intermediate_timestep_count) |
---|
123 | ENDIF |
---|
124 | |
---|
125 | ! |
---|
126 | !-- Multigrid method expects array d to have one ghost layer. |
---|
127 | !-- |
---|
128 | IF ( psolver(1:9) == 'multigrid' ) THEN |
---|
129 | |
---|
130 | DEALLOCATE( d ) |
---|
131 | ALLOCATE( d(nzb:nzt+1,nys-1:nyn+1,nxl-1:nxr+1) ) |
---|
132 | |
---|
133 | ! |
---|
134 | !-- Since p is later used to hold the weighted average of the substeps, it |
---|
135 | !-- cannot be used in the iterative solver. Therefore, its initial value is |
---|
136 | !-- stored on p_loc, which is then iteratively advanced in every substep. |
---|
137 | IF ( intermediate_timestep_count <= 1 ) THEN |
---|
138 | DO i = nxl-1, nxr+1 |
---|
139 | DO j = nys-1, nyn+1 |
---|
140 | DO k = nzb, nzt+1 |
---|
141 | p_loc(k,j,i) = p(k,j,i) |
---|
142 | ENDDO |
---|
143 | ENDDO |
---|
144 | ENDDO |
---|
145 | ENDIF |
---|
146 | |
---|
147 | ELSEIF ( psolver == 'sor' .AND. intermediate_timestep_count <= 1 ) THEN |
---|
148 | |
---|
149 | ! |
---|
150 | !-- Since p is later used to hold the weighted average of the substeps, it |
---|
151 | !-- cannot be used in the iterative solver. Therefore, its initial value is |
---|
152 | !-- stored on p_loc, which is then iteratively advanced in every substep. |
---|
153 | p_loc = p |
---|
154 | |
---|
155 | ENDIF |
---|
156 | |
---|
157 | ! |
---|
158 | !-- Conserve the volume flow at the outflow in case of non-cyclic lateral |
---|
159 | !-- boundary conditions |
---|
160 | !-- WARNING: so far, this conservation does not work at the left/south |
---|
161 | !-- boundary if the topography at the inflow differs from that at the |
---|
162 | !-- outflow! For this case, volume_flow_area needs adjustment! |
---|
163 | ! |
---|
164 | !-- Left/right |
---|
165 | IF ( conserve_volume_flow .AND. ( bc_radiation_l .OR. & |
---|
166 | bc_radiation_r ) ) THEN |
---|
167 | |
---|
168 | volume_flow(1) = 0.0_wp |
---|
169 | volume_flow_l(1) = 0.0_wp |
---|
170 | |
---|
171 | IF ( bc_radiation_l ) THEN |
---|
172 | i = 0 |
---|
173 | ELSEIF ( bc_radiation_r ) THEN |
---|
174 | i = nx+1 |
---|
175 | ENDIF |
---|
176 | |
---|
177 | DO j = nys, nyn |
---|
178 | ! |
---|
179 | !-- Sum up the volume flow through the south/north boundary |
---|
180 | DO k = nzb+1, nzt |
---|
181 | volume_flow_l(1) = volume_flow_l(1) + u(k,j,i) * dzw(k) & |
---|
182 | * MERGE( 1.0_wp, 0.0_wp, & |
---|
183 | BTEST( wall_flags_0(k,j,i), 1 ) & |
---|
184 | ) |
---|
185 | ENDDO |
---|
186 | ENDDO |
---|
187 | |
---|
188 | #if defined( __parallel ) |
---|
189 | IF ( collective_wait ) CALL MPI_BARRIER( comm1dy, ierr ) |
---|
190 | CALL MPI_ALLREDUCE( volume_flow_l(1), volume_flow(1), 1, MPI_REAL, & |
---|
191 | MPI_SUM, comm1dy, ierr ) |
---|
192 | #else |
---|
193 | volume_flow = volume_flow_l |
---|
194 | #endif |
---|
195 | volume_flow_offset(1) = ( volume_flow_initial(1) - volume_flow(1) ) & |
---|
196 | / volume_flow_area(1) |
---|
197 | |
---|
198 | DO j = nysg, nyng |
---|
199 | DO k = nzb+1, nzt |
---|
200 | u(k,j,i) = u(k,j,i) + volume_flow_offset(1) & |
---|
201 | * MERGE( 1.0_wp, 0.0_wp, & |
---|
202 | BTEST( wall_flags_0(k,j,i), 1 ) & |
---|
203 | ) |
---|
204 | ENDDO |
---|
205 | ENDDO |
---|
206 | |
---|
207 | ENDIF |
---|
208 | |
---|
209 | ! |
---|
210 | !-- South/north |
---|
211 | IF ( conserve_volume_flow .AND. ( bc_radiation_n .OR. bc_radiation_s ) ) THEN |
