1 | !> @file timestep.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 1997-2020 Leibniz Universitaet Hannover |
---|
18 | !------------------------------------------------------------------------------! |
---|
19 | ! |
---|
20 | ! Current revisions: |
---|
21 | ! ------------------ |
---|
22 | ! |
---|
23 | ! |
---|
24 | ! Former revisions: |
---|
25 | ! ----------------- |
---|
26 | ! $Id: timestep.f90 4360 2020-01-07 11:25:50Z Giersch $ |
---|
27 | ! Added missing OpenMP directives |
---|
28 | ! |
---|
29 | ! 4233 2019-09-20 09:55:54Z knoop |
---|
30 | ! OpenACC data update host removed |
---|
31 | ! |
---|
32 | ! 4182 2019-08-22 15:20:23Z scharf |
---|
33 | ! Corrected "Former revisions" section |
---|
34 | ! |
---|
35 | ! 4101 2019-07-17 15:14:26Z gronemeier |
---|
36 | ! - consider 2*Km within diffusion criterion as Km is considered twice within |
---|
37 | ! the diffusion of e, |
---|
38 | ! - in RANS mode, instead of considering each wind component individually use |
---|
39 | ! the wind speed of 3d wind vector in CFL criterion |
---|
40 | ! - do not limit the increase of dt based on its previous value in RANS mode |
---|
41 | ! |
---|
42 | ! 3658 2019-01-07 20:28:54Z knoop |
---|
43 | ! OpenACC port for SPEC |
---|
44 | ! |
---|
45 | ! Revision 1.1 1997/08/11 06:26:19 raasch |
---|
46 | ! Initial revision |
---|
47 | ! |
---|
48 | ! |
---|
49 | ! Description: |
---|
50 | ! ------------ |
---|
51 | !> Compute the time step under consideration of the FCL and diffusion criterion. |
---|
52 | !------------------------------------------------------------------------------! |
---|
53 | SUBROUTINE timestep |
---|
54 | |
---|
55 | |
---|
56 | USE arrays_3d, & |
---|
57 | ONLY: dzu, dzw, kh, km, u, u_stokes_zu, v, v_stokes_zu, w |
---|
58 | |
---|
59 | USE control_parameters, & |
---|
60 | ONLY: cfl_factor, coupling_mode, dt_3d, dt_fixed, dt_max, & |
---|
61 | galilei_transformation, message_string, rans_mode, & |
---|
62 | stop_dt, terminate_coupled, terminate_coupled_remote, & |
---|
63 | timestep_reason, u_gtrans, use_ug_for_galilei_tr, v_gtrans |
---|
64 | |
---|
65 | USE cpulog, & |
---|
66 | ONLY: cpu_log, log_point |
---|
67 | |
---|
68 | USE grid_variables, & |
---|
69 | ONLY: dx, dx2, dy, dy2 |
---|
70 | |
---|
71 | USE indices, & |
---|
72 | ONLY: nxl, nxlg, nxr, nxrg, nyn, nyng, nys, nysg, nzb, nzt |
---|
73 | |
---|
74 | USE interfaces |
---|
75 | |
---|
76 | USE kinds |
---|
77 | |
---|
78 | USE bulk_cloud_model_mod, & |
---|
79 | ONLY: dt_precipitation |
---|
80 | |
---|
81 | USE pegrid |
---|
82 | |
---|
83 | USE pmc_interface, & |
---|
84 | ONLY: nested_run |
---|
85 | |
---|
86 | USE statistics, & |
---|
87 | ONLY: flow_statistics_called, hom, u_max, u_max_ijk, v_max, v_max_ijk,& |
---|
88 | w_max, w_max_ijk |
---|
89 | |
---|
90 | USE vertical_nesting_mod, & |
---|
91 | ONLY: vnested, vnest_timestep_sync |
---|
92 | |
---|
93 | IMPLICIT NONE |
---|
94 | |
---|
95 | INTEGER(iwp) :: i !< |
---|
96 | INTEGER(iwp) :: j !< |
---|
97 | INTEGER(iwp) :: k !< |
---|
98 | INTEGER(iwp) :: km_max_ijk(3) = -1 !< index values (i,j,k) of location where km_max occurs |
