Home

Context Navigation

source: palm/trunk/SOURCE/timestep.f90 @ 3248

Last change on this file since 3248 was 3241, checked in by raasch, 6 years ago
various changes to avoid compiler warnings (mainly removal of unused variables)
Property svn:keywords set to `Id`
File size: 16.1 KB

Line
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 1997-2018 Leibniz Universitaet Hannover
18	!------------------------------------------------------------------------------!
19	!
20	! Current revisions:
21	! ------------------
22	!
23	!
24	! Former revisions:
25	! -----------------
26	! $Id: timestep.f90 3241 2018-09-12 15:02:00Z sward $
27	! unused variables removed
28	!
29	! 3120 2018-07-11 18:30:57Z gronemeier
30	! Put ABS( km ) in computation of time step according to the diffusion criterion
31	!
32	! 3084 2018-06-19 15:30:55Z gronemeier
33	! limit increase of dt_3d only in case of RANS mode
34	!
35	! 3083 2018-06-19 14:03:12Z gronemeier
36	! limit dt_3d to be at maximum 2*old_dt; define old_dt at beginning of routine
37	! Add km/kh_max
38	!
39	! 3049 2018-05-29 13:52:36Z Giersch
40	! Error messages revised
41	!
42	! 3045 2018-05-28 07:55:41Z Giersch
43	! Error message revised
44	!
45	! 2718 2018-01-02 08:49:38Z maronga
46	! Corrected "Former revisions" section
47	!
48	! 2696 2017-12-14 17:12:51Z kanani
49	! Change in file header (GPL part)
50	!
51	! 2365 2017-08-21 14:59:59Z kanani
52	! Vertical grid nesting: Sync fine and coarse grid timestep (SadiqHuq)
53	!
54	! 2258 2017-06-08 07:55:13Z suehring
55	! Bugfix, add pre-preprocessor directives to enable non-parrallel mode
56	!
57	! 2168 2017-03-06 13:08:38Z suehring
58	!
59	! 2130 2017-01-24 16:25:39Z raasch
60	! bugfix: in case of nested runs the stop condition in case of too small
61	! timesteps is communicated to all parent/child processes,
62	! some formatting done
63	!
64	! 2118 2017-01-17 16:38:49Z raasch
65	! OpenACC directives and related part of code removed
66	!
67	! 2000 2016-08-20 18:09:15Z knoop
68	! Forced header and separation lines into 80 columns
69	!
70	! 1849 2016-04-08 11:33:18Z hoffmann
71	! Adapted for modularization of microphysics
72	!
73	! 1682 2015-10-07 23:56:08Z knoop
74	! Code annotations made doxygen readable
75	!
76	! 1484 2014-10-21 10:53:05Z kanani
77	! Changes due to new module structure of the plant canopy model:
78	! calculations and parameters related to the plant canopy model removed
79	! (the limitation of the canopy drag, i.e. that the canopy drag itself should
80	! not change the sign of the velocity components, is now assured for in the
81	! calculation of the canopy tendency terms in subroutine plant_canopy_model)
82	!
83	! 1342 2014-03-26 17:04:47Z kanani
84	! REAL constants defined as wp-kind
85	!
86	! 1322 2014-03-20 16:38:49Z raasch
87	! REAL functions provided with KIND-attribute
88	!
89	! 1320 2014-03-20 08:40:49Z raasch
90	! ONLY-attribute added to USE-statements,
91	! kind-parameters added to all INTEGER and REAL declaration statements,
92	! kinds are defined in new module kinds,
93	! old module precision_kind is removed,
94	! revision history before 2012 removed,
95	! comment fields (!:) to be used for variable explanations added to
96	! all variable declaration statements
97	!
98	! 1257 2013-11-08 15:18:40Z raasch
99	! openacc porting
100	! bugfix for calculation of advective timestep in case of vertically stretched
101	! grids
102	!
103	! 1092 2013-02-02 11:24:22Z raasch
104	! unused variables removed
105	!
106	! 1053 2012-11-13 17:11:03Z hoffmann