---|
212 | |
---|
213 | volume_flow(2) = 0.0_wp |
---|
214 | volume_flow_l(2) = 0.0_wp |
---|
215 | |
---|
216 | IF ( bc_radiation_s ) THEN |
---|
217 | j = 0 |
---|
218 | ELSEIF ( bc_radiation_n ) THEN |
---|
219 | j = ny+1 |
---|
220 | ENDIF |
---|
221 | |
---|
222 | DO i = nxl, nxr |
---|
223 | ! |
---|
224 | !-- Sum up the volume flow through the south/north boundary |
---|
225 | DO k = nzb+1, nzt |
---|
226 | volume_flow_l(2) = volume_flow_l(2) + v(k,j,i) * dzw(k) & |
---|
227 | * MERGE( 1.0_wp, 0.0_wp, & |
---|
228 | BTEST( wall_flags_0(k,j,i), 2 ) & |
---|
229 | ) |
---|
230 | ENDDO |
---|
231 | ENDDO |
---|
232 | |
---|
233 | #if defined( __parallel ) |
---|
234 | IF ( collective_wait ) CALL MPI_BARRIER( comm1dx, ierr ) |
---|
235 | CALL MPI_ALLREDUCE( volume_flow_l(2), volume_flow(2), 1, MPI_REAL, & |
---|
236 | MPI_SUM, comm1dx, ierr ) |
---|
237 | #else |
---|
238 | volume_flow = volume_flow_l |
---|
239 | #endif |
---|
240 | volume_flow_offset(2) = ( volume_flow_initial(2) - volume_flow(2) ) & |
---|
241 | / volume_flow_area(2) |
---|
242 | |
---|
243 | DO i = nxlg, nxrg |
---|
244 | DO k = nzb+1, nzt |
---|
245 | v(k,j,i) = v(k,j,i) + volume_flow_offset(2) & |
---|
246 | * MERGE( 1.0_wp, 0.0_wp, & |
---|
247 | BTEST( wall_flags_0(k,j,i), 2 ) & |
---|
248 | ) |
---|
249 | ENDDO |
---|
250 | ENDDO |
---|
251 | |
---|
252 | ENDIF |
---|
253 | |
---|
254 | ! |
---|
255 | !-- Remove mean vertical velocity in case that Neumann conditions are used both at bottom and top |
---|
256 | !-- boundary. With Neumann conditions at both vertical boundaries, the solver cannot remove |
---|
257 | !-- mean vertical velocities. They should be removed, because incompressibility requires that |
---|
258 | !-- the vertical gradient of vertical velocity is zero. Since w=0 at the solid surface, it must be |
---|
259 | !-- zero everywhere. |
---|
260 | !-- This must not be done in case of a 3d-nesting child domain, because a mean vertical velocity |
---|
261 | !-- can physically exist in such a domain. |
---|
262 | !-- Also in case of offline nesting, mean vertical velocities may exist (and must not be removed), |
---|
263 | !-- caused by horizontal divergence/convergence of the large scale flow that is prescribed at the |
---|
264 | !-- side boundaries. |
---|
265 | !-- The removal cannot be done before the first initial time step because ngp_2dh_outer |
---|
266 | !-- is not yet known then. |
---|
267 | IF ( ibc_p_b == 1 .AND. ibc_p_t == 1 .AND. .NOT. nesting_offline & |
---|
268 | .AND. .NOT. ( child_domain .AND. nesting_mode /= 'vertical' ) & |
---|
269 | .AND. intermediate_timestep_count /= 0 ) & |
---|
270 | THEN |
---|
271 | w_l = 0.0_wp; w_l_l = 0.0_wp |
---|
272 | DO i = nxl, nxr |
---|
273 | DO j = nys, nyn |
---|
274 | DO k = nzb+1, nzt |
---|
275 | w_l_l(k) = w_l_l(k) + w(k,j,i) & |
---|
276 | * MERGE( 1.0_wp, 0.0_wp, BTEST( wall_flags_0(k,j,i), 3 ) ) |
---|
277 | ENDDO |
---|
278 | ENDDO |
---|
279 | ENDDO |
---|
280 | #if defined( __parallel ) |
---|
281 | IF ( collective_wait ) CALL MPI_BARRIER( comm2d, ierr ) |
---|
282 | CALL MPI_ALLREDUCE( w_l_l(1), w_l(1), nzt, MPI_REAL, MPI_SUM, comm2d, ierr ) |
---|
283 | #else |
---|
284 | w_l = w_l_l |
---|
285 | #endif |
---|
286 | DO k = 1, nzt |
---|
287 | w_l(k) = w_l(k) / ngp_2dh_outer(k,0) |
---|
288 | ENDDO |
---|
289 | DO i = nxlg, nxrg |
---|
