Chassis Design VDS Fuel Cell Car

This instructable deals with the chassis design and manufacturing of the Fuel Cell Electric Vehicle from the Vehicle Design Summit at MIT.

The Fuel Cell car is a small two person, three wheeled (2 in the front, 1 in the back) car for commuter purposes.

Step 1: Anthropometry and Component Parts

An important part of the chassis design is to check if a human being fits inside. Anthropometry means the measurement of humans; it deals with the ergonomics of a human being inside a car.

In literature there is information available about anthropometry in the automotive industry:
The standard J1100 of the Society of Automotive Engineers (SAE) has standardized different dimensions for the space available for driver and passengers in cars: it is difficult to find values for these dimensions in J1100
Automotive Ergonomics (Peacock & Karwowski, 1993): deals with several topics of ergonomics and anthropometry in vehicle design; also has values of important dimensions!

At www.3dcontentcentral.com (a free resource for 3D CAD drawings) there is a car seat model available for SolidWorks. With this model and the data available from literature a new model was made in SolidWorks, the so-called forbidden space model. This is a model with a layout of the seat and the space that is needed to accommodate the driver and passenger: this space can not be used for the structure.

Also in this step you need to have a good idea about all other big componentes in the car and how to accomodate them.
(pictures copied from "Automotive Ergonomics (Peacock & Karwowski, 1993")

Step 2: Concept Design

The next step is to generate a concept. To save time and weight we chose to design a spaceframe using aluminum square tubes. The concept was made using the 3DSketch option in SolidWorks and mechanically validated (see next step) using ANSYS software.

Generating a concept is a real trial-and-error proces. Keep in mind that the chassis design depends on every other component in the car and vice versa. The best thing to start with is to make a list of all major components and requirements that are important for the chassis design. For our case:

1) Suspension mounting point (the biggests mechanical loads on the chassis will be there)
2) Weight estimate of the complete car (you need this for FEM calculation)
3) Load scenarios (see next step)
4) Body shell (affects the outer shape of the chassis)
5) Fuel Cell (big and heavy)
6) Battery Pack (also big and heavy)
7) Driver and passenger position and space (see previous step)

Step 3: FEM Validation 1/3

F.E.M. (finite element method) can be used relatively easy and fast to check your concept for mechanical requirements. There are a lot of software packages available for this purpose, I used ANSYS. Some software packages are really easy to use, but I think you always need somebody with experience in mechanical design to check your model and calculations. For novice users it can be hard to understand the limitation of FEM software and mistakes are made easily (even by professional users)!!

The goal of the FEM calculations is to detremine a good layout for the tubes in order to bring the weight down as much as possible.

STEPS FOR ANSYS

MODEL
I started by exporting the 3Dsketch from SolidWorks to ANSYS (convert SW model to IGES file).

ELEMENTS
We are going to use square or rectangular tubes because they are easier to handle than round tubes. This implies the use of BEAM4 elements, because the easier to use PIPE16 only works for round tubes.
The BEAM4 element is an uniaxial element with tension, compression, torsion, and bending capabilities. input parameters (real constants) for this element are:
-Cross-sectional area
-IZZ Area moment of inertia
-IYY Area moment of inertia
-TKZ Thickness along Z axis
-TKY Thickness along Y axis
-IXX Torsional moment of inertia
-ADDMAS Added mass/unit length

MATERIAL
We are using aluminum 6061-T6 tubes. At www.matweb.com the following material properties can be found:
Density: 2700 kg/m3
Elastic modulus (Young's Modulus): 68.9 GPa
Poisson's ratio (PRXY): 0.33
Tensile Yield Strength: 276 MPa

Step 4: FEM Validation 2/3

LOAD CASES
In this section the load cases will be defined as well as how to implement these load cases in ANSYS.

4G BUMP
When the car hits a bump in the road the chassis has to endure a load of 4 times the mass in vertical direction at the suspension mounting point. The target mass of the car is 500 kg, together with the mass of driver and passenger this is 650 kg. This means the force at the mounting points equals 4*g*650/3 = 8502 N.

Failure criteria:
The maximum stress (Von Mises) should be less then the yield stress divided by a safety factor.
The maximum displacement (displacement vector sum) should be less than 1/400 of the wheelbase. This is a rule of thumb for the bending stiffness of the chassis.

1G STEER
When the car has to make a sudden turn (e.g. to avoid a collision) the chassis has to endure a load of 1 times the mass in horizontal direction at the front suspension mounting points. This means equals a force of 1*g*650/2 = 3188 N.

Failure criteria:
The maximum stress (Von Mises) should be less then the yield stress divided by a safety factor.

1G BRAKE
When the car has to make an emergency stop (e.g. to avoid a collision) the chassis has to endure a load of 1 times the mass in horizontal direction at the front suspension mounting points. This means equals a force of 1*g*650/2 = 3188 N.

Failure criteria:
The maximum stress (Von Mises) should be less then the yield stress divided by a safety factor.