107	! timestep is reduced in two-moment cloud scheme according to the maximum
108	! terminal velocity of rain drops
109	!
110	! 1036 2012-10-22 13:43:42Z raasch
111	! code put under GPL (PALM 3.9)
112	!
113	! 1001 2012-09-13 14:08:46Z raasch
114	! all actions concerning leapfrog scheme removed
115	!
116	! 978 2012-08-09 08:28:32Z fricke
117	! restriction of the outflow damping layer in the diffusion criterion removed
118	!
119	! 866 2012-03-28 06:44:41Z raasch
120	! bugfix for timestep calculation in case of Galilei transformation,
121	! special treatment in case of mirror velocity boundary condition removed
122	!
123	! Revision 1.1 1997/08/11 06:26:19 raasch
124	! Initial revision
125	!
126	!
127	! Description:
128	! ------------
129	!> Compute the time step under consideration of the FCL and diffusion criterion.
130	!------------------------------------------------------------------------------!
131	SUBROUTINE timestep
132
133
134	USE arrays_3d, &
135	ONLY: dzu, dzw, kh, km, u, v, w
136
137	USE control_parameters, &
138	ONLY: cfl_factor, coupling_mode, dt_3d, dt_fixed, dt_max, &
139	galilei_transformation, old_dt, message_string, rans_mode, &
140	stop_dt, terminate_coupled, terminate_coupled_remote, &
141	timestep_reason, u_gtrans, use_ug_for_galilei_tr, v_gtrans
142
143	USE cpulog, &
144	ONLY: cpu_log, log_point
145
146	USE grid_variables, &
147	ONLY: dx, dx2, dy, dy2
148
149	USE indices, &
150	ONLY: nxl, nxlg, nxr, nxrg, nyn, nyng, nys, nysg, nzb, nzt
151
152	USE interfaces
153
154	USE kinds
155
156	USE microphysics_mod, &
157	ONLY: dt_precipitation
158
159	USE pegrid
160
161	USE pmc_interface, &
162	ONLY: nested_run
163
164	USE statistics, &
165	ONLY: flow_statistics_called, hom, u_max, u_max_ijk, v_max, v_max_ijk,&
166	w_max, w_max_ijk
167
168	USE vertical_nesting_mod, &
169	ONLY: vnested, vnest_timestep_sync
170
171	IMPLICIT NONE
172
173	INTEGER(iwp) :: i !<
174	INTEGER(iwp) :: j !<
175	INTEGER(iwp) :: k !<
176	INTEGER(iwp) :: km_max_ijk(3) = -1 !< index values (i,j,k) of location where km_max occurs
177	INTEGER(iwp) :: kh_max_ijk(3) = -1 !< index values (i,j,k) of location where kh_max occurs
178
179	LOGICAL :: stop_dt_local !< local switch for controlling the time stepping
180
181	REAL(wp) :: div !<
182	REAL(wp) :: dt_diff !<
183	REAL(wp) :: dt_diff_l !<
184	REAL(wp) :: dt_u !<
185	REAL(wp) :: dt_u_l !<
186	REAL(wp) :: dt_v !<
187	REAL(wp) :: dt_v_l !<
188	REAL(wp) :: dt_w !<
189	REAL(wp) :: dt_w_l !<
190	REAL(wp) :: km_max !< maximum of Km in entire domain
191	REAL(wp) :: kh_max !< maximum of Kh in entire domain
192	REAL(wp) :: u_gtrans_l !<
193	REAL(wp) :: v_gtrans_l !<
194
195	REAL(wp), DIMENSION(2) :: uv_gtrans !<
196	REAL(wp), DIMENSION(2) :: uv_gtrans_l !<
197	REAL(wp), DIMENSION(3) :: reduce !<
198	REAL(wp), DIMENSION(3) :: reduce_l !<
199	REAL(wp), DIMENSION(nzb+1:nzt) :: dxyz2_min !<
200
201
202	CALL cpu_log( log_point(12), 'calculate_timestep', 'start' )
203	!
204	!-- Save former time step as reference
205	old_dt = dt_3d
206
207	!
208	!-- In case of Galilei-transform not using the geostrophic wind as translation
209	!-- velocity, compute the volume-averaged horizontal velocity components, which
210	!-- will then be subtracted from the horizontal wind for the time step and
211	!-- horizontal advection routines.
212	IF ( galilei_transformation .AND. .NOT. use_ug_for_galilei_tr ) THEN
213	IF ( flow_statistics_called ) THEN
214	!
215	!-- Horizontal averages already existent, just need to average them
216	!-- vertically.
217	u_gtrans = 0.0_wp
218	v_gtrans = 0.0_wp