290 | DO j = nysg, nyng |
---|
291 | DO k = nzb+1, nzt |
---|
292 | w(k,j,i) = w(k,j,i) - w_l(k) & |
---|
293 | * MERGE( 1.0_wp, 0.0_wp, BTEST( wall_flags_0(k,j,i), 3 ) ) |
---|
294 | ENDDO |
---|
295 | ENDDO |
---|
296 | ENDDO |
---|
297 | ENDIF |
---|
298 | |
---|
299 | ! |
---|
300 | !-- Compute the divergence of the provisional velocity field. |
---|
301 | CALL cpu_log( log_point_s(1), 'divergence', 'start' ) |
---|
302 | |
---|
303 | IF ( psolver(1:9) == 'multigrid' ) THEN |
---|
304 | !$OMP PARALLEL DO SCHEDULE( STATIC ) PRIVATE (i,j,k) |
---|
305 | DO i = nxl-1, nxr+1 |
---|
306 | DO j = nys-1, nyn+1 |
---|
307 | DO k = nzb, nzt+1 |
---|
308 | d(k,j,i) = 0.0_wp |
---|
309 | ENDDO |
---|
310 | ENDDO |
---|
311 | ENDDO |
---|
312 | ELSE |
---|
313 | !$OMP PARALLEL DO SCHEDULE( STATIC ) PRIVATE (i,j,k) |
---|
314 | !$ACC PARALLEL LOOP COLLAPSE(3) PRIVATE(i, j, k) & |
---|
315 | !$ACC PRESENT(d) |
---|
316 | DO i = nxl, nxr |
---|
317 | DO j = nys, nyn |
---|
318 | DO k = nzb+1, nzt |
---|
319 | d(k,j,i) = 0.0_wp |
---|
320 | ENDDO |
---|
321 | ENDDO |
---|
322 | ENDDO |
---|
323 | ENDIF |
---|
324 | |
---|
325 | localsum = 0.0_wp |
---|
326 | threadsum = 0.0_wp |
---|
327 | |
---|
328 | #if defined( __ibm ) |
---|
329 | !$OMP PARALLEL PRIVATE (i,j,k) FIRSTPRIVATE(threadsum) REDUCTION(+:localsum) |
---|
330 | !$OMP DO SCHEDULE( STATIC ) |
---|
331 | DO i = nxl, nxr |
---|
332 | DO j = nys, nyn |
---|
333 | DO k = nzb+1, nzt |
---|
334 | d(k,j,i) = ( ( u(k,j,i+1) - u(k,j,i) ) * rho_air(k) * ddx + & |
---|
335 | ( v(k,j+1,i) - v(k,j,i) ) * rho_air(k) * ddy + & |
---|
336 | ( w(k,j,i) * rho_air_zw(k) - & |
---|
337 | w(k-1,j,i) * rho_air_zw(k-1) ) * ddzw(k) & |
---|
338 | ) * ddt_3d * d_weight_pres & |
---|
339 | * MERGE( 1.0_wp, 0.0_wp, & |
---|
340 | BTEST( wall_flags_0(k,j,i), 0 ) & |
---|
341 | ) |
---|
342 | ENDDO |
---|
343 | ! |
---|
344 | !-- Compute possible PE-sum of divergences for flow_statistics |
---|
345 | DO k = nzb+1, nzt |
---|
346 | threadsum = threadsum + ABS( d(k,j,i) ) & |
---|
347 | * MERGE( 1.0_wp, 0.0_wp, & |
---|
348 | BTEST( wall_flags_0(k,j,i), 0 ) & |
---|
349 | ) |
---|
350 | ENDDO |
---|
351 | |
---|
352 | ENDDO |
---|
353 | ENDDO |
---|
354 | |
---|
355 | IF ( intermediate_timestep_count == intermediate_timestep_count_max .OR. & |
---|
356 | intermediate_timestep_count == 0 ) THEN |
---|
357 | localsum = localsum + threadsum * dt_3d * weight_pres_l |
---|
358 | ENDIF |
---|
359 | !$OMP END PARALLEL |
---|
360 | #else |
---|
361 | |
---|
362 | !$OMP PARALLEL PRIVATE (i,j,k) |
---|
363 | !$OMP DO SCHEDULE( STATIC ) |
---|
364 | !$ACC PARALLEL LOOP COLLAPSE(3) PRIVATE(i, j, k) & |
---|
365 | !$ACC PRESENT(u, v, w, rho_air, rho_air_zw, ddzw, wall_flags_0) & |
---|
366 | !$ACC PRESENT(d) |
---|
367 | DO i = nxl, nxr |
---|
368 | DO j = nys, nyn |
---|
369 | DO k = 1, nzt |
---|
370 | d(k,j,i) = ( ( u(k,j,i+1) - u(k,j,i) ) * rho_air(k) * ddx + & |
---|
371 | ( v(k,j+1,i) - v(k,j,i) ) * rho_air(k) * ddy + & |
---|
372 | ( w(k,j,i) * rho_air_zw(k) - & |
---|
373 | w(k-1,j,i) * rho_air_zw(k-1) ) * ddzw(k) & |
---|
374 | ) * ddt_3d * d_weight_pres & |
---|
375 | * MERGE( 1.0_wp, 0.0_wp, & |
---|
376 | BTEST( wall_flags_0(k,j,i), 0 ) & |
---|
377 | ) |
---|
378 | ENDDO |
---|
379 | ENDDO |
---|
380 | ENDDO |
---|
381 | !$OMP END PARALLEL |
---|
382 | |
---|
383 | ! |
---|