---|
99 | INTEGER(iwp) :: kh_max_ijk(3) = -1 !< index values (i,j,k) of location where kh_max occurs |
---|
100 | |
---|
101 | LOGICAL :: stop_dt_local !< local switch for controlling the time stepping |
---|
102 | |
---|
103 | REAL(wp) :: div !< |
---|
104 | REAL(wp) :: dt_diff !< |
---|
105 | REAL(wp) :: dt_diff_l !< |
---|
106 | REAL(wp) :: dt_u !< |
---|
107 | REAL(wp) :: dt_u_l !< |
---|
108 | REAL(wp) :: dt_v !< |
---|
109 | REAL(wp) :: dt_v_l !< |
---|
110 | REAL(wp) :: dt_w !< |
---|
111 | REAL(wp) :: dt_w_l !< |
---|
112 | REAL(wp) :: km_max !< maximum of Km in entire domain |
---|
113 | REAL(wp) :: kh_max !< maximum of Kh in entire domain |
---|
114 | REAL(wp) :: u_gtrans_l !< |
---|
115 | REAL(wp) :: v_gtrans_l !< |
---|
116 | |
---|
117 | REAL(wp), DIMENSION(2) :: uv_gtrans !< |
---|
118 | REAL(wp), DIMENSION(2) :: uv_gtrans_l !< |
---|
119 | REAL(wp), DIMENSION(3) :: reduce !< |
---|
120 | REAL(wp), DIMENSION(3) :: reduce_l !< |
---|
121 | REAL(wp), DIMENSION(nzb+1:nzt) :: dxyz2_min !< |
---|
122 | !$ACC DECLARE CREATE(dxyz2_min) |
---|
123 | |
---|
124 | |
---|
125 | CALL cpu_log( log_point(12), 'calculate_timestep', 'start' ) |
---|
126 | |
---|
127 | ! |
---|
128 | !-- In case of Galilei-transform not using the geostrophic wind as translation |
---|
129 | !-- velocity, compute the volume-averaged horizontal velocity components, which |
---|
130 | !-- will then be subtracted from the horizontal wind for the time step and |
---|
131 | !-- horizontal advection routines. |
---|
132 | IF ( galilei_transformation .AND. .NOT. use_ug_for_galilei_tr ) THEN |
---|
133 | IF ( flow_statistics_called ) THEN |
---|
134 | ! |
---|
135 | !-- Horizontal averages already existent, just need to average them |
---|
136 | !-- vertically. |
---|
137 | u_gtrans = 0.0_wp |
---|
138 | v_gtrans = 0.0_wp |
---|
139 | DO k = nzb+1, nzt |
---|
140 | u_gtrans = u_gtrans + hom(k,1,1,0) |
---|
141 | v_gtrans = v_gtrans + hom(k,1,2,0) |
---|
142 | ENDDO |
---|
143 | u_gtrans = u_gtrans / REAL( nzt - nzb, KIND=wp ) |
---|
144 | v_gtrans = v_gtrans / REAL( nzt - nzb, KIND=wp ) |
---|
145 | ELSE |
---|
146 | ! |
---|
147 | !-- Averaging over the entire model domain. |
---|
148 | u_gtrans_l = 0.0_wp |
---|
149 | v_gtrans_l = 0.0_wp |
---|
150 | DO i = nxl, nxr |
---|
151 | DO j = nys, nyn |
---|
152 | DO k = nzb+1, nzt |
---|
153 | u_gtrans_l = u_gtrans_l + u(k,j,i) |
---|
154 | v_gtrans_l = v_gtrans_l + v(k,j,i) |
---|
155 | ENDDO |
---|
156 | ENDDO |
---|
157 | ENDDO |
---|
158 | uv_gtrans_l(1) = u_gtrans_l / & |
---|
159 | REAL( (nxr-nxl+1)*(nyn-nys+1)*(nzt-nzb), KIND=wp ) |
---|
160 | uv_gtrans_l(2) = v_gtrans_l / & |
---|
161 | REAL( (nxr-nxl+1)*(nyn-nys+1)*(nzt-nzb), KIND=wp ) |
---|
162 | #if defined( __parallel ) |
---|
163 | IF ( collective_wait ) CALL MPI_BARRIER( comm2d, ierr ) |
---|
164 | CALL MPI_ALLREDUCE( uv_gtrans_l, uv_gtrans, 2, MPI_REAL, MPI_SUM, & |
---|
165 | comm2d, ierr ) |
---|
166 | u_gtrans = uv_gtrans(1) / REAL( numprocs, KIND=wp ) |
---|
167 | v_gtrans = uv_gtrans(2) / REAL( numprocs, KIND=wp ) |
---|
