Specifying cumulative particle size distribution

Raïsa Roeplal
Raïsa Roeplal Altair Community Member
edited April 26 in Community Q&A

Dear EDEM community,

I am modeling materials with different size distributions in EDEM. To my knowledge, there are three different methods for specifying a particle size distribution such that there is NOT an equal probability of each size being generated:

(1) Using a normal size distribution

(2) Using a log-normal size distribution

(3) Using a user defined size distribution

The shortcoming of methods (1) and (2) is obviously that these types of distributions should be able to fit to your particle data, which is not the case for my material. Method (3) is the next option, but I have a number-based size distribution as shown in the figure below. So, ideally I would like to specify the PSD using a cumulative distribution function. Is there any way to accomplish this?

image

Thanks in advance for any help/suggestions!

Cheers,

Raisa

Tagged:

Answers

  • Renan
    Renan
    Altair Employee
    edited March 11

    Hi Raisa,

    It looks like you have the total number of particles and the frequency of each size class, which means you can calculate the total number of particles in each size class. With that information and the material density you can calculate the total mass of particles in each class from which you can easily calculate the mass density for each size class.

    Best regards,
    Renan

  • Raïsa Roeplal
    Raïsa Roeplal Altair Community Member
    edited April 26

    Dear Renan,

    Thank you for your kind suggestion. The "frequency" is not really a frequency (sorry for the confusion); it represents the probability. So for example, for the data point indicated in my picture below, it just means that 1.1% of the total number of particles has a diameter smaller than 0.47 cm. This does not tell me how many particles have a diameter of 0.47 cm. 

    Of course, I can think of a solution to this problem: discretize the data by making bins (a histogram) based on size and determining the mass percentage of particles in each bin. However, I would rather not do this here and was just wondering if there is any option for specifying a cumulative size distribution.

    Cheers,

    Raisa

  • jerrinjobs
    jerrinjobs
    Altair Employee
    edited April 26
    Raïsa Roeplal said:

    Dear Renan,

    Thank you for your kind suggestion. The "frequency" is not really a frequency (sorry for the confusion); it represents the probability. So for example, for the data point indicated in my picture below, it just means that 1.1% of the total number of particles has a diameter smaller than 0.47 cm. This does not tell me how many particles have a diameter of 0.47 cm. 

    Of course, I can think of a solution to this problem: discretize the data by making bins (a histogram) based on size and determining the mass percentage of particles in each bin. However, I would rather not do this here and was just wondering if there is any option for specifying a cumulative size distribution.

    Cheers,

    Raisa

    Hi Raisa

    The input needed for EDEM size distribution is the size and mass percentage. You will have to modify your data in a way for EDEM to read it correctly.

    The cumulative size distribution needs to be modified, and there is no way to import it directly.

    Thanks,
    Jerrin Job

Categories