Hello everyone!
I am doing a frequency analysis of a train car in ABAQUS. I have to consider mass from the passengers. I have two options:
1) Use Coupling structural to distribute the point mass of all passengers to a floor
2) Increase density of the floor to account the passenger's mass.
I tried both and got different results (got less eigenfrequency, using "density" approach).
Which one is more correct? What can you suggest me? Does the Coupling structural add some rigidity (I think it should as RBE3)?
Hi Aleksandr,
Why are you adding the passenger mass, and what aspects of the dynamic behaviour of the carriage interests you?
Neither method would be suitable if you are interested in the behaviour of the floor because you are changing its mass distribution.
People don't really behave like masses that are fixed by springs.
Regards,
Simon
Simon Smith I need to find minimal natural frequency of the car with passengers (the frequency value should be below a certain number specified in regulations).
Hi Aleksandr,
Is the minimum natural frequency of the car affected by the way in which you add the passenger mass?
Simon
Aleksandr Elkin Can you share your Abaqus model (.inp format)?
You have different options to consider the passenger mass. You can either distribute the mass over a surface or couple discrete mass elements to the structure via MPCs.
Simon Smith Yes, the way the load applied has a huge effect on natural frequency, e.g. MPCs type or the area of distribution can cause stiffening of the whole model.
Nils Wagner , unfortunately, I cannot (NDA and it is rather big). Yes, tried 3 ways:
1) Distribute through a one RBE-3 (MPC coupling structural)
2) Distribute as an extra density in the floor
3) Use 5 RBE-3 (MPC coupling structural) elements along the train car.
Results of 2,3 are pretty similar (2 is provide a bit higher frequency than 3).
Results of 1 are the highest frequency. It seems it adds "kinda" stiffening or just incorrect load distribution, but RBE-3 should not add any stiffness.
So, I am a bit confused what to choose and want to refer to maybe some regulations related to train (subway) cars.
Hi Aleksandr,
What is the sort of deformation of the first mode, and what is its frequency?
Simon
Aleksandr Elkin What about an experimental or operational modal analysis ?
Afterwards you can correlate the mode shapes using the MAC matrix.
Imho, your first approach "Distribute through a one RBE-3" is not suitable.
Nils Wagner I didn't do that, there is no data for EMA or OMA due to it is design stage, so just linear frequency analysis was done only.
Yes, I agree that just a one rbe3 over the whole floor is questionable (at least this approach should be substantiated).
I also though about another way: Create a block of elements that represents the passengers (a crowd) with properties of a solid body with average human density and young's modulus of 1-10 MPa.
Simon Smith the first mode is when the car have kind of lateral shear swing around 6 Hz, but the only vertical bending mode of the interest, so it is ahould be at least 8 Hz.
Hi Aleksandr, I'd be surprised if these modes are affected by the way the mass of the passengers is added. Are you also predicting the natural frequencies of the car floor, and if so, what are these? (i.e. the vibration of the floor itself rather than the whole car).
Simon
- Badges
- Science topic
- Similar topics
- Chemistry
- Electrochemistry