168 | #else |
---|
169 | u_gtrans = uv_gtrans_l(1) |
---|
170 | v_gtrans = uv_gtrans_l(2) |
---|
171 | #endif |
---|
172 | ENDIF |
---|
173 | ENDIF |
---|
174 | |
---|
175 | ! |
---|
176 | !-- Determine the maxima of the velocity components, including their |
---|
177 | !-- grid index positions. |
---|
178 | CALL global_min_max( nzb, nzt+1, nysg, nyng, nxlg, nxrg, u, 'abs', 0.0_wp, & |
---|
179 | u_max, u_max_ijk ) |
---|
180 | CALL global_min_max( nzb, nzt+1, nysg, nyng, nxlg, nxrg, v, 'abs', 0.0_wp, & |
---|
181 | v_max, v_max_ijk ) |
---|
182 | CALL global_min_max( nzb, nzt+1, nysg, nyng, nxlg, nxrg, w, 'abs', 0.0_wp, & |
---|
183 | w_max, w_max_ijk ) |
---|
184 | |
---|
185 | IF ( .NOT. dt_fixed ) THEN |
---|
186 | ! |
---|
187 | !-- Variable time step: |
---|
188 | !-- Calculate the maximum time step according to the CFL-criterion |
---|
189 | dt_u_l = 999999.9_wp |
---|
190 | dt_v_l = 999999.9_wp |
---|
191 | dt_w_l = 999999.9_wp |
---|
192 | |
---|
193 | IF ( .NOT. rans_mode ) THEN |
---|
194 | ! |
---|
195 | !-- Consider each velocity component individually |
---|
196 | |
---|
197 | !$ACC PARALLEL LOOP COLLAPSE(3) PRIVATE(i,j,k) & |
---|
198 | !$ACC COPY(dt_u_l, dt_v_l, dt_w_l, u_stokes_zu, v_stokes_zu) & |
---|
199 | !$ACC REDUCTION(MIN: dt_u_l, dt_v_l, dt_w_l) & |
---|
200 | !$ACC PRESENT(u, v, w, dzu) |
---|
201 | !$OMP PARALLEL DO PRIVATE(i,j,k) & |
---|
202 | !$OMP REDUCTION(MIN: dt_u_l, dt_v_l, dt_w_l) |
---|
203 | DO i = nxl, nxr |
---|
204 | DO j = nys, nyn |
---|
205 | DO k = nzb+1, nzt |
---|
206 | dt_u_l = MIN( dt_u_l, ( dx / & |
---|
207 | ( ABS( u(k,j,i) - u_gtrans + u_stokes_zu(k) ) & |
---|
208 | + 1.0E-10_wp ) ) ) |
---|
209 | dt_v_l = MIN( dt_v_l, ( dy / & |
---|
210 | ( ABS( v(k,j,i) - v_gtrans + v_stokes_zu(k) ) & |
---|
211 | + 1.0E-10_wp ) ) ) |
---|
212 | dt_w_l = MIN( dt_w_l, ( dzu(k) / & |
---|
213 | ( ABS( w(k,j,i) ) + 1.0E-10_wp ) ) ) |
---|
214 | ENDDO |
---|
215 | ENDDO |
---|
216 | ENDDO |
---|
217 | |
---|
218 | ELSE |
---|
219 | ! |
---|
220 | !-- Consider the wind speed at the scalar-grid point |
---|
221 | !-- !> @note considering the wind speed instead of each individual wind |
---|
222 | !-- !> component is only a workaround so far. This might has to be |
---|
223 | !-- !> changed in the future. |
---|
224 | |
---|
225 | !$ACC PARALLEL LOOP COLLAPSE(3) PRIVATE(i,j,k) & |
---|
226 | !$ACC COPY(dt_u_l, u_stokes_zu, v_stokes_zu) & |
---|
227 | !$ACC REDUCTION(MIN: dt_u_l) & |
---|
228 | !$ACC PRESENT(u, v, w, dzu) |
---|
229 | !$OMP PARALLEL DO PRIVATE(i,j,k) & |
---|
230 | !$OMP REDUCTION(MIN: dt_u_l) |
---|
231 | DO i = nxl, nxr |
---|
232 | DO j = nys, nyn |
---|
233 | DO k = nzb+1, nzt |
---|
234 | dt_u_l = MIN( dt_u_l, ( MIN( dx, dy, dzu(k) ) / ( & |
---|
235 | SQRT( ( 0.5 * ( u(k,j,i) + u(k,j,i+1) ) - u_gtrans + u_stokes_zu(k) )**2 & |
---|
236 | + ( 0.5 * ( v(k,j,i) + v(k,j+1,i) ) - v_gtrans + v_stokes_zu(k) )**2 & |
---|
237 | + ( 0.5 * ( w(k,j,i) + w(k-1,j,i) ) )**2 ) & |
---|
238 | + 1.0E-10_wp ) ) ) |
---|
239 | ENDDO |
---|
240 | ENDDO |
---|
241 | ENDDO |
---|
242 | |
---|
243 | dt_v_l = dt_u_l |
