Lightning Motorcycles Revisits Generative Design for Mass Production
By Peter Simpson, Richard Hatfield, and Nick Markovic for Autodesk University
Lightning Motorcycles has been breaking records in the world of e-bikes since the company entered the market. The LS-218 is pushing electric and gasoline-powered motorcycles to be the fastest production bike on the market. But pure performance is not the only goal; Lightning is focused on providing consumers with the best quality and value in every product it creates. Key to meeting these goals and staying ahead of the competition are new technology and ways of thinking.
Lightning Motorcycles was an early adopter of generative design. Collaboratively, we produced a part-consolidation and lightweighting workflow that saved 30% of target component mass. In this article, we’ll show you how developments within generative design can be used to gain these benefits, while ensuring cost-effective manufacturing. We’ll cover workflows used within Fusion 360, maximizing the integrated simulation, generative design, and manufacturing tools to better understand the part and its performance.
Fusion 360 Generative Design
Fusion 360 Generative Design is a design exploration tool, aiming to cut down the necessary steps within the standard design exploration cycle. It