'Differential Mount right 4. iteration' simulation project by steinccf


#1

I created a new simulation project called 'Differential Mount right 4. iteration':

Differential mount of a formula student car. The goal is to reduce the weight.


More of my public projects can be found here.


#2

Hi @steinccf!

That looks like a very interesting FE simulation! Can you give me more background information on that? Maybe you have a blog where you describe this project more precisely. Would love to know more in order to pick that for our Project spotlight :slight_smile:

All the best!

Jousef


#3

Hi @jousefm,

so this is the differential mount of the RUB 18 the next formula student car of RUB Motorsport. The FSAE team of Ruhr-University Bochum in Germany.

The Problem:
Since 2016 our team is using a 1-cylinder engine. The main drawback is the enhanced engine vibration and because the old differential mount was fixed to the frame it broke. We clearly underestimated the effect of the vibration.
In 2017 we used an entirely new design similar to what you see here. It holds engine and differential in a decoupled system which is damped by rubber mounts. The method for the FE analysis was to make a static model an design the part for a minimum safety of 4,5. Brutally inaccurate but the part didnt brake. The goal this year was to reduce the weight since the solution of last year was pretty heavy. A more accurate FEA was needed.

The input data for this simulation:
The aluminum alloy we are using (7075-T651) has a fatigue strength of 159MPa, an elasticity modulus of 71GPa and a poisson ration of 0,32.
The engine vibration was simplified to one direction and the acceleration data was sourced from a similar sized 1-cylinder engine. We planned to do acceleration measurements of the engine setup this year but for now this is what we have to work with. The movement of 4mm was calculated from the sinus-function of the acceleration data.
The boundary conditions were sourced from the datasheet of the dampening mount and the forces applied to the differential at full load.
There are a couple factors that are not considered in this to make the simulation easier. Gravity, cornering acceleration and acceleration from driving over bumps would be examples of this. They would probably require a multibody simulation of engine and drivetrain.

The simulation:
A “fine” mesh was chosen although it could have been 2D. The dynamic forces were modeled with a simple sinus function and the simulation interval and timesteps were chosen so that one point will hit the maximum of all sinus functions. This is to ensure that the worst case is included in the simulation. All other values were put in according to the data input.
There are 4 iterations of this simulation with geometry optimization after every FE analysis.


A topology optimization would have been nice but because of the dynamic nature of the problem it is more efficient to do it manually. I hope to do a much more precise simulation for the year.