---|
244 | dt_w_l = dt_u_l |
---|
245 | |
---|
246 | ENDIF |
---|
247 | |
---|
248 | #if defined( __parallel ) |
---|
249 | reduce_l(1) = dt_u_l |
---|
250 | reduce_l(2) = dt_v_l |
---|
251 | reduce_l(3) = dt_w_l |
---|
252 | IF ( collective_wait ) CALL MPI_BARRIER( comm2d, ierr ) |
---|
253 | CALL MPI_ALLREDUCE( reduce_l, reduce, 3, MPI_REAL, MPI_MIN, comm2d, ierr ) |
---|
254 | dt_u = reduce(1) |
---|
255 | dt_v = reduce(2) |
---|
256 | dt_w = reduce(3) |
---|
257 | #else |
---|
258 | dt_u = dt_u_l |
---|
259 | dt_v = dt_v_l |
---|
260 | dt_w = dt_w_l |
---|
261 | #endif |
---|
262 | |
---|
263 | ! |
---|
264 | !-- Compute time step according to the diffusion criterion. |
---|
265 | !-- First calculate minimum grid spacing which only depends on index k. |
---|
266 | !-- When using the dynamic subgrid model, negative km are possible. |
---|
267 | dt_diff_l = 999999.0_wp |
---|
268 | |
---|
269 | !$ACC PARALLEL LOOP PRESENT(dxyz2_min, dzw) |
---|
270 | DO k = nzb+1, nzt |
---|
271 | dxyz2_min(k) = MIN( dx2, dy2, dzw(k)*dzw(k) ) * 0.125_wp |
---|
272 | ENDDO |
---|
273 | |
---|
274 | !$OMP PARALLEL private(i,j,k) reduction(MIN: dt_diff_l) |
---|
275 | !$OMP DO |
---|
276 | !$ACC PARALLEL LOOP COLLAPSE(3) PRIVATE(i,j,k) & |
---|
277 | !$ACC COPY(dt_diff_l) REDUCTION(MIN: dt_diff_l) & |
---|
278 | !$ACC PRESENT(dxyz2_min, kh, km) |
---|
279 | DO i = nxl, nxr |
---|
280 | DO j = nys, nyn |
---|
281 | DO k = nzb+1, nzt |
---|
282 | dt_diff_l = MIN( dt_diff_l, & |
---|
283 | dxyz2_min(k) / & |
---|
284 | ( MAX( kh(k,j,i), 2.0_wp * ABS( km(k,j,i) ) ) & |
---|
285 | + 1E-20_wp ) ) |
---|
286 | ENDDO |
---|
287 | ENDDO |
---|
288 | ENDDO |
---|
289 | !$OMP END PARALLEL |
---|
290 | #if defined( __parallel ) |
---|
291 | IF ( collective_wait ) CALL MPI_BARRIER( comm2d, ierr ) |
---|
292 | CALL MPI_ALLREDUCE( dt_diff_l, dt_diff, 1, MPI_REAL, MPI_MIN, comm2d, & |
---|
293 | ierr ) |
---|
294 | #else |
---|
295 | dt_diff = dt_diff_l |
---|
296 | #endif |
---|
297 | |
---|
298 | ! |
---|
299 | !-- The time step is the minimum of the 3-4 components and the diffusion time |
---|
300 | !-- step minus a reduction (cfl_factor) to be on the safe side. |
---|
301 | !-- The time step must not exceed the maximum allowed value. |
---|
302 | dt_3d = cfl_factor * MIN( dt_diff, dt_u, dt_v, dt_w, dt_precipitation ) |
---|
303 | dt_3d = MIN( dt_3d, dt_max ) |
---|
304 | |
---|
305 | ! |
---|
306 | !-- Remember the restricting time step criterion for later output. |
---|
307 | IF ( MIN( dt_u, dt_v, dt_w ) < dt_diff ) THEN |
---|
308 | timestep_reason = 'A' |
---|
309 | ELSE |
---|
310 | timestep_reason = 'D' |
---|
311 | ENDIF |
---|
312 | |
---|
313 | ! |
---|
314 | !-- Set flag if the time step becomes too small. |
---|
315 | IF ( dt_3d < ( 0.00001_wp * dt_max ) ) THEN |
---|
316 | stop_dt = .TRUE. |
---|
317 | |
---|
318 | ! |
---|
319 | !-- Determine the maxima of the diffusion coefficients, including their |
---|
320 | !-- grid index positions. |
---|
321 | CALL global_min_max( nzb, nzt+1, nysg, nyng, nxlg, nxrg, km, 'abs', & |
---|
322 | 0.0_wp, km_max, km_max_ijk ) |
