Learning Forum

[ouyangxin]

Hello there, recently I was simulating a 50-micron-long structure with FDTD, but I am doubtful about used settings of the simulation, especially settings related to results accuracy.

First, the default setting about simualtion time is 1000fs and the autoshutoff is 1e-5. I didn't change those settings then started the simulation. After 2 hours, the autoshutoff condition was satisfied and the process was closed.  

I reviewed the log file of simution, about 8.86%  the process was closed. So I am not sure the results is reliable or not. Is there anyway to confirm or realted tutorial,  that the simulated results satify the accuracy? 

 

Thanks in advance.

 

Followed is log file of my simulation:

 

``` 

current time: Fri Apr 21 09:47:13 2023

Starting meshing
Simulation size in gridpoints: 1 x 490 x 229
Refined 0 mesh cells (0%)
Completed meshing!
Meshing complete in 0.21875 seconds of CPU time
Starting meshing
Simulation size in gridpoints: 1 x 490 x 229
Refined 0 mesh cells (0%)
Completed meshing!
Meshing complete in 0 seconds of CPU time
Starting meshing
Simulation size in gridpoints: 1 x 490 x 229
Refined 0 mesh cells (0%)
Completed meshing!
Meshing complete in 0.015625 seconds of CPU time
number of processors is 1
processors are distributed as: 1x1x1
Running 3D Simulation
Starting meshing
Simulation size in gridpoints: 1191 x 504 x 243
Refined 0 mesh cells (0%)
Completed meshing!
Meshing complete in 83.2344 seconds of CPU time
0 0
Starting meshing
Simulation size in gridpoints: 1191 x 504 x 243
Refined 0 mesh cells (0%)
Completed meshing!
Meshing complete in 82.125 seconds of CPU time
0 0
Starting meshing
Simulation size in gridpoints: 1191 x 504 x 243
Refined 0 mesh cells (0%)
Completed meshing!
Meshing complete in 85.1094 seconds of CPU time
0 0
finished material properties
Meshing Time: 34.225 seconds, Init Electrical Material Time: 7.46482 seconds, Init Magnetic Material Time: 0.347179
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 Mode source propagating in X
Starting to add boundary conditions to simulation
Adding 3D PML BC
Adding 3D PML BC
Adding 3D PML BC
Adding 3D PML BC
Adding 3D PML BC
Adding 3D PML BC
Starting to add monitors to Simulation
Adding analysis monitor
Adding analysis monitor
Adding 3D DFT monitor
Adding 3D movie monitor
Adding 3D movie monitor
Finished initialization of 3D Simulation
-----------------------------------------
All processes are communicating and ready to proceed to simulation...
Starting 89918 total iterations
0.0266911% complete. Max time remaining: 17 hrs, 47 mins, 34 secs. Auto Shutoff: 1
0.498234% complete. Max time remaining: 17 hrs, 37 mins, 54 secs. Auto Shutoff: 1
0.967553% complete. Max time remaining: 17 hrs, 34 mins, 30 secs. Auto Shutoff: 1
1.4391% complete. Max time remaining: 17 hrs, 28 mins, 17 secs. Auto Shutoff: 1
1.91286% complete. Max time remaining: 17 hrs, 21 mins, 47 secs. Auto Shutoff: 0.99891
2.38441% complete. Max time remaining: 17 hrs, 16 mins, 37 secs. Auto Shutoff: 0.996123
2.8604% complete. Max time remaining: 17 hrs, 9 mins, 60 secs. Auto Shutoff: 0.9923
3.33639% complete. Max time remaining: 17 hrs, 3 mins, 46 secs. Auto Shutoff: 0.987186
3.80571% complete. Max time remaining: 17 hrs, 4 secs. Auto Shutoff: 0.98056
4.25946% complete. Max time remaining: 16 hrs, 59 mins, 37 secs. Auto Shutoff: 0.975436
4.68652% complete. Max time remaining: 17 hrs, 4 mins, 19 secs. Auto Shutoff: 0.971829
5.15139% complete. Max time remaining: 16 hrs, 59 mins, 28 secs. Auto Shutoff: 0.965228
5.61181% complete. Max time remaining: 16 hrs, 55 mins, 42 secs. Auto Shutoff: 0.959446
6.08558% complete. Max time remaining: 16 hrs, 49 mins, 24 secs. Auto Shutoff: 0.953803
6.53933% complete. Max time remaining: 16 hrs, 46 mins, 28 secs. Auto Shutoff: 0.948919
6.98195% complete. Max time remaining: 16 hrs, 44 mins, 55 secs. Auto Shutoff: 0.783849
7.45794% complete. Max time remaining: 16 hrs, 38 mins, 5 secs. Auto Shutoff: 0.0112362
7.93394% complete. Max time remaining: 16 hrs, 31 mins, 28 secs. Auto Shutoff: 0.00200478
8.39436% complete. Max time remaining: 16 hrs, 27 mins, 11 secs. Auto Shutoff: 4.18526e-05
8.85478% complete. Max time remaining: 16 hrs, 22 mins, 50 secs. Auto Shutoff: 2.1205e-05
100% complete. Max time remaining: 0 mins, 0 secs. Auto Shutoff: 7.2876e-06
Early termination of simulation, the autoshutoff criteria are satisfied.
Completed 8365 iterations, or 3.721187e-13s of Simulation Time
Process 0 starting to collect data
    collecting monitor data
    collecting source data
Finished collecting data
Simulation completed successfully at: Fri Apr 21 11:28:46 2023
Overall wall time measurements in seconds: 6094.39
    time to mesh and initialize: 66.6501
    time to run FDTD simulation: 6024.43
    time to checkpoint simulation: 3e-07
    time to finalize data and save to files: 1.50545
    miscellaneous time: 1.81085
    FDTD solver speed on process 0: 202.534 Mnodes/s
    total FDTD solver speed on 1 processes: 202.534 Mnodes/s