219	DO k = nzb+1, nzt
220	u_gtrans = u_gtrans + hom(k,1,1,0)
221	v_gtrans = v_gtrans + hom(k,1,2,0)
222	ENDDO
223	u_gtrans = u_gtrans / REAL( nzt - nzb, KIND=wp )
224	v_gtrans = v_gtrans / REAL( nzt - nzb, KIND=wp )
225	ELSE
226	!
227	!-- Averaging over the entire model domain.
228	u_gtrans_l = 0.0_wp
229	v_gtrans_l = 0.0_wp
230	DO i = nxl, nxr
231	DO j = nys, nyn
232	DO k = nzb+1, nzt
233	u_gtrans_l = u_gtrans_l + u(k,j,i)
234	v_gtrans_l = v_gtrans_l + v(k,j,i)
235	ENDDO
236	ENDDO
237	ENDDO
238	uv_gtrans_l(1) = u_gtrans_l / &
239	REAL( (nxr-nxl+1)(nyn-nys+1)(nzt-nzb), KIND=wp )
240	uv_gtrans_l(2) = v_gtrans_l / &
241	REAL( (nxr-nxl+1)(nyn-nys+1)(nzt-nzb), KIND=wp )
242	#if defined( __parallel )
243	IF ( collective_wait ) CALL MPI_BARRIER( comm2d, ierr )
244	CALL MPI_ALLREDUCE( uv_gtrans_l, uv_gtrans, 2, MPI_REAL, MPI_SUM, &
245	comm2d, ierr )
246	u_gtrans = uv_gtrans(1) / REAL( numprocs, KIND=wp )
247	v_gtrans = uv_gtrans(2) / REAL( numprocs, KIND=wp )
248	#else
249	u_gtrans = uv_gtrans_l(1)
250	v_gtrans = uv_gtrans_l(2)
251	#endif
252	ENDIF
253	ENDIF
254
255	!
256	!-- Determine the maxima of the velocity components, including their
257	!-- grid index positions.
258	CALL global_min_max( nzb, nzt+1, nysg, nyng, nxlg, nxrg, u, 'abs', 0.0_wp, &
259	u_max, u_max_ijk )
260	CALL global_min_max( nzb, nzt+1, nysg, nyng, nxlg, nxrg, v, 'abs', 0.0_wp, &
261	v_max, v_max_ijk )
262	CALL global_min_max( nzb, nzt+1, nysg, nyng, nxlg, nxrg, w, 'abs', 0.0_wp, &
263	w_max, w_max_ijk )
264
265	IF ( .NOT. dt_fixed ) THEN
266	!
267	!-- Variable time step:
268	!-- Calculate the maximum time step according to the CFL-criterion,
269	!-- individually for each velocity component
270	dt_u_l = 999999.9_wp
271	dt_v_l = 999999.9_wp
272	dt_w_l = 999999.9_wp
273	DO i = nxl, nxr
274	DO j = nys, nyn
275	DO k = nzb+1, nzt
276	dt_u_l = MIN( dt_u_l, ( dx / &
277	( ABS( u(k,j,i) - u_gtrans ) + 1.0E-10_wp ) ) )
278	dt_v_l = MIN( dt_v_l, ( dy / &
279	( ABS( v(k,j,i) - v_gtrans ) + 1.0E-10_wp ) ) )
280	dt_w_l = MIN( dt_w_l, ( dzu(k) / &
281	( ABS( w(k,j,i) ) + 1.0E-10_wp ) ) )
282	ENDDO
283	ENDDO
284	ENDDO
285
286	#if defined( __parallel )
287	reduce_l(1) = dt_u_l
288	reduce_l(2) = dt_v_l
289	reduce_l(3) = dt_w_l
290	IF ( collective_wait ) CALL MPI_BARRIER( comm2d, ierr )
291	CALL MPI_ALLREDUCE( reduce_l, reduce, 3, MPI_REAL, MPI_MIN, comm2d, ierr )
292	dt_u = reduce(1)
293	dt_v = reduce(2)
294	dt_w = reduce(3)
295	#else
296	dt_u = dt_u_l
297	dt_v = dt_v_l
298	dt_w = dt_w_l
299	#endif
300
301	!
302	!-- Compute time step according to the diffusion criterion.
303	!-- First calculate minimum grid spacing which only depends on index k.
304	!-- When using the dynamic subgrid model, negative km are possible.
305	dt_diff_l = 999999.0_wp
306
307	DO k = nzb+1, nzt
308	dxyz2_min(k) = MIN( dx2, dy2, dzw(k)dzw(k) ) 0.125_wp
309	ENDDO
310
311	!$OMP PARALLEL private(i,j,k) reduction(MIN: dt_diff_l)
312	!$OMP DO
313	DO i = nxl, nxr
314	DO j = nys, nyn
315	DO k = nzb+1, nzt
316	dt_diff_l = MIN( dt_diff_l, dxyz2_min(k) / &
317	( MAX( kh(k,j,i), ABS( km(k,j,i) ) ) &
318	+ 1E-20_wp ) )
319	ENDDO
320	ENDDO
321	ENDDO
322	!$OMP END PARALLEL
323	#if defined( __parallel )
324	IF ( collective_wait ) CALL MPI_BARRIER( comm2d, ierr )
325	CALL MPI_ALLREDUCE( dt_diff_l, dt_diff, 1, MPI_REAL, MPI_MIN, comm2d, &
326	ierr )
327	#else
328	dt_diff = dt_diff_l
329	#endif
330
331	!
332	!-- The time step is the minimum of the 3-4 components and the diffusion time
333	!-- step minus a reduction (cfl_factor) to be on the safe side.