---|
323 | CALL global_min_max( nzb, nzt+1, nysg, nyng, nxlg, nxrg, kh, 'abs', & |
---|
324 | 0.0_wp, kh_max, kh_max_ijk ) |
---|
325 | |
---|
326 | WRITE( message_string, * ) 'Time step has reached minimum limit.', & |
---|
327 | '&dt = ', dt_3d, ' s Simulation is terminated.', & |
---|
328 | '&dt_u = ', dt_u, ' s', & |
---|
329 | '&dt_v = ', dt_v, ' s', & |
---|
330 | '&dt_w = ', dt_w, ' s', & |
---|
331 | '&dt_diff = ', dt_diff, ' s', & |
---|
332 | '&u_max = ', u_max, ' m/s k=', u_max_ijk(1), & |
---|
333 | ' j=', u_max_ijk(2), ' i=', u_max_ijk(3), & |
---|
334 | '&v_max = ', v_max, ' m/s k=', v_max_ijk(1), & |
---|
335 | ' j=', v_max_ijk(2), ' i=', v_max_ijk(3), & |
---|
336 | '&w_max = ', w_max, ' m/s k=', w_max_ijk(1), & |
---|
337 | ' j=', w_max_ijk(2), ' i=', w_max_ijk(3), & |
---|
338 | '&km_max = ', km_max, ' m2/s2 k=', km_max_ijk(1), & |
---|
339 | ' j=', km_max_ijk(2), ' i=', km_max_ijk(3), & |
---|
340 | '&kh_max = ', kh_max, ' m2/s2 k=', kh_max_ijk(1), & |
---|
341 | ' j=', kh_max_ijk(2), ' i=', kh_max_ijk(3) |
---|
342 | CALL message( 'timestep', 'PA0312', 0, 1, 0, 6, 0 ) |
---|
343 | ! |
---|
344 | !-- In case of coupled runs inform the remote model of the termination |
---|
345 | !-- and its reason, provided the remote model has not already been |
---|
346 | !-- informed of another termination reason (terminate_coupled > 0) before. |
---|
347 | #if defined( __parallel ) |
---|
348 | IF ( coupling_mode /= 'uncoupled' .AND. terminate_coupled == 0 ) THEN |
---|
349 | terminate_coupled = 2 |
---|
350 | IF ( myid == 0 ) THEN |
---|
351 | CALL MPI_SENDRECV( & |
---|
352 | terminate_coupled, 1, MPI_INTEGER, target_id, 0, & |
---|
353 | terminate_coupled_remote, 1, MPI_INTEGER, target_id, 0, & |
---|
354 | comm_inter, status, ierr ) |
---|
355 | ENDIF |
---|
356 | CALL MPI_BCAST( terminate_coupled_remote, 1, MPI_INTEGER, 0, & |
---|
357 | comm2d, ierr) |
---|
358 | ENDIF |
---|
359 | #endif |
---|
360 | ENDIF |
---|
361 | |
---|
362 | ! |
---|
363 | !-- In case of nested runs all parent/child processes have to terminate if |
---|
364 | !-- one process has set the stop flag, i.e. they need to set the stop flag |
---|
365 | !-- too. |
---|
366 | IF ( nested_run ) THEN |
---|
367 | stop_dt_local = stop_dt |
---|
368 | #if defined( __parallel ) |
---|
369 | CALL MPI_ALLREDUCE( stop_dt_local, stop_dt, 1, MPI_LOGICAL, MPI_LOR, & |
---|
370 | MPI_COMM_WORLD, ierr ) |
---|
371 | #endif |
---|
372 | ENDIF |
---|
373 | |
---|
374 | ! |
---|
375 | !-- Ensure a smooth value (two significant digits) of the timestep. |
---|
376 | div = 1000.0_wp |
---|
377 | DO WHILE ( dt_3d < div ) |
---|
378 | div = div / 10.0_wp |
---|
379 | ENDDO |
---|
380 | dt_3d = NINT( dt_3d * 100.0_wp / div ) * div / 100.0_wp |
---|
381 | |
---|
382 | ENDIF |
---|
383 | |
---|
384 | ! |
---|
385 | !-- Vertical nesting: coarse and fine grid timestep has to be identical |
---|
386 | IF ( vnested ) CALL vnest_timestep_sync |
---|
387 | |
---|
388 | CALL cpu_log( log_point(12), 'calculate_timestep', 'stop' ) |
---|
389 | |
---|
390 | END SUBROUTINE timestep |
---|