Mass scaling to reduce the run time of a model and ensure the stability of the System using Ls-Dyna

To determine where and when mass is automatically added, write GLSTAT and MATSUM files. These files will allow you to plot added mass vs. time for the complete model and for individual parts, respectively. To produce fringe plots of added mass in parts comprised of shell elements ), set STSSZ=3 in  *DATABASE_EXTENT_BINARY . You can then fringe the added mass (per element) using LS-POST by choosing Fcomp > Misc > time step size. (Here, the label "time step size" is really the element added mass.)

 

Difference between Negative mass scaling & Positive mass scaling.

 

Negative: Mass is added to only those elements whose timestep would otherwise be less than TSSFAC* Dt2MS. When mass scaling is appropriate, I recommend this method.

Positive: Mass is added or taken away from elements so that the timestep of every element is the same. My opinion is there is no advantage to using this method over the negative DT2MS method and I find it harder to rationalize.

The parameter MS1ST in *Control_Timestep controls whether the mass is added only once during initialization (MS1ST=1)  or anytime as necessary to maintain the desired timestep specified via DT2MS (MS1ST=0).

You can use ENDMAS it in *Control_Termination to stop the calculation after a certain amount of mass has been added (active for automatic mass scaling only).

Procedure

Run the .k file as it is 

The Timestep parameters in that file are as follows

DTINT = 0.001

DT2MS = 3.5 X E-05 

TSSFAC = 0.9 (Default)

  

 The Estimated time shown by the solver is 37hr 29 min, which is very high for such a problem.

One more observation is the time step followed by the solver i.e. 4.40-4.68 XE-5 and so on, is more than the Dt2MS value mentioned in the TimeStep Control Card and as we are using Negative value Mass has added only those elements whose timestep would be less than TSSF * DT2MS.

If we calculate,

Mass Scaling = TSSF * DT2MS

                    = 0.9 * (-3.5*E-5)

                    =-3.2*E-5

And if we compare this value with the obtaining Timestep value (4.40*E-5) the Mass scaling value (-3.2*E-5) is less, and the Timestep is higher so no mass is added in the system.

To Reduced the Computational Time we have to do Mass Scaling, It will improve the speed and if we are able to mass Scale within the Limit Provided (Which is 8% is allowable) then it would not affect the Stability either.

Case 2

Increase the Dt2MS value to -5*E-05 And the Simulation till 10ms to see if any mass addition takes place

  

As we can see the Mass is added to the system and its value is 375 and the percentage increase is 0.001% of the total mass. This is much lesser than the allowable value but still, it has reduced the simulation time by 10hr, as you can see new simulation time by Mass Scaling is 29hrs 17min.

It means there is a Scope to Increase the Dt2MS value further till we reach near the allowable limit. Like this, I have performed No of Dt2MS calculation to reduce the simulation time by Mass Scaling.

 

As the DT2MS value decreases there is an increase in the Percentage of Mass.

As the DT2MS value Increases, there is a Decrease in the Computational time.

I had got the percentage of mass to 8% at the DTMS of 0.0001029 with a Simulation time of 15 hr 23 min which is much better than the result with no mass Scaling. Also the Similar result I have got with the Dt2MS of 0.001025 with the percentage of mass increase 7.76% with the same simulation time. 

I will surely choose the DT2MS = 0.0001025 for the safer side of the Simulation.

The result Marked with the Red color showing too much increase in mass that's why these values couldn't take for further simulation. 

In the explicit method, the time step usually is very small to maintain numerical stability. However, small step size prevents this method from being useful for routine analysis work. To reduce the CPU cost and improve the performance, mass scaling is often used to increase the time step size in each cycle.

Varying TSSFAC (Scale factor) for computed time step -Explicit Analysis.

For this, we are just varying the Scale factor by keeping the DT2MS = 0.0001029 as constant.

  

One noticeable difference is as we decreased the Scale factor in the context of DT2MS value the Simulation / Computational Time Increased.

  

with 0.9 scale Factor = 15hr 23 Min

       0.8 scale Factor = 18hr 24 Min

       0.7 scale Factor = 19hr 46 Min

The changing the scale factor only changing the Computational Time and the Mass Added is remain Constant, Which means the Scale factor doesn't have any effect on mass scaling. mass Scaling completely depends on the Value of DT2MS and Time Step.

Implicit Solver Method.

There is no Mass Scaling like Explicite is available in the case of the Implicite method. The mass Scaling Increase the Timestep by increasing the mass over the element that's why we can say that the mass scaling is Time step dependent. but in Implicite Module we work with the Steps and iterations, also the Implicite itself Independent of time so if there is no variable related to time present then we cant add mass also by Mass Scaling.

Here we have added

Using *CONTROL_IMPLICIT_AUTO card with IAUTO=1, automatically adjust time step size.

Using *CONTROL_IMPLICIT_GENERAL card, IMFLAG=1 (implicit Module is activativation) and with initial timestep as 1 (i.e.DT0=6).

 

CASE1:- Where keep the DT2MS = -3.25E-5 and Run for the 600ms

  

It just took some sec. to solve the problem with Normal Termination.

CASE2:-

Where keep the DT2MS = 1.029E-4 and Run for the 600ms. At this value, we have got the best Mass scaling result with the Explicite Module.

  

In this case, also solver is able to solve the problem without adding artificial Mass. The iteration taken by the solver is two numbers max to solve the problem in each step.

 

CONCLUSION
From the comparison, in different cases, it can be concluded that using a mass scale can significantly bring down the analysis run time without affecting the outcomes.

It is always a good practice to run the analysis without mass scaling to get an idea of the minimum time step computed by the solver and then scale it up using mass scaling if required.

As the DT2MS value decreases there is an increase in the Percentage of Mass.

As the DT2MS value Increases, there is a Decrease in the Computational time.

Scale factor doesn't have any effect on mass scaling. mass Scaling depends on the Value of DT2MS and Time Step and not on Scale Factor..

In this case, also solver is able to solve the problem without adding artificial Mass. The iteration taken by the solver is two numbers max to solve the problem in each step.

Implicte Analysis is good where the Time is more and deformation is less as we have seen in our problem therefore Implicite solver is good for this simulation.