TORSION
The chassis should have a torsional stiffness of 2000 Nm/deg: when a moment of 2000 Nm is applied on the center axis of the car (the axis in the center plane of the car between rear suspension and front suspension) the chassis is allowed to have a rotation of 1 degree.

The way to do this in ANSYS is to apply a force in vertical (opposite direction) direction at the front suspension points and constrain the displacements of the rear suspension point. Run the simulation and examine the displacement vector sum (USUM): convert this translation to rotation.

Again the 2000 Nm/deg is a rule of thumb.
The importance of chassis torsional rigidity is very simply the ability to maintain a stable platform for the suspension to operate from and relate to the road surface (Source: http://www.jblmotor.com/JBLchassis.html, July 7 2006)

STATIC LOADING
Static loading is to check if the chassis is capable of bearing all the loads of the components in the car. We have the following list of components (based of estimation):

Component Weight (kg)
Propulsion 80
Fuel Cell system 200
Power storage 20
Composite body 50
Interior (seats, dashboard) 100
Driver and passenger 150

The problem is that we do not know yet where all components will be placed in the car. Normally load requirements for torsional- and bending stiffness are dominant over static loading so at this point we do not have to worry about static loading to much.
When there is more information available about the placement of the components a static analysis can be made to check for local stress concentration.

5G FRONTAL COLLISION
A distributed load of 5*g*650 = 31882 N will be applied to the frontal tubes of the car. The maximum stress should be less than the yield stress divided by a safety factor.

5G SIDE SOLLICION
Same as front collision, except that forces act on the side tubes of the chassis.

5G REAR COLLISION
Same as front collision, except that forces act on the rear tubes of the chassis.

3G ROLLOVER
A distributed load of 3*g*650 = 19129 N will be applied to the tubes on the roof of the chassis. The maximum stress should be less than the yield stress divided by a safety factor.

Step 5: FEM Validation 3/3

Some final remarks on the FEM validation:

- Make sure you know what you are doing using FEM otherwise you could end up with unreliable results.

- Load scenario's are dependent on the situation where you are designing for. In our case the load scenarios were based on "rules of thumb" for solarcar design and we used a sfaetyfactor of 2.

Step 6: Finalizing Design

Having a 3d sketch and a FEM model of your concept is just the start. Making the drawings for production is probably the hardest work, or at least the most time consuming work.

Although a FEM model or 3D sketch looks "finished" it still needs a lot of work before you can start building (without having to deal with unpleasant suprises). A good way to experience this is to make a scalemodel of your spaceframe using small wooden bars.

In the file "assembly building steps.doc" you can see how we designed and constructed the chassis.

Comments

[alxram]

Where can I find out more about the fuel cells that will be used to power this car?

[Pal]

The Tesla Roadster is being developed with banks of cell phone batteries as an option over traditional batteries, it seems like an ideal option as you leave the research of battery efficiency to larger companies already working on it for their own benefit and use a lighter, modular system for power.

http://www.teslamotors.com/index.php

[VanBee]

Thanks for your the usefull feedback. The fuel cell people in our group should post a report here, I don't know if they already made one but hopefully they will do it soon. Due to time constraints we were not really able to optimize the design of the frame. All tubes are welded by a very skilled welder. He knew where to drill small holes to avoid local stresses, and there was actually no need for a jig -> I don't know why (I also expected more frame distortion), maybe because of the modular design and the experienced welder... The picture above is not the final design: we've added some extra members for torsional and bending stifness. The central tunnel was designed for extra torsional stifness (the idea was to make a box out of it for stifness and to store samller h2 tanks and cables in). The structural requirements came from solarcar design and are more or less standard. In solarcar design I used the skin of the car to contribute to the stiffness, but in this case there was no time to design the body so we only had the frame to work with...

[foote]

Beam 188 or 189 elements will provide you with a much better visualization of your model, as well as more post-processing options. 188 and 189 use the actual cross-section to calculate stresses, etc. while beam 4 elements can only calculate bending stresses in two orthogonal directions. Also, when you weld a heat treated aluminum, you loose much of the strength about the weld locations and you need to account for this in your chosen design strength. You could heat treat the entire frame back to a T6 state, but it would be VERY difficult to prevent warping during the quench. And I am talking about a lot of warping. I'm really looking forward to some articles on the actual drive train of this thing. Hope they are coming...

[gsport george]

How are you planning on joining the aluminium sections together? Welding is likely to lead to a lot of frame distortion if you dont have a strong jig to restrain it during welding. I also think you will have problems achieving the torsional stiffness you specify unless you add some diagonal members to the floor, sides and roof. Since weight is a major factor you might want to try to let the skin of the vehicle contribute to the torsional stiffness. It would also be interesting to know the specific requirements that led to this design. The central tunnel seems like a slightly frivolous use of material unless there is some obvious purpose for it that you havent mentioned. Good Luck.

[saites2001]

Anything on DIY fuel cells?