384 | !-- Compute possible PE-sum of divergences for flow_statistics. Carry out |
---|
385 | !-- computation only at last Runge-Kutta substep. |
---|
386 | IF ( intermediate_timestep_count == intermediate_timestep_count_max .OR. & |
---|
387 | intermediate_timestep_count == 0 ) THEN |
---|
388 | !$OMP PARALLEL PRIVATE (i,j,k) FIRSTPRIVATE(threadsum) REDUCTION(+:localsum) |
---|
389 | !$OMP DO SCHEDULE( STATIC ) |
---|
390 | !$ACC PARALLEL LOOP COLLAPSE(3) PRIVATE(i,j,k) & |
---|
391 | !$ACC REDUCTION(+:threadsum) COPY(threadsum) & |
---|
392 | !$ACC PRESENT(d) |
---|
393 | DO i = nxl, nxr |
---|
394 | DO j = nys, nyn |
---|
395 | DO k = nzb+1, nzt |
---|
396 | threadsum = threadsum + ABS( d(k,j,i) ) |
---|
397 | ENDDO |
---|
398 | ENDDO |
---|
399 | ENDDO |
---|
400 | localsum = localsum + threadsum * dt_3d * weight_pres_l |
---|
401 | !$OMP END PARALLEL |
---|
402 | ENDIF |
---|
403 | #endif |
---|
404 | |
---|
405 | ! |
---|
406 | !-- For completeness, set the divergence sum of all statistic regions to those |
---|
407 | !-- of the total domain |
---|
408 | IF ( intermediate_timestep_count == intermediate_timestep_count_max .OR. & |
---|
409 | intermediate_timestep_count == 0 ) THEN |
---|
410 | sums_divold_l(0:statistic_regions) = localsum |
---|
411 | ENDIF |
---|
412 | |
---|
413 | CALL cpu_log( log_point_s(1), 'divergence', 'stop' ) |
---|
414 | |
---|
415 | ! |
---|
416 | !-- Compute the pressure perturbation solving the Poisson equation |
---|
417 | IF ( psolver == 'poisfft' ) THEN |
---|
418 | |
---|
419 | ! |
---|
420 | !-- Solve Poisson equation via FFT and solution of tridiagonal matrices |
---|
421 | CALL poisfft( d ) |
---|
422 | |
---|
423 | ! |
---|
424 | !-- Store computed perturbation pressure and set boundary condition in |
---|
425 | !-- z-direction |
---|
426 | !$OMP PARALLEL DO PRIVATE (i,j,k) |
---|
427 | !$ACC PARALLEL LOOP COLLAPSE(3) PRIVATE(i, j, k) & |
---|
428 | !$ACC PRESENT(d, tend) |
---|
429 | DO i = nxl, nxr |
---|
430 | DO j = nys, nyn |
---|
431 | DO k = nzb+1, nzt |
---|
432 | tend(k,j,i) = d(k,j,i) |
---|
433 | ENDDO |
---|
434 | ENDDO |
---|
435 | ENDDO |
---|
436 | |
---|
437 | ! |
---|
438 | !-- Bottom boundary: |
---|
439 | !-- This condition is only required for internal output. The pressure |
---|
440 | !-- gradient (dp(nzb+1)-dp(nzb))/dz is not used anywhere else. |
---|
441 | IF ( ibc_p_b == 1 ) THEN |
---|
442 | ! |
---|
443 | !-- Neumann (dp/dz = 0). Using surfae data type, first for non-natural |
---|
444 | !-- surfaces, then for natural and urban surfaces |
---|
445 | !-- Upward facing |
---|
446 | !$OMP PARALLEL DO PRIVATE( i, j, k ) |
---|
447 | !$ACC PARALLEL LOOP PRIVATE(i, j, k) & |
---|
448 | !$ACC PRESENT(bc_h, tend) |
---|
449 | DO m = 1, bc_h(0)%ns |
---|
450 | i = bc_h(0)%i(m) |
---|
451 | j = bc_h(0)%j(m) |
---|
452 | k = bc_h(0)%k(m) |
---|
453 | tend(k-1,j,i) = tend(k,j,i) |
---|
454 | ENDDO |
---|
455 | ! |
---|
456 | !-- Downward facing |
---|
457 | !$OMP PARALLEL DO PRIVATE( i, j, k ) |
---|
458 | !$ACC PARALLEL LOOP PRIVATE(i, j, k) & |
---|
459 | !$ACC PRESENT(bc_h, tend) |
---|
460 | DO m = 1, bc_h(1)%ns |
---|
461 | i = bc_h(1)%i(m) |
---|
462 | j = bc_h(1)%j(m) |
---|
463 | k = bc_h(1)%k(m) |
---|
464 | tend(k+1,j,i) = tend(k,j,i) |
---|
465 | ENDDO |
---|
466 | |
---|
467 | ELSE |
---|
468 | ! |
---|
469 | !-- Dirichlet. Using surface data type, first for non-natural |
