Generative Optimization of a Wheel
Generative optimization is pretty fascinating, I've already written one article on it here. The idea is to define loads and constraints, and let the computer design the part to optimize the weight and stiffness, by only leaving material in the most critical areas.
Below are a few parts that I ran through the shape optimization process. There is a cantilever beam, bridge, chair, bike frame, and c-clamp.
This article is a look into how a wheel would be designed using generative design. It's a high level overview, but the concept can be applied to many different wheel geometries. The software used is Autodesk Fusion 360
All of this design was done in Fusion 360, a powerful and economic modeling solution offered by Autodesk. Check it out, it has free versions for students, hobby use and start-ups.
Step 1: Draw the Defining Solid
The defining solid should be as large as possible, fill the part dimensions out to their maximums based on space constraints. We're going to let the computer optimize the ideal shape, we just need to let it know the boundaries it can work in:
Step 2: Forced Symmetry
A wheel is radially symmetric, however Fusion 360 solver doesn't let us define that. We'll force it into giving a radially symmetric design by only drawing 1/3 of the wheel. The 1/3 is a bit arbitrary, it could be 1/4, 1/5, etc. Play around with it a bit too see how different portions affect the output
Step 3: Define The Areas to Keep
The first step in setting up the simulation is setting the central hole as our fixed constraint. This area needs to be defined as a geometry the model needs to maintain. We'll generate a keep zone around this central hole, as a bearing or shaft will most likely exist here.
Step 4: Define Loads
The outer perimeter of the wheel should be defined as a pressure. The wheel needs to be equally strong at any point as it rotates. To accomplish this, we'll apply a constant pressure to the whole outer perimeter. The magnitude is not important, since we're only applying once force. The magnitudes are only important when multiple forces are applied, since the ratio of the forces will dictate how the solver creates the optimized shape.
Step 5: Simulation Symmetry
We want to maintain some amount of symmetry for the remaining solid. This keeps the wheel balanced, and will also reduce simulation time.
Step 6: Setup Mesh Details
Next jump into the settings panel and reduce the size of the mesh to help provide a more accurate result. Do one final check to make sure the simulation setup is complete, and then run the simulation.
Step 7: Result
When the result is ready, turn off some of the simulation setup visualizations. Analyze the result to ensure the solution makes sense. Did the solver generate a symmetric shape that would be balanced? Check out the slider to see the different weight savings, and how the shape changes
Step 8: Model It
Promote the simulation result back to the model space. This will create a new body that can be used as a reference to re-draw the optimized part. When drawing the shape, it's important to think about how the part will be made. If you're machining it, check out my video on machining for some tips and tricks. The key is to keep things simple, an approximation, and always consider the manufacturing method that will be used.
Step 9: Validate Result
Apply a material, and run a FEA simulation to make sure the final shape can support the loads needed. To do this, model a small flat on the outer wheel perimeter. Use this as your fixed support, and apply a vertical bearing load to the center hole. Make sure the resulting Von Mises stress is less than the material yield strength.
Interested in learning more about mechanical design and materials? Here are two fantastic books that I own both of, and have been bibles for a lot of my design projects:
Shigley's Mechanical Design: https://amzn.to/31L3eg2
Ashby's Material Selection: https://amzn.to/2F9pkPG
(These are affiliate links and don't cost you a thing, but help me keep making content)