EdgeSmooth weird behavior
7 views
Skip to first unread message
Aurel
4:30 AM (15 hours ago) 4:30 AM
to mumax2
Dear all,
I am currently using the Mumax3 software to simulate nanowires with circular cross-section. I am exploring different materials parameters and varying diameters. Due to the circular geometry, I performed simulations with the EdgeSmooth option to analyze the impact on the energy of my system and assess its effectiveness, and an appropriate value for the EdgeSmooth interpolation value.
I had expected a sizable relative decrease in energy for small diameters, and lower decrease for larger diameters, for a given cell size. However, while this is the general trend, an increase in energy occurs for certain diameters (see figures).
This increase in energy appears to be reproducible. I have attempted this process several times, using the following parameters:
Msat = 8e5 A/m
Aex = 1e-11 J/m
L = 1 µm
Cell size = 1x1x1 nm
Diameters are between 20 and 50 nm.
I used Minimize() and Relax() to run my simulations, yielding similar results.
I visualized my wires to see if any issues would become apparent during the simulation. Nothing seemed wrong when I looked at the different wires.
Additionally, simulations with larger cells were conducted to emphasize the impact of EdgeSmooth on the energy of the system, for further testing, although I would not use such large cells for my final simulation. A peak (energy maximum) versus EdgeSmooth is also present for a diameter of 28nm (see figures).
The figures represent the percentage of energy variation compared to EdgeSmooth = 0 as a function of diameter for different EdgeSmooth values. The first figure was run with a cell size of 1 nm, and the second with a cell size of 4 nm.
I understand that the EdgeSmooth option isn't perfect. It might not be very effective for 3D systems, and it could lead to problems related to exchange. Even so, I was surprised by the non-monotonic behavior that resulted from only small changes in system size.
Has anybody had some experience with a similar behavior, and could contribute to explain this increase in energy?
Thank you,
A Mumax3 user
I am currently using the Mumax3 software to simulate nanowires with circular cross-section. I am exploring different materials parameters and varying diameters. Due to the circular geometry, I performed simulations with the EdgeSmooth option to analyze the impact on the energy of my system and assess its effectiveness, and an appropriate value for the EdgeSmooth interpolation value.
I had expected a sizable relative decrease in energy for small diameters, and lower decrease for larger diameters, for a given cell size. However, while this is the general trend, an increase in energy occurs for certain diameters (see figures).
This increase in energy appears to be reproducible. I have attempted this process several times, using the following parameters:
Msat = 8e5 A/m
Aex = 1e-11 J/m
L = 1 µm
Cell size = 1x1x1 nm
Diameters are between 20 and 50 nm.
I used Minimize() and Relax() to run my simulations, yielding similar results.
I visualized my wires to see if any issues would become apparent during the simulation. Nothing seemed wrong when I looked at the different wires.
Additionally, simulations with larger cells were conducted to emphasize the impact of EdgeSmooth on the energy of the system, for further testing, although I would not use such large cells for my final simulation. A peak (energy maximum) versus EdgeSmooth is also present for a diameter of 28nm (see figures).
The figures represent the percentage of energy variation compared to EdgeSmooth = 0 as a function of diameter for different EdgeSmooth values. The first figure was run with a cell size of 1 nm, and the second with a cell size of 4 nm.
I understand that the EdgeSmooth option isn't perfect. It might not be very effective for 3D systems, and it could lead to problems related to exchange. Even so, I was surprised by the non-monotonic behavior that resulted from only small changes in system size.
Has anybody had some experience with a similar behavior, and could contribute to explain this increase in energy?
Thank you,
A Mumax3 user
Reply all
Reply to author
Forward
0 new messages