---|
470 | !-- surfaces, then for natural and urban surfaces |
---|
471 | !-- Upward facing |
---|
472 | !$OMP PARALLEL DO PRIVATE( i, j, k ) |
---|
473 | DO m = 1, bc_h(0)%ns |
---|
474 | i = bc_h(0)%i(m) |
---|
475 | j = bc_h(0)%j(m) |
---|
476 | k = bc_h(0)%k(m) |
---|
477 | tend(k-1,j,i) = 0.0_wp |
---|
478 | ENDDO |
---|
479 | ! |
---|
480 | !-- Downward facing |
---|
481 | !$OMP PARALLEL DO PRIVATE( i, j, k ) |
---|
482 | DO m = 1, bc_h(1)%ns |
---|
483 | i = bc_h(1)%i(m) |
---|
484 | j = bc_h(1)%j(m) |
---|
485 | k = bc_h(1)%k(m) |
---|
486 | tend(k+1,j,i) = 0.0_wp |
---|
487 | ENDDO |
---|
488 | |
---|
489 | ENDIF |
---|
490 | |
---|
491 | ! |
---|
492 | !-- Top boundary |
---|
493 | IF ( ibc_p_t == 1 ) THEN |
---|
494 | ! |
---|
495 | !-- Neumann |
---|
496 | !$OMP PARALLEL DO PRIVATE (i,j,k) |
---|
497 | DO i = nxlg, nxrg |
---|
498 | DO j = nysg, nyng |
---|
499 | tend(nzt+1,j,i) = tend(nzt,j,i) |
---|
500 | ENDDO |
---|
501 | ENDDO |
---|
502 | |
---|
503 | ELSE |
---|
504 | ! |
---|
505 | !-- Dirichlet |
---|
506 | !$OMP PARALLEL DO PRIVATE (i,j,k) |
---|
507 | !$ACC PARALLEL LOOP COLLAPSE(2) PRIVATE(i, j) & |
---|
508 | !$ACC PRESENT(tend) |
---|
509 | DO i = nxlg, nxrg |
---|
510 | DO j = nysg, nyng |
---|
511 | tend(nzt+1,j,i) = 0.0_wp |
---|
512 | ENDDO |
---|
513 | ENDDO |
---|
514 | |
---|
515 | ENDIF |
---|
516 | |
---|
517 | ! |
---|
518 | !-- Exchange boundaries for p |
---|
519 | CALL exchange_horiz( tend, nbgp ) |
---|
520 | |
---|
521 | ELSEIF ( psolver == 'sor' ) THEN |
---|
522 | |
---|
523 | ! |
---|
524 | !-- Solve Poisson equation for perturbation pressure using SOR-Red/Black |
---|
525 | !-- scheme |
---|
526 | CALL sor( d, ddzu_pres, ddzw, p_loc ) |
---|
527 | tend = p_loc |
---|
528 | |
---|
529 | ELSEIF ( psolver(1:9) == 'multigrid' ) THEN |
---|
530 | |
---|
531 | ! |
---|
532 | !-- Solve Poisson equation for perturbation pressure using Multigrid scheme, |
---|
533 | !-- array tend is used to store the residuals. |
---|
534 | |
---|
535 | !-- If the number of grid points of the gathered grid, which is collected |
---|
536 | !-- on PE0, is larger than the number of grid points of an PE, than array |
---|
537 | !-- tend will be enlarged. |
---|
538 | IF ( gathered_size > subdomain_size ) THEN |
---|
539 | DEALLOCATE( tend ) |
---|
540 | ALLOCATE( tend(nzb:nzt_mg(mg_switch_to_pe0_level)+1,nys_mg( & |
---|
541 | mg_switch_to_pe0_level)-1:nyn_mg(mg_switch_to_pe0_level)+1,& |
---|
542 | nxl_mg(mg_switch_to_pe0_level)-1:nxr_mg( & |
---|
543 | mg_switch_to_pe0_level)+1) ) |
---|
544 | ENDIF |
---|
545 | |
---|
546 | IF ( psolver == 'multigrid' ) THEN |
---|
547 | CALL poismg( tend ) |
---|
548 | ELSE |
---|
549 | CALL poismg_noopt( tend ) |
---|
550 | ENDIF |
---|
551 | |
---|
552 | IF ( gathered_size > subdomain_size ) THEN |
---|
553 | DEALLOCATE( tend ) |
---|
554 | ALLOCATE( tend(nzb:nzt+1,nysg:nyng,nxlg:nxrg) ) |
---|
555 | ENDIF |
---|
556 | |
---|
557 | ! |
---|
558 | !-- Restore perturbation pressure on tend because this array is used |
---|
559 | !-- further below to correct the velocity fields |
---|
560 | DO i = nxl-1, nxr+1 |
---|
561 | DO j = nys-1, nyn+1 |
---|
562 | DO k = nzb, nzt+1 |
---|
563 | tend(k,j,i) = p_loc(k,j,i) |
---|
564 | ENDDO |
---|
565 | ENDDO |
---|
566 | ENDDO |
---|
567 | |
---|
568 | ENDIF |
---|
569 | |
---|
570 | ! |
---|
