How can you optimize vehicle components using topology?

One of the key things when working on FEA analyses for topology optimization is that after you define the design space/volume. You then need to define “non-design” and “design” space where you instruct the optistruct algorithm to only play around with the “design” space and not the “non-design” space. These non-design space typically are the installation/mounting locations of the part or manufacturing constraint points which are fixed locations and cannot be altered.

Topology optimization Applications 1. Performance improvement cases: Existing design Modification case. a. Topology optimization suggests reinforcement locations ( weak regions) of the subsystems. b. Topology optimization suggests additional connection required between parts or subsystems. New design case Topology optimization helps in capturing the optimal loadpath for constrained loadcases. 2. Mass reduction cases. Existing design Modification case Topology optimization helps in finding the locations where material has to be removed from existing parts (Cutouts). New design case Topology optimization helps in capturing the optimal loadpath for constrained loadcases with minimum mass.

I have worked on multiple projects where I’ve used topology optimization and I can say this is one of the best methods industry uses to reduce the weight/stress and increasing the stiffness without compromising on the functionality performance of the part. One of my achievements in this space: I had to reduce the weight of a knuckle/upright of a racing car as the driver was complaining about understeer due to heavy knuckle/upright. I had used topology optimization to reduce about 1.5kgs from it without effecting its stress and stiffness and the car won second place in the following race.

One of the algorithmic bottleneck that is rarely talked about in TO is the fact that when problems become multi-objective multi-constrained in nature which is quite often in today's engineering environment. These problems require finding global minima instead of local minima. Traditional algorithms (like gradient-based) can only find local minima. Hence, we need exploration of non-gradient algorithms which can find global minima. The challenge with such algorithms is that they are computationally very expensive. Recently, there are demonstration of novel accelerated non-gradient algorithms that can find global minima in less iterations and compute resources. Check BosonQ Psi's quantum-inspired evolutionary optimization algorithm.

1. Understanding of manufacturing process of sheet metal n Casted parts is required to define Topology Manufacturing constraints. 2. Interpretation of the Topology results is a biggest challenge. 3. Each element of the design space will be design variable, so HPC may be required to solve complex problem.

For designers who have not used FEA softwares like ANSYS, or meshing tools like hypermesh, you can use softwares like INSPIRE or optistruct. These tools are simple to use with less complexity in controls(which reflects in less accuracy compared to manual mesh + solvers). You can define your manufacturing technology(casting, Sheet metal, forging, 3D Printing) in these softwares to optimise the topology to suit your specific requirement. My favourite approach is to use unconstrained optimisation (Suited for 3D printing) and based on the load paths, design a part to suit other requirements.

To start with topology optimization, we follow following steps: 1. Define the problem statement - What do you want to achieve? Reduce weight? Reduce Cost? Improve life? and by how much? 2. Define constraints, environment and flexibilities - Set up your design space. Set up the hard points. 3. Define loads 4. Select realistic target - By defining % reduction in volume, you can check the max loads in the structure or vice versa. Use this to adjust your target. 5. Select material and manufacturing tech to meet your target. 6. Add manufacturability constraints to your model to re-calculate the topology for sheet metal/casting/3d printing. 6. Design a finished part around the optimized map and run detailed analysis of your final design.