I am trying to simulate punching of tube process in Abaqus. Speed of punch is 10 mm/min and it have to travel total 7 mm including the thickness of tube (1.5 mm).
The time period in Explicit simulation is the time within which you want to apply the defined load and b.c's. Say, you have 100kN and you have specified a time step of 1, then all the 100kN is applied within the Time period of 1. If you specify Time period = 2, then all the 100kN is applied in the Time period = 2. The only thing which changes in the % of load that you are applying in each increment.
Say you are at an increment = 0.1 with time period = 1, then the % of load applied = 10%.
The same, if you are at an increment = 0.1 with time period = 2, then the & of load applied = 5%.
The time increment in ABAQUS Explicit is auto decided by ABAQUS normally (you can also choose to override this and define your own time increment) and the only condition that should be noted is that this time increment should be less than the Stability Limit. Thus, the only consideration which you should do in an ABAQUS Explicit analysis is to make sure that the time increment is lesser than the Stability limit. The stability time increment is based upon Wave speed, Characteristic Element Length and Young's modulus of the material.
The stability limit is the ratio between Characteristic length of the element and the Wave speed. The wave speed is the square root of the ratio between Young’s modulus and Density. (Note : It is dilatational wave speed)
If you have 10mm/min, your process is almost surely quasi-static.
You will probably want to apply mass scaling to keep the computing time reasonable (or simulate a much faster version of the process).
If convergence problems are not too great, an implicit method might be better.
Martin Baeker Yes process is quasi-static and convergence is required. Total process time is 40 sec. I tried with the by default setting, it is taking much time, so I terminated the simulation. Will time scaling help to reduce the simulation time?
Yes, mass sclaing will help but you need to monitor kinetic energy, it should remain small compared to other energies in the model.
Or switch to implicit and don't bother with dynamics...
You can try mass scaling as suggested by Sir Mr. Martin. In status file ( status file will be available if you atleast one completed simulation ) , you can see the list of the top 10 controlling elements. Controlling elements means to say the list of the elements that are reducing your simulation speed. You can try to try to increase the mass of these elements, this you can do by changing the density, which will then affect the wave speed and ultimately, the Stable Time.
Or check for the mesh in these controlling elements. It so happens that you may have very good mesh, but somewhere one poorly shaped element may have one edge very small. This is the characteristic element size that ABAQUS will consider for calculation of the Stable Time. Hence, try to remesh this.
if i remember correctly ... it depends on the size of smallest element in the model and dilatation speed of material (material properties) - see attachment.
What you could try (for trial simulation runs - results may not be valid but you may at least see if your model is set properly)... is to set the critical time increment not bigger than 1e-006 then increase it step by step up to 1e-008 if things dont work out. Another thing (somebody allready suggested) is to use mass scaling .... or add so called virtual time scalling: set the time of the loading step to 0,01s instead of 1s ... or 0,4 instead of 40s...etc. and simoultaneously increase the velocity of the punching process (16,7mm/s instead of 0,167mm/s) ... etc.
Also good to know ... with mass/time scaling ... the kinetic energy (of the model used for proper results) should not exceed more than 10% of the total energy (? somebody correct me if i am wrong)...
kind regards, Matej
Matej Hudovernik Thank you for your reply..I will run some simulations as per your suggestions.
Sagar Pawar I hope it helps ... and works .... one correction: by critical time incr. not bigger than 1e-006 and then DECREASE it down to 1e-008 ... that would be proper phrasing :)
Critical time increment should be as:
Δ𝑡𝑐𝑟≤𝑙/𝑐𝑑
where 𝑙 is the dimension of the smallest discrete element in the model and 𝑐𝑑 is dilatation speed of material, defined as:
𝑐𝑑=√2𝐺(1−𝜈)/(1−2𝜈)≈√𝐸/𝜌
Critical time increment should be as:
Δ𝑡𝑐𝑟≤𝑙/𝑐𝑑
where 𝑙 is the dimension of the smallest discrete element in the model and 𝑐𝑑 is dilatation speed of material, defined as:
𝑐𝑑=√2𝐺(1−𝜈)/(1−2𝜈)≈√𝐸/𝜌
- Badges
- Science topic