Which Is Better for Analyzing Natural Frequency in Flywheel Systems: FEM or MBS?

[bhaazee]

I am working on a project of developing a Flywheel energy storage (powered by electric motor) for a vehicle.

The natural frequency of the system has to be calculated.

Simple project description:
The flywheel is held by two bearings inside a housing. The electric motor is fixed to the housing bottom and holds the clutch plate. The other end of the clutch is held by flywheel. When the clutch is engaged the electric motor drives the flywheel. All components as a whole is considered as a single assembly to be mounted on the vehicle.

please refer the diagram attached here.

I have less experience with FEM/Multi body simulation.
However, started the analysis with simpack. All 4 bearings can be assumed to be spring models. Also, the shaft, flywheel and motor's shaft tolerance have to be considered for the analysis (means both the shaft to be considered as elastic bodies when working in simpack). The torque transfer from motor to flywheel during the assumed driving cycle also has to be taken into consideration for determining the natural frequency. But currently I am uncertain of the correctness of my decision.

I would like to know whether FEM or MBS best suits for the analysis in determining the natural frequency of the complete system?

 

[Mech_Engineer]

What speed are you planning on operating the flywheel at? 1krpm? 10krpm? 100krpm? I'm suspicious that you've overlooked important issues concerning the speed this will have to operate at to store appreciable amounts of energy; how much energy do you hope to store?

 

[bhaazee]

min. speed - 2200 rpm
max. speed - 4000 rpm

it is designed to store 30 watthour of energy, power - 60 kw. A low density model.

 

[Mech_Engineer]

Sounds like you've got a handle on the flywheel, how much does your vehicle weigh?

When you say you want to find the natural frequency, I assume you mean a modal analysis? A modal analysis can be done using most (if not all) FEA packages. You'll need to be careful in modeling boundary conditions which approximate the bearings, since they only constrain specific degrees of freedom.

 

[bhaazee]

The weight of the vehicle is 3.2 tonnes. The weight of the flywheel set up alone mounts to 250 Kg approximately. Well the objective of this simulation is, the flywheel setup is to be mounted on the vehicle using damping elements. Hence we would like to know the eigen frequency of the flywheel set up and select the damping elements accordingly.

Yes modal analysis can be done with FEA. However, I am not sure if the driving cycles (moment of motor vs time, flywheel speed vs engagement time of clutch, etc) can be executed in FEA since I don't have much knowledge of FEA softwares. Because even the driving cycle factors counts when determining the eigen frequency of the entire system.

 

[AlephZero]

Your model needs to include gyroscopic effects in the rotordynamics to model this properly. Some FE systems will do that. Possibly simulink will also do it, and a quick Google search for "simulink rotordynamics" found a few presentations on wind turbines and one on whirl flutter, but it wasn't very clear what had actually been included in the models. From my own long experience, it's dangerous to assume anything has been done right unless it actually the write-up shows clearly that it has been done right!

 

[bhaazee]

Considering various flywheel layout models, the current model suits both in gyroscopic effects and costs. Although its possible with simulink (in control engineering means), use of simulink mutiplies the complexity as a separate FEA software is needed for stiffness calculations.
Also, not every aspect has been considered while designing the system, the important properties have been taken into account for designing. And btw regarding my question on Abaqus/SIMPACK, I think such rotor dynamics are better solvable with FEA rather than SIMPACK. Just going through the topic of explicit dynamic analysis.

Thnx everyone for their suggestion and time.

Regards.