Strain rate effect on composite material (/MAT/COMPSH)

Matteo Castagna_20400
Matteo Castagna_20400 Altair Community Member
edited March 6 in Community Q&A

Hello everyone,

I would like to enrich the material card of a composite material with the effect of strain rate, using the parameter c, but I have the problem that both conditions of low strain rate and high strain rate are affected by this introduction. In particular, the curves of both specimens get higher.

In this particular material, the only experimental test with a significant difference between quasi static and high strain rate conditions is the tensile 45° test. For this reason I decided to introduce this effect defining the global reference strain rate and the parameter c12 (instead of the global one c).

The global reference strain rate is defined according to the strain rate of the quasi static FEM model, and I have evaluated this value with the time history request EPSD of shell elements.

Is there any other aspect I should take into account or any errors I made?

In case I can share materials to solve this issue,

Thanks in advance.

 

Tagged:

Answers

  • Polyvios Romanidis
    Polyvios Romanidis New Altair Community Member
    edited March 6

    Hello Matteo,

    Based on your description the setup of the strain rate effect is correct.

    The only thing I propose you add is a strain rate filtering.  This will prevent the strain rates oscillations, that caused by numerical application.  To do that, you have to set Fsmooth equal to 1 and define an Fcut  value.  The smoothing is applied is an exponential moving average, given by the following equation.

    image

     

     

    Where a = 2π · dt · Fcut and dt is the time step of the simulation.

    The cutoff frequency is a function of the model timestep and its value should be determined.  For many applications, Fcut value belongs in the range between 1 and 10 kHz, but there are high-speed effects in some applications that require higher values of cutoff frequency.  For these applications, the Fcut can reach the value of 1 GHz.

    In all applications the Fcut value should be estimated by the simulation engineer, most of the time with a trial-and-error process.

    All in all, I recommend you enable smoothing filtering by setting Fsmooth = 1 and then try different Fcut values to determine the exact value that removes noise from your results.  I do not know your exact application, but probably values in the range between 1 and 10 kHz should work.

    Polyvios

Categories