571 | !-- Store perturbation pressure on array p, used for pressure data output. |
---|
572 | !-- Ghost layers are added in the output routines (except sor-method: see below) |
---|
573 | IF ( intermediate_timestep_count <= 1 ) THEN |
---|
574 | !$OMP PARALLEL PRIVATE (i,j,k) |
---|
575 | !$OMP DO |
---|
576 | !$ACC PARALLEL LOOP COLLAPSE(3) PRIVATE(i, j, k) & |
---|
577 | !$ACC PRESENT(p, tend) |
---|
578 | DO i = nxl-1, nxr+1 |
---|
579 | DO j = nys-1, nyn+1 |
---|
580 | DO k = nzb, nzt+1 |
---|
581 | p(k,j,i) = tend(k,j,i) * & |
---|
582 | weight_substep_l |
---|
583 | ENDDO |
---|
584 | ENDDO |
---|
585 | ENDDO |
---|
586 | !$OMP END PARALLEL |
---|
587 | |
---|
588 | ELSEIF ( intermediate_timestep_count > 1 ) THEN |
---|
589 | !$OMP PARALLEL PRIVATE (i,j,k) |
---|
590 | !$OMP DO |
---|
591 | !$ACC PARALLEL LOOP COLLAPSE(3) PRIVATE(i, j, k) & |
---|
592 | !$ACC PRESENT(p, tend) |
---|
593 | DO i = nxl-1, nxr+1 |
---|
594 | DO j = nys-1, nyn+1 |
---|
595 | DO k = nzb, nzt+1 |
---|
596 | p(k,j,i) = p(k,j,i) + tend(k,j,i) * & |
---|
597 | weight_substep_l |
---|
598 | ENDDO |
---|
599 | ENDDO |
---|
600 | ENDDO |
---|
601 | !$OMP END PARALLEL |
---|
602 | |
---|
603 | ENDIF |
---|
604 | |
---|
605 | ! |
---|
606 | !-- SOR-method needs ghost layers for the next timestep |
---|
607 | IF ( psolver == 'sor' ) CALL exchange_horiz( p, nbgp ) |
---|
608 | |
---|
609 | ! |
---|
610 | !-- Correction of the provisional velocities with the current perturbation |
---|
611 | !-- pressure just computed |
---|
612 | IF ( conserve_volume_flow .AND. ( bc_lr_cyc .OR. bc_ns_cyc ) ) THEN |
---|
613 | volume_flow_l(1) = 0.0_wp |
---|
614 | volume_flow_l(2) = 0.0_wp |
---|
615 | ENDIF |
---|
616 | ! |
---|
617 | !-- Add pressure gradients to the velocity components. Note, the loops are |
---|
618 | !-- running over the entire model domain, even though, in case of non-cyclic |
---|
619 | !-- boundaries u- and v-component are not prognostic at i=0 and j=0, |
---|
620 | !-- respectiveley. However, in case of Dirichlet boundary conditions for the |
---|
621 | !-- velocities, zero-gradient conditions for the pressure are set, so that |
---|
622 | !-- no modification is imposed at the boundaries. |
---|
623 | !$OMP PARALLEL PRIVATE (i,j,k) |
---|
624 | !$OMP DO |
---|
625 | !$ACC PARALLEL LOOP COLLAPSE(2) PRIVATE(i, j, k) & |
---|
626 | !$ACC PRESENT(u, v, w, tend, ddzu, wall_flags_0) |
---|
627 | DO i = nxl, nxr |
---|
628 | DO j = nys, nyn |
---|
629 | |
---|
630 | DO k = nzb+1, nzt |
---|
631 | w(k,j,i) = w(k,j,i) - dt_3d * & |
---|
632 | ( tend(k+1,j,i) - tend(k,j,i) ) * ddzu(k+1) & |
---|
633 | * weight_pres_l & |
---|
634 | * MERGE( 1.0_wp, 0.0_wp, & |
---|
635 | BTEST( wall_flags_0(k,j,i), 3 ) & |
---|
636 | ) |
---|
637 | ENDDO |
---|
638 | |
---|
639 | DO k = nzb+1, nzt |
---|
640 | u(k,j,i) = u(k,j,i) - dt_3d * & |
---|
641 | ( tend(k,j,i) - tend(k,j,i-1) ) * ddx & |
---|
642 | * weight_pres_l & |
---|
643 | * MERGE( 1.0_wp, 0.0_wp, & |
---|
644 | BTEST( wall_flags_0(k,j,i), 1 ) & |
---|
645 | ) |
---|
646 | ENDDO |
---|
647 | |
---|
648 | DO k = nzb+1, nzt |
---|
649 | v(k,j,i) = v(k,j,i) - dt_3d * & |
---|
650 | ( tend(k,j,i) - tend(k,j-1,i) ) * ddy & |
---|
651 | * weight_pres_l & |
---|
652 | * MERGE( 1.0_wp, 0.0_wp, & |
---|
653 | BTEST( wall_flags_0(k,j,i), 2 ) & |
---|
654 | ) |
---|
655 | ENDDO |
---|
656 | |
---|
657 | ENDDO |
---|
658 | ENDDO |
