Hi Amin,

The stable time increment in an explicit analysis is related to the wave propagation speed in a certain element, which is, among others, related to the density of the material and element size. What time scaling does is artificially increase the density of the material. Now time scaling algorithms have more sophisticated features like non-uniformly increasing the density of elements according to the element size etc. So if you use time scaling in an impact analysis, basically you increase the mass of the structure, which will change the dynamic behavior of the problem. In a dynamic problem this will definitely have an effect on your result. If you are using time scaling to reduce the analysis time because you want to improve your model and don't want to wait days for a single result, I think it would be better to run a coarse mesh first (without time scaling) and when you are confident with the model, increase the mesh density to get the final desired result. 

The problem of abaqus terminating the job prematurely while using time scaling is also related to the stress wave propagation. In the status file (.sta) generally an error is issued stating something like "The ratio of deformation speed to wave speed exceeds 1.0000 in at least one element....", which is a result of unrealistic wave propagation due to mass scaling.

Regards,

Farid 

I think you can compare the result with some experimental results to know witch condition is near to reality. 

I would suggest initialy try to solve the problem with smaller time increment. for example, perform a frequncy analysis to retrieve the natural frequencies of the construction and use as Dt=1/(max(NF))/8. if the problem still doesnt converge then try to change the time scalling factor.

According to abaqus help,  time scaling factor is acceptable for explicit analysis if kinetic energy is one-tenth(0.1) of total energy in job results table. thus you have to increase the time scaling factor until error in job. 

Dear Reza, thanks a lot for your response, can you address me to it in the help, I haven't seen it.

Please check it >>>>>>

Abaqus analysis user's manual

6. analysis procedures

6.3. dynamic stress/displacement analysis

6.3.3. explicit dynamic

computational cost

--------------------------------------

In a quasi-static analysis it is expedient to reduce the computational cost by either speeding up the simulation or by scaling the mass. In either case the kinetic energy should be monitored to ensure that the ratio of kinetic energy to internal energy does not get too large—typically less than 10%.

--------------------------------------

Good Luck

See the following paper

Effect of Time scaling and Mass scaling in Numerical Simulation of

Incremental Forming - Applied Mechanics and Materials Vol. 612 (2014) pp 105-110

Hi Amin,

The stable time increment in an explicit analysis is related to the wave propagation speed in a certain element, which is, among others, related to the density of the material and element size. What time scaling does is artificially increase the density of the material. Now time scaling algorithms have more sophisticated features like non-uniformly increasing the density of elements according to the element size etc. So if you use time scaling in an impact analysis, basically you increase the mass of the structure, which will change the dynamic behavior of the problem. In a dynamic problem this will definitely have an effect on your result. If you are using time scaling to reduce the analysis time because you want to improve your model and don't want to wait days for a single result, I think it would be better to run a coarse mesh first (without time scaling) and when you are confident with the model, increase the mesh density to get the final desired result. 

The problem of abaqus terminating the job prematurely while using time scaling is also related to the stress wave propagation. In the status file (.sta) generally an error is issued stating something like "The ratio of deformation speed to wave speed exceeds 1.0000 in at least one element....", which is a result of unrealistic wave propagation due to mass scaling.

Regards,

Farid 

You said you ran with different velocity (medium impact)so its obvious you get different results or you just need to correct your question

Dear Reza Naseri, As I you have mentioned in one of your comment regarding the Abaqus manual, kinetic energy should be less than 10% of the internal energy or the energy due to external work for quasi-static analysis. It is NOT 10% of the total energy! Please correct me if I have got it wrong. Thanks!

Dear Amin,

You asked which time scaling factor should we use? My answer is: it depends. First, you must choose a value for your time step period, and then a value of mass scaling factor. If you choose a time period of 1 second, applying a time scaling of 1.10^-5 is a good start. But for example, if you choose a time period of 1500 seconds, you must use a higher time scaling like 0.1 or 0.2 (Please refer to the Figures 1 and 2, attached below), otherwise you will increase significantly your computational costs during the analysis.

As Mr. Reza Nazeri had mentioned, after the analysis, you must always check if the kinetic energy is considerably lower than the internal energy of your problem (some sources consider that kinetic energy must be around 1-2% of the internal energy). You can find these variables in the ODB filed output menu, looking for ALLIE (Internal energy for whole model) and ALLKE (value of the kinetic energy), as it can be seen in the attached Figures 3 and 4.

Regards,

Hermano.