Speed up calculation time DEM

jeroendemto · October 2, 2018, 05:17

Hey everyone!

I'm trying to simulate particles with a size of 3*10^-5 diameter and a youngh's modulus of 2 Gpa. They flow through a tube filled with liquid. The amount of particles I try to simulate are around 20,000. The velocity magnitude is not that high, around 0.3 m/s. The time step I use for the rest of the calculations is 10^-3s.

All this combined makes my calculations really slow because of the substepping. Will increasing the substepping time influence my calculations much?

I thought the program takes a really small substepping time because of the rayleigh wave speed. However, my particles don't interact much with eachother, only sometimes with the wall. Which would mean increasing the substepping time manually does not influence my calculations much?

I hope I gave enough information, if I did not feel free to ask.

kind regards,
Jeroen

Cassian · October 14, 2022, 15:58

Hello I know its years later, but has anyone found any solutions or tips to increase the DEM simulation speed (other than decreasing the Young's modulus)?

Bovy · November 16, 2022, 09:20

The interactons are important, for example play whit normal spring stiffness if it is too small the particle go trow the wall like a phantom and change the sub step times.... use the more simple model and then try to change for other.

or try to use hertz mindlin and the only variable important is the young moduls ....

the "mesh" of the geometry is important more triangles more interactions more time... in meshless too

October 2, 2018, 05:17	Speed up calculation time DEM	#1
jeroendemto New Member Join Date: Sep 2018 Posts: 25 Rep Power: 8	Hey everyone! I'm trying to simulate particles with a size of 3*10^-5 diameter and a youngh's modulus of 2 Gpa. They flow through a tube filled with liquid. The amount of particles I try to simulate are around 20,000. The velocity magnitude is not that high, around 0.3 m/s. The time step I use for the rest of the calculations is 10^-3s. All this combined makes my calculations really slow because of the substepping. Will increasing the substepping time influence my calculations much? I thought the program takes a really small substepping time because of the rayleigh wave speed. However, my particles don't interact much with eachother, only sometimes with the wall. Which would mean increasing the substepping time manually does not influence my calculations much? I hope I gave enough information, if I did not feel free to ask. kind regards, Jeroen TommyM likes this.

October 14, 2022, 15:58	Any solutions?	#2
Cassian New Member Cassian Join Date: May 2022 Posts: 5 Rep Power: 4	Hello I know its years later, but has anyone found any solutions or tips to increase the DEM simulation speed (other than decreasing the Young's modulus)?

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Elastic or Plastic Deformations in OpenFOAM	NickolasPl	OpenFOAM Pre-Processing	36	September 23, 2023 09:22
AMI speed performance	danny123	OpenFOAM	21	October 24, 2020 05:13
LES, Courant Number, Crash, Sudden	Alhasan	OpenFOAM Running, Solving & CFD	5	November 22, 2019 03:05
Micro Scale Pore, icoFoam	gooya_kabir	OpenFOAM Running, Solving & CFD	2	November 2, 2013 14:58
dynamic Mesh is faster than MRF????	sharonyue	OpenFOAM Running, Solving & CFD	14	August 26, 2013 08:47

November 16, 2022, 09:20		#3
Bovy New Member Valparaiso Join Date: Dec 2019 Posts: 3 Rep Power: 6	The interactons are important, for example play whit normal spring stiffness if it is too small the particle go trow the wall like a phantom and change the sub step times.... use the more simple model and then try to change for other. or try to use hertz mindlin and the only variable important is the young moduls .... the "mesh" of the geometry is important more triangles more interactions more time... in meshless too