---|
659 | !$OMP END PARALLEL |
---|
660 | |
---|
661 | ! |
---|
662 | !-- The vertical velocity is not set to zero at nzt + 1 for nested domains |
---|
663 | !-- Instead it is set to the values of nzt (see routine vnest_boundary_conds |
---|
664 | !-- or pmci_interp_tril_t) BEFORE calling the pressure solver. To avoid jumps |
---|
665 | !-- while plotting profiles w at the top has to be set to the values in the |
---|
666 | !-- height nzt after above modifications. Hint: w level nzt+1 does not impact |
---|
667 | !-- results. |
---|
668 | IF ( child_domain .OR. coupling_mode == 'vnested_fine' ) THEN |
---|
669 | w(nzt+1,:,:) = w(nzt,:,:) |
---|
670 | ENDIF |
---|
671 | |
---|
672 | ! |
---|
673 | !-- Sum up the volume flow through the right and north boundary |
---|
674 | IF ( conserve_volume_flow .AND. bc_lr_cyc .AND. bc_ns_cyc .AND. & |
---|
675 | nxr == nx ) THEN |
---|
676 | |
---|
677 | !$OMP PARALLEL PRIVATE (j,k) |
---|
678 | !$OMP DO |
---|
679 | DO j = nys, nyn |
---|
680 | !$OMP CRITICAL |
---|
681 | DO k = nzb+1, nzt |
---|
682 | volume_flow_l(1) = volume_flow_l(1) + u(k,j,nxr) * dzw(k) & |
---|
683 | * MERGE( 1.0_wp, 0.0_wp, & |
---|
684 | BTEST( wall_flags_0(k,j,nxr), 1 )& |
---|
685 | ) |
---|
686 | ENDDO |
---|
687 | !$OMP END CRITICAL |
---|
688 | ENDDO |
---|
689 | !$OMP END PARALLEL |
---|
690 | |
---|
691 | ENDIF |
---|
692 | |
---|
693 | IF ( conserve_volume_flow .AND. bc_ns_cyc .AND. bc_lr_cyc .AND. & |
---|
694 | nyn == ny ) THEN |
---|
695 | |
---|
696 | !$OMP PARALLEL PRIVATE (i,k) |
---|
697 | !$OMP DO |
---|
698 | DO i = nxl, nxr |
---|
699 | !$OMP CRITICAL |
---|
700 | DO k = nzb+1, nzt |
---|
701 | volume_flow_l(2) = volume_flow_l(2) + v(k,nyn,i) * dzw(k) & |
---|
702 | * MERGE( 1.0_wp, 0.0_wp, & |
---|
703 | BTEST( wall_flags_0(k,nyn,i), 2 )& |
---|
704 | ) |
---|
705 | ENDDO |
---|
706 | !$OMP END CRITICAL |
---|
707 | ENDDO |
---|
708 | !$OMP END PARALLEL |
---|
709 | |
---|
710 | ENDIF |
---|
711 | |
---|
712 | ! |
---|
713 | !-- Conserve the volume flow |
---|
714 | IF ( conserve_volume_flow .AND. ( bc_lr_cyc .AND. bc_ns_cyc ) ) THEN |
---|
715 | |
---|
716 | #if defined( __parallel ) |
---|
717 | IF ( collective_wait ) CALL MPI_BARRIER( comm2d, ierr ) |
---|
718 | CALL MPI_ALLREDUCE( volume_flow_l(1), volume_flow(1), 2, MPI_REAL, & |
---|
719 | MPI_SUM, comm2d, ierr ) |
---|
720 | #else |
---|
721 | volume_flow = volume_flow_l |
---|
722 | #endif |
---|
723 | |
---|
724 | volume_flow_offset(1:2) = ( volume_flow_initial(1:2) - volume_flow(1:2) ) / & |
---|
725 | volume_flow_area(1:2) |
---|
726 | |
---|
727 | !$OMP PARALLEL PRIVATE (i,j,k) |
---|
728 | !$OMP DO |
---|
729 | DO i = nxl, nxr |
---|
730 | DO j = nys, nyn |
---|
731 | DO k = nzb+1, nzt |
---|
732 | u(k,j,i) = u(k,j,i) + volume_flow_offset(1) & |
---|
733 | * MERGE( 1.0_wp, 0.0_wp, & |
---|
734 | BTEST( wall_flags_0(k,j,i), 1 ) & |
---|
735 | ) |
---|
736 | ENDDO |
---|
737 | DO k = nzb+1, nzt |
---|
738 | v(k,j,i) = v(k,j,i) + volume_flow_offset(2) & |
---|
739 | * MERGE( 1.0_wp, 0.0_wp, & |
---|
740 | BTEST( wall_flags_0(k,j,i), 2 ) & |
---|
741 | ) |
---|
742 | ENDDO |
---|
743 | ENDDO |
---|
744 | ENDDO |
---|
745 | |
---|
746 | !$OMP END PARALLEL |
---|
747 | |
---|
748 | ENDIF |
---|
749 | |
---|
750 | ! |
---|
751 | !-- Exchange of boundaries for the velocities |
---|
752 | CALL exchange_horiz( u, nbgp ) |
