How to judge mass scaling results?

[M MOTAAL]

Hi everybody. I have a quasi static problem that I need to use mass scaling to increase the stable time increment and it's the first time i use mass scaling. I used two mass scaling factors as I followed a strategy in assigning mass scaling factor to selective element sets with larger value of scaling factor (200) for the those elements with smaller stable time increment than 1e-5 and another value (150) of the same factor for the rest of the model elements.
Taking mass scaling factors in consideration made the expenses of computations much lesser. But I have some worries about the results. How can I judge the results is the issue. I have been surfing the net but I didn't find much about this topic. Some of the few results while searching, was talking about some criteria through which I can accept the results of the model after applying mass scaling. They mentioned that total energy must be constant and for quasi static in particular, and has to equal zero or a value close to zero. But I can find the words "close to zero" are very ambiguous. In my model the total energy was constant but equal to (0.225) my units are in N and mm. I also found another criterion related to kinetic and internal energy. It states that the kinetic energy has not to increase more than a ratio from Internal Energy equals to 5%.
My questions are:
a) Is the value of 0.225 for total energy for quasi static in my model accepted?
b) Is there anyway to judge mass scaling factor or making an initial gesture of it's value?
c) Do anybody has a reference in this topic that I can depend on in this topic?
d) What kind of effect may result from assigning more than mass scaling factor in the same model on selective set basis? has anybody tried it before?
Thank you in advance.

 

[Dave442]

When mass scaling in quasi static analyses I target a single value for the stable time increment and then adjust this value up/down by an order of magnitude until the ratio of kinetic/internal energy is < 5%.

For reference, this stuff is described in a lot of detail with examples in the documentation. For Abaqus 6.13 you will find the following topics:

In the Abaqus Users guide there is an entire section on Mass-scaling:

11.6.1 Mass Scaling

In the Getting started with Abaqus guide there's also a section on quasi-static analyses where mass-scaling is discussed, Methods for interrogating the energy history are described, An example is provided of a forming analysis where the effect of mass scaling is shown. In this example they show how to interrogate the energy history and adjust the scaling definition based on the analysis output. they go through this in a lot of detail.

13 Quasi Static Analysis with Abaqus Explicit

 

[IceBreakerSours]

In a physical quasi-static test (on a bench), at a macroscopic scale, material deforms at slow rate i.e., the kinetic energy is small in comparison to the deformation energy. At a microscopic scale, a whole is happening that we ignore, which can easily come back to bite us when we are trying to predict localized phenomena. Now, on to the numerical test, if the kinetic energy in the model is small relative to the physical energy, then you have a realistic result (assuming the model is an appropriate approximation of the physical reality). Furthermore, if the added mass is away from the area of interest, then that works even better.

*********************************************************
Are you new to this forum? If so, please read these FAQs:

http://www.eng-tips.com/faqs.cfm?fid=376

http://www.eng-tips.com/faqs.cfm?fid=1083

 

[M MOTAAL]

Thank you Dave442 and IceBreakerSours for response. I want to ask also about the 5% ratio between ALLKE and ALLIE what is meant by this ratio, is it to compare individual xy-data from abaqus in a spread sheet for example?

 

[IceBreakerSours]

I suggest you seek some one-on-one expert supervision.

*********************************************************
Are you new to this forum? If so, please read these FAQs:

http://www.eng-tips.com/faqs.cfm?fid=376

http://www.eng-tips.com/faqs.cfm?fid=1083

 

[Dave442]

That's the idea, create an XY plot of IE vs. KE. KE should remain below 5% of IE.

There's an example in the documentation I referred to that shows the effect of the scaling factor on this plot:

Getting Started with Abaqus: Interactive Edition

13 Quasi Static Analysis with Abaqus/Explicit​

Example: Forming a channel in Abaqus/Explicit​

Figure 13-21: Kinetic and internal energy histories for mass scaling factors of 5, 10, and 25.​