```

 

[Guilin Sun]

Ealier than the preset simulation time usually mean the simulation reaches the autoshutoff min and the result is about good. However, in some circumstances the autoshutoff min might not be small enough. So you will see the transmission/reflection curves are not smooth.

Please refer to :

Ansys Insight: 为什么说仿真时间要足够长才能得到正确的频域结果

Ansys Insight: FDTD仿真过程进度条结束显示100%，这样结果可靠吗？

Ansys Insight: 关于Auto shut off min &auto shut of max

Ansys Insight: 如何知道FDTD的仿真时间足够长

[ouyangxin]

 

Hi, Dr. Sun, thank you so much for your answer.

Here are more details about the purpose of simulation:

A linear sin taper wavegide inside the SiO2 cladding and a Si-based substrate added as well.  The background index of FDTD is setted to 1 refers to air.  The dimension of simulation region is 12um wide and 14um tall, upper boundary of FDTD region contains “air layer”, and lower boundary is inside Si-substrate.   

I want to use FDTD solver to check the after 50-micron-long propagation, if there any radiation loss or absorbtion happens, mainly for TM-mode. Another simulation using the same stucture with EME solver has been conducted. However, from EME results, there is no clear answer to indicate that substrate-absorbtion or radiation loss happen. Because after a taper-length sweep, the transmission could always be convergent. From EME solver, for this taper, the transmission is stablized after 100 microns. FDTD simulation length is setted to half of the stucture, it is just to simplify and reduce the time of simulation. 

 

Follwing the given links, I have double-checked settings in FDTD simulation:

 

1. simulation time, it is changed to 4000fs, that should be enough for my simulation. Meanwhile, the autoshutoff is applied to 1e-5 and the mode source is setted to min/max at 1.54/1.55;
2. source mode, because the simulation mainly focus on wavelength around 1.55um, so the wavelength min/max is 1.54/1.56 um, and from the mode source editor, it only needs few fs to propagate;
3. power monitor, because of the small range of min/max wavelength, the transmission curve of a mode expansion monitor(same location as power monitor at 50-micron-position) looks smooth

 

But I didn’t add time-monitor inside simulation. However from the settings and curve above, It looks like the result is correct.  If the EME results are reliable, which after 100 microns lang the transmission in EME is convergent to nearly 1,  in FDTD  simuation, the transmission is only about 95%. I am not clear about the difference of those two results. 

above is the EME results and follwed is FDTD results with mode expansion：

and this is the T directly exported from power monitor without mode expansion:

I have captured the T from power monitor and T-total/T-forward in same image:

 

Thanks in advance, Dr Sun.