---|
753 | CALL exchange_horiz( v, nbgp ) |
---|
754 | CALL exchange_horiz( w, nbgp ) |
---|
755 | |
---|
756 | ! |
---|
757 | !-- Compute the divergence of the corrected velocity field, |
---|
758 | !-- A possible PE-sum is computed in flow_statistics. Carry out computation |
---|
759 | !-- only at last Runge-Kutta step. |
---|
760 | IF ( intermediate_timestep_count == intermediate_timestep_count_max .OR. & |
---|
761 | intermediate_timestep_count == 0 ) THEN |
---|
762 | CALL cpu_log( log_point_s(1), 'divergence', 'start' ) |
---|
763 | sums_divnew_l = 0.0_wp |
---|
764 | |
---|
765 | ! |
---|
766 | !-- d must be reset to zero because it can contain nonzero values below the |
---|
767 | !-- topography |
---|
768 | IF ( topography /= 'flat' ) d = 0.0_wp |
---|
769 | |
---|
770 | localsum = 0.0_wp |
---|
771 | threadsum = 0.0_wp |
---|
772 | |
---|
773 | !$OMP PARALLEL PRIVATE (i,j,k) FIRSTPRIVATE(threadsum) REDUCTION(+:localsum) |
---|
774 | #if defined( __ibm ) |
---|
775 | !$OMP DO SCHEDULE( STATIC ) |
---|
776 | DO i = nxl, nxr |
---|
777 | DO j = nys, nyn |
---|
778 | DO k = nzb+1, nzt |
---|
779 | d(k,j,i) = ( ( u(k,j,i+1) - u(k,j,i) ) * rho_air(k) * ddx + & |
---|
780 | ( v(k,j+1,i) - v(k,j,i) ) * rho_air(k) * ddy + & |
---|
781 | ( w(k,j,i) * rho_air_zw(k) - & |
---|
782 | w(k-1,j,i) * rho_air_zw(k-1) ) * ddzw(k) & |
---|
783 | ) * MERGE( 1.0_wp, 0.0_wp, & |
---|
784 | BTEST( wall_flags_0(k,j,i), 0 ) & |
---|
785 | ) |
---|
786 | ENDDO |
---|
787 | DO k = nzb+1, nzt |
---|
788 | threadsum = threadsum + ABS( d(k,j,i) ) |
---|
789 | ENDDO |
---|
790 | ENDDO |
---|
791 | ENDDO |
---|
792 | #else |
---|
793 | !$OMP DO SCHEDULE( STATIC ) |
---|
794 | !$ACC PARALLEL LOOP COLLAPSE(3) PRIVATE(i, j, k) & |
---|
795 | !$ACC PRESENT(u, v, w, rho_air, rho_air_zw, ddzw, wall_flags_0) & |
---|
796 | !$ACC PRESENT(d) |
---|
797 | DO i = nxl, nxr |
---|
798 | DO j = nys, nyn |
---|
799 | DO k = nzb+1, nzt |
---|
800 | d(k,j,i) = ( ( u(k,j,i+1) - u(k,j,i) ) * rho_air(k) * ddx + & |
---|
801 | ( v(k,j+1,i) - v(k,j,i) ) * rho_air(k) * ddy + & |
---|
802 | ( w(k,j,i) * rho_air_zw(k) - & |
---|
803 | w(k-1,j,i) * rho_air_zw(k-1) ) * ddzw(k) & |
---|
804 | ) * MERGE( 1.0_wp, 0.0_wp, & |
---|
805 | BTEST( wall_flags_0(k,j,i), 0 ) & |
---|
806 | ) |
---|
807 | ENDDO |
---|
808 | ENDDO |
---|
809 | ENDDO |
---|
810 | ! |
---|
811 | !-- Compute possible PE-sum of divergences for flow_statistics |
---|
812 | !$OMP DO SCHEDULE( STATIC ) |
---|
813 | !$ACC PARALLEL LOOP COLLAPSE(3) PRIVATE(i, j, k) & |
---|
814 | !$ACC REDUCTION(+:threadsum) COPY(threadsum) & |
---|
815 | !$ACC PRESENT(d) |
---|
816 | DO i = nxl, nxr |
---|
817 | DO j = nys, nyn |
---|
818 | DO k = nzb+1, nzt |
---|
819 | threadsum = threadsum + ABS( d(k,j,i) ) |
---|
820 | ENDDO |
---|
821 | ENDDO |
---|
822 | ENDDO |
---|
823 | #endif |
---|
824 | |
---|
825 | localsum = localsum + threadsum |
---|
826 | !$OMP END PARALLEL |
---|
827 | |
---|
828 | ! |
---|
829 | !-- For completeness, set the divergence sum of all statistic regions to those |
---|
830 | !-- of the total domain |
---|
831 | sums_divnew_l(0:statistic_regions) = localsum |
---|
832 | |
---|
833 | CALL cpu_log( log_point_s(1), 'divergence', 'stop' ) |
---|
834 | |
---|
835 | ENDIF |
---|
836 | |
---|
837 | CALL cpu_log( log_point(8), 'pres', 'stop' ) |
---|
838 | |
---|
839 | END SUBROUTINE pres |
---|