334	!-- The time step must not exceed the maximum allowed value.
335	dt_3d = cfl_factor * MIN( dt_diff, dt_u, dt_v, dt_w, dt_precipitation )
336	dt_3d = MIN( dt_3d, dt_max )
337	!
338	!-- In RANS mode, the time step must not increase by more than a factor of 2
339	IF ( rans_mode ) dt_3d = MIN( dt_3d, dt_max, 2.0_wp * old_dt )
340
341	!
342	!-- Remember the restricting time step criterion for later output.
343	IF ( MIN( dt_u, dt_v, dt_w ) < dt_diff ) THEN
344	timestep_reason = 'A'
345	ELSE
346	timestep_reason = 'D'
347	ENDIF
348
349	!
350	!-- Set flag if the time step becomes too small.
351	IF ( dt_3d < ( 0.00001_wp * dt_max ) ) THEN
352	stop_dt = .TRUE.
353
354	!
355	!-- Determine the maxima of the diffusion coefficients, including their
356	!-- grid index positions.
357	CALL global_min_max( nzb, nzt+1, nysg, nyng, nxlg, nxrg, km, 'abs', &
358	0.0_wp, km_max, km_max_ijk )
359	CALL global_min_max( nzb, nzt+1, nysg, nyng, nxlg, nxrg, kh, 'abs', &
360	0.0_wp, kh_max, kh_max_ijk )
361
362	WRITE( message_string, * ) 'Time step has reached minimum limit.', &
363	'&dt = ', dt_3d, ' s Simulation is terminated.', &
364	'&old_dt = ', old_dt, ' s', &
365	'&dt_u = ', dt_u, ' s', &
366	'&dt_v = ', dt_v, ' s', &
367	'&dt_w = ', dt_w, ' s', &
368	'&dt_diff = ', dt_diff, ' s', &
369	'&u_max = ', u_max, ' m/s k=', u_max_ijk(1), &
370	' j=', u_max_ijk(2), ' i=', u_max_ijk(3), &
371	'&v_max = ', v_max, ' m/s k=', v_max_ijk(1), &
372	' j=', v_max_ijk(2), ' i=', v_max_ijk(3), &
373	'&w_max = ', w_max, ' m/s k=', w_max_ijk(1), &
374	' j=', w_max_ijk(2), ' i=', w_max_ijk(3), &
375	'&km_max = ', km_max, ' m2/s2 k=', km_max_ijk(1), &
376	' j=', km_max_ijk(2), ' i=', km_max_ijk(3), &
377	'&kh_max = ', kh_max, ' m2/s2 k=', kh_max_ijk(1), &
378	' j=', kh_max_ijk(2), ' i=', kh_max_ijk(3)
379	CALL message( 'timestep', 'PA0312', 0, 1, 0, 6, 0 )
380	!
381	!-- In case of coupled runs inform the remote model of the termination
382	!-- and its reason, provided the remote model has not already been
383	!-- informed of another termination reason (terminate_coupled > 0) before.
384	#if defined( __parallel )
385	IF ( coupling_mode /= 'uncoupled' .AND. terminate_coupled == 0 ) THEN
386	terminate_coupled = 2
387	IF ( myid == 0 ) THEN
388	CALL MPI_SENDRECV( &
389	terminate_coupled, 1, MPI_INTEGER, target_id, 0, &
390	terminate_coupled_remote, 1, MPI_INTEGER, target_id, 0, &
391	comm_inter, status, ierr )
392	ENDIF
393	CALL MPI_BCAST( terminate_coupled_remote, 1, MPI_INTEGER, 0, &
394	comm2d, ierr)
395	ENDIF
396	#endif
397	ENDIF
398
399	!
400	!-- In case of nested runs all parent/child processes have to terminate if
401	!-- one process has set the stop flag, i.e. they need to set the stop flag
402	!-- too.
403	IF ( nested_run ) THEN
404	stop_dt_local = stop_dt
405	#if defined( __parallel )
406	CALL MPI_ALLREDUCE( stop_dt_local, stop_dt, 1, MPI_LOGICAL, MPI_LOR, &
407	MPI_COMM_WORLD, ierr )
408	#endif
409	ENDIF
410
411	!
412	!-- Ensure a smooth value (two significant digits) of the timestep.
413	div = 1000.0_wp
414	DO WHILE ( dt_3d < div )
415	div = div / 10.0_wp
416	ENDDO
417	dt_3d = NINT( dt_3d * 100.0_wp / div ) * div / 100.0_wp
418
419	ENDIF
420
421	!
422	!-- Vertical nesting: coarse and fine grid timestep has to be identical
423	IF ( vnested ) CALL vnest_timestep_sync
424
425	CALL cpu_log( log_point(12), 'calculate_timestep', 'stop' )
426
427	END SUBROUTINE timestep

Note: See TracBrowser for help on using the repository browser.

Download in other formats:

| Impressum | ©Leibniz Universität Hannover |