Balance between volume fraction and element density threshold

Altair Forum User
Altair Forum User
Altair Employee
edited October 2020 in Community Q&A

Hello everyone, I'm trying to find an optimal beam layout to design an engine mount through topology optimization. The original mass of the design space is 360 kg. I want to reduce it to an ideal mass of below 6 kg. However, topology optimization won't allow the volume fraction to be less than three times the element density threshold (MINDENS) which by default is 0.01. That means the lowest volume fraction I can achieve is 0.03. But 0.03 of the original mass is just 10 kg. To achieve the mass of 6 kg, I will need to reduce the MINDENS to at most 0.0056 that will achieve volume fraction of 0.0167.

 

However, this will produce a visually unfeasible result even when the solution converged. I think the default value 0.01 is the lowest I can go. I can't reduce the design space anymore as this is region that I want the solution to cover. Is there any other possible ways that I can use to reduce the volume fraction without altering the MINDENS value?

The first picture is my design space model. Four non-design space, two at front an two at the back.

problem1.PNG.92c1e3cb74ed830fa10013698ecf8416.PNG

The second picture is when MINDENS=0.009 and volume fraction of 0.027.

problem2.PNG.56dbf14748efd2c505bc10734121755d.PNG

The third picture is when the MINDENS=0.01 and volume fraction of 0.05. Here, plates exist in the solution and the mass is 18 kg.

problem3.PNG.29314eeed66e8096ebf6fae7f092955a.PNG

 

Any advice is much appreciated. Thank you.

Answers

  • Altair Forum User
    Altair Forum User
    Altair Employee
    edited August 2018

    Hi @Izzdham Shah

     

    Can you share a reference of MINDENS and Volume fraction relation?

     

    Why not use Mass with absolute target value instead of a fraction?

  • Altair Forum User
    Altair Forum User
    Altair Employee
    edited September 2018

    Hi Prakash,

     

    I'm sorry for the late reply. The reference for the relation is in the hyperworks solver view when I tried to run Optistruct. It is an error message.

    Messages for the job:

     
     *** ERROR # 2150 ***
     Invalid Volume or Mass Constraint.
     The response type is a property VOLFRAC. 
     the upper bound value = 2.0000e-002
          the total volume = 4.5993e+007
         the design volume = 4.5993e+007
            the total mass = 3.6105e+002
           the design mass = 3.6105e+002
     The design material fraction (percent) should be at least 3 times
     the DOPTPRM MINDENS (default 0.01).

     

    I think using mass constraint is a good idea. I will have a go with that first. Thanks.

     

  • Altair Forum User
    Altair Forum User
    Altair Employee
    edited September 2018

    Hi Prakash,

     

    I'm sorry for the late reply. The reference for the relation is in the hyperworks solver view when I tried to run Optistruct. It is an error message.

    Messages for the job:

     
     *** ERROR # 2150 ***
     Invalid Volume or Mass Constraint.
     The response type is a property VOLFRAC. 
     the upper bound value = 2.0000e-002
          the total volume = 4.5993e+007
         the design volume = 4.5993e+007
            the total mass = 3.6105e+002
           the design mass = 3.6105e+002
     The design material fraction (percent) should be at least 3 times
     the DOPTPRM MINDENS (default 0.01).

     

    I think using mass constraint is a good idea. I will have a go with that first. Thanks.

     

    Yes, as suggested, please use mass instead of mass fraction and check the results. 

  • Altair Forum User
    Altair Forum User
    Altair Employee
    edited October 2018

    Apparently, using mass constraint is the same case as applying volume fraction or mass fraction constraint. The error still popped up. I think this is also true for volume constraint as they are all related by density formula. I think the minimum volume fraction is acceptable for my work. Thank you for replying.

  • Altair Forum User
    Altair Forum User
    Altair Employee
    edited October 2018

    HI,

     

    Can you check the mass in your model?

     

    If mass is not available, please provide density on Material card and try again!!!