有关auto shutoff min 的问题

autoshutoff
fdtd

#1

老师,您好。
我首先用5nm网格来模拟auto shutoff min :1e-6,但是它在auto shutoff min :1e-5振荡了很久后,程序运行到30000fs就停了,
用2nm网格来模拟auto shutoff min :1e-6,但是它在auto shutoff min :1e-5振荡了一会,程序运行到18000fs就停了。
两个结果相差挺大的,那它们哪个才比较准确呢?
谢谢!!!


#2

没有源文件以及Log文件,很难说哪个准确。
这种振荡很可能是PML引起的,你加大仿真区,用Meshaccuracy2,需要细网格的地方用细化网格(Override),设置相同的仿真时间,例如35000fs,然后再比较。加大仿真区的目的是为了保证用不同的细化网格时,PML所在位置的网格尺寸不变,否则更难诊断。
当PML和其它设置没有问题时,一般是细网格的精度高。


#3

老师,我的PML层离结构都快1.5微米远了。我的结构有6微米大。是因为我结果里的谐振太强了,电场都衰减完了,但是谐振还是很强。原文件在超算都要算一天,太大了。


#4

设置仿真时间没什么用,我的仿真时间足够长的。它怎么会一直在还没达到auto shutoff min就停下来了,而且我设置10-6,在10-5就停了。


#5

没有源文件难以找到问题的根源,请参看这个帖子给与所有信息

如果是结构本身在仿真波段的谐振很强,一般也用不了一天的时间。很可能是光源波长之外的谐振,你可以添加时间监视器看频谱找谐振,要针对原因才能解决问题。

如果设置10-6,在10-5就停了,说明是达到10-6啦,你查看一些Log文件。


#6

我看的就是Log文件,它并没有达到就提前停下来的。添加时间监视器看频谱找谐振,这是怎样找的?谢谢


#7

FDTD Solutions Version 8.12.590 (Linux 64bit)
expires (month/day/year): 12/3/2017
current time: Sun Jun 25 21:33:22 2017
Running on host: cn11194
All processes are communicating and ready to proceed to simulation…
Running fsp file: di19_mesh2.fsp
number of processors is 480
processors are distributed as: 10x12x4
Running 3D Simulation
Starting meshing
Simulation size in gridpoints: 70 x 58 x 66
Refined 4238 mesh cells (1.58158%)
Completed meshing!
Meshing complete in 0.22 seconds of CPU time
Starting meshing
Simulation size in gridpoints: 70 x 58 x 66
Refined 4242 mesh cells (1.58307%)
Completed meshing!
Meshing complete in 0.19 seconds of CPU time
Starting meshing
Simulation size in gridpoints: 70 x 58 x 66
Refined 4246 mesh cells (1.58456%)
Completed meshing!
Meshing complete in 0.18 seconds of CPU time
finished material properties
Using selected optimization level: 2
Selecting 3d kernel for splitFieldDirection=0 and optLevel=2

Beginning initialization of 3D Simulation
Starting to add sources to simulation
Adding 3D point Source
Starting to add boundary conditions to simulation
Adding 3D PML BC
Adding 3D Parallel BC
Adding 3D PML BC
Adding 3D Parallel BC
Adding 3D PML BC
Adding 3D Parallel BC
Starting to add monitors to Simulation
Adding analysis monitor
Adding 3D Time monitor
Finished initialization of 3D Simulation

All processes are communicating and ready to proceed to simulation…
Starting 52450078 total iterations
0% complete. Max time remaining: 919 hrs, 26 mins, 12 secs. Auto Shutoff: 1
1% complete. Max time remaining: 430 hrs, 56 mins, 26 secs. Auto Shutoff: 0.104973
2% complete. Max time remaining: 424 hrs, 13 secs. Auto Shutoff: 0.0577912
3% complete. Max time remaining: 419 hrs, 47 mins, 12 secs. Auto Shutoff: 0.0324062
4% complete. Max time remaining: 415 hrs, 4 mins, 16 secs. Auto Shutoff: 0.0146229
5% complete. Max time remaining: 412 hrs, 4 mins, 40 secs. Auto Shutoff: 0.00537093
6% complete. Max time remaining: 410 hrs, 2 mins, 27 secs. Auto Shutoff: 0.00107564
7% complete. Max time remaining: 407 hrs, 17 mins, 36 secs. Auto Shutoff: 4.31727e-05
8% complete. Max time remaining: 404 hrs, 21 mins, 3 secs. Auto Shutoff: 0.000426475
9% complete. Max time remaining: 400 hrs, 37 mins, 30 secs. Auto Shutoff: 0.00108991
10% complete. Max time remaining: 399 hrs, 18 mins, 56 secs. Auto Shutoff: 0.00150619
11% complete. Max time remaining: 394 hrs, 27 mins, 27 secs. Auto Shutoff: 0.00154255
12% complete. Max time remaining: 398 hrs, 27 mins, 46 secs. Auto Shutoff: 0.000399
Early termination of simulation, the autoshutoff criteria are satisfied.
Completed 6621892 iterations, or 2.52503e-11s of Simulation Time
Process 0 starting to collect data
collecting monitor data
collecting source data
Finished collecting data
Simulation completed successfully at: Wed Jun 28 06:44:04 2017
Overall wall time measurements in seconds:
time to mesh and initialize: 8
time to run FDTD simulation: 205820
time to finalize data and save to files: 12
FDTD solver speed: 8.62114 Mnodes/s
total FDTD solver speed on 480 processes: 4138.15 Mnodes/s


#8

这句话说明已经达到了你预先设置的autoshutoff;

用Visualizer察看时间监视器的SPectrum就可以。


#9

老师,时间监视器的光谱情况就是这样的,我需要的卸载是900那个,意思就是他在500-800直接有很多谐振吗?那我应该怎么办?
谢谢老师!


#10

但是我设置的是10-6啊

12% complete. Max time remaining: 398 hrs, 27 mins, 46 secs. Auto Shutoff: 0.000399
Early termination of simulation, the autoshutoff criteria are satisfied.
它这不是在10-4就停了吗,差这么远。


#11

你可以将光源的光谱范围设置为850 到950纳米,不需要的波长没有必要包含在光源中。[quote=“liuzhj, post:10, topic:7112”]
它这不是在10-4就停了吗,差这么远
[/quote]
400
不是,是达到了你预先设置的。你看看,软件估计需要400小时,大概每 4小时显示一次结果。但是在这 4小时内就达到了 autoshutoff ,上次不是说了,重点看另外一句:
#Early termination of simulation, the autoshutoff criteria are satisfied.

这就够了,不要指望它再给你显示1E-6。

如果每次都是这样,你可以减小仿真时间的设置。


#12


仅供参考


#13

我本来就是这样的设置了,但是他还是这样的情况出现。


#14

你说的“这样的情况出现”,是指“它并没有达到就提前停下来的”吗?
如果是,请在参考昨天的回复,不理解的地方请指出来。这种提示每1%才显示一次,不是每个仿真时间上都显示。