How to output the correct nodal acceleration in Radioss

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

Hi all,

I want to compare the acceleration result of my simulation with that of experiment. So I extract the nodal acceleration from animation result by using Build Plots tool in HyperView (show in the below figure). 

2017-10-19_224116.png.c96869d7d81f1609d90ca013e8f3400f.png

 

However, the acceleration is very big comparing to the experiment result. I suspect if this simulation acceleration result could be directly compared with testing acceleration or not. Should I have to do some filtering with the acceleration simulation result?

2017-10-19_224829.png.ef3460c712f8ebc1eface1be3b1c50c4.png

 

My model is an ABS block drop test. Please see the attachment model file. The model unit system is kg/mm/ms/GPa. The drop height is 50mm and corresponding initial velocity is 0.99m/s. The total mass of the ABS block is about 850 gram.

 

Could you help me how could I get the top center nodal acceleration which can compare with the experiment? Thank you.

 

Roy 

 

 

Unable to find an attachment - read this blog

Tagged:

Answers

  • Altair Forum User
    Altair Forum User
    Altair Employee
    edited October 2017

    Hi Roy,

    I suspect with the material values and unit consistency followed. Please review the same in the model.

    For nodal accelerations you can define an accelerometer in the model using sensors and create an output block for the same. From the time history file you can plot the curve. 

  • Altair Forum User
    Altair Forum User
    Altair Employee
    edited October 2017

    Hi Roy,

    I suspect with the material values and unit consistency followed. Please review the same in the model.

    For nodal accelerations you can define an accelerometer in the model using sensors and create an output block for the same. From the time history file you can plot the curve. 

    Hi George,

    My unit system is kg-mm-ms-GPa. The block of my model is made up of ABS material. Density=1080kg/m3, E=2GPa, u=0.35, yield stress=45MPa, tangent modules=0.02GPa. I think there is no problem in the unit system. 

    For accelerations, is there any difference between the method you mentioned and the method I used in the first post? What's more, how to define the node accelerometer in HyperMesh?

     

    Thank you

    Roy

  • Altair Forum User
    Altair Forum User
    Altair Employee
    edited November 2020

    Hi Roy,

    Please check the files attached. Please go through this videoseries, https://altairuniversity.com/conceptual-design-of-a-3-wheeler-motorbike/conceptual-design-of-a-3-wheeler-motorbike-crash-box-analysis/ where defining an accelerometer is shown.

    In the acceleration curve you can see that typical spikes are happening at a particular time step. So, normally what we do is we will filter this result in HyperGraph. In experimental testing also they will filter this curve and we normally use the same filter they have used.

    Unable to find an attachment - read this blog

  • Altair Forum User
    Altair Forum User
    Altair Employee
    edited October 2017

    Hi Roy,

    Please check the files attached. Please go through this videoseries, https://altairuniversity.com/conceptual-design-of-a-3-wheeler-motorbike/conceptual-design-of-a-3-wheeler-motorbike-crash-box-analysis/ where defining an accelerometer is shown.

    In the acceleration curve you can see that typical spikes are happening at a particular time step. So, normally what we do is we will filter this result in HyperGraph. In experimental testing also they will filter this curve and we normally use the same filter they have used.

    Hi George,

    I have compared the two acceleration results by using Accelerometer sensor (/accel and /th/accel) and node history (/th/node) for the same node. And the result shown that the two results were the same. I have uploaded my model file.

    Is there any parameters that influence rigidwall force mostly? How to do filter in HyperGraph?

    Roy

    Unable to find an attachment - read this blog

  • Altair Forum User
    Altair Forum User
    Altair Employee
    edited October 2017

    Hi Roy,

    For accelerometer sensor we select a rigid node, that is we are ensuring that it is not deformable and will not get disturbed due to impact. That's why a rigid body was created in the block and the sensor was placed on the master node.

    In HyperGraph, right click the curve, select filter option and select the required filter.

    <?xml version="1.0" encoding="UTF-8"?>acc.PNG

  • Altair Forum User
    Altair Forum User
    Altair Employee
    edited October 2017

    Hi George,

    From the help of Accelerometer /accel, the filter is Butterworth. That is to say the filter is a low pass filter. However, in many data acquisition device, there is high pass filter. How can we do this?

     

    Roy

  • Altair Forum User
    Altair Forum User
    Altair Employee
    edited October 2017

    Hi Roy,

    We have many options available for curve filtering.

    I recommend you to go through HG-1030: Referencing and Filtering Curves tutorial in Help which describes on how to reference and filter curves and it shows several options.

  • Altair Forum User
    Altair Forum User
    Altair Employee
    edited October 2017

    Hi Roy,

    We have many options available for curve filtering.

    I recommend you to go through HG-1030: Referencing and Filtering Curves tutorial in Help which describes on how to reference and filter curves and it shows several options.

    Hi George,

    It's a very useful example. May be we can do filter by post-processing. By using /ACCEL directly, the software can only do low pass filter I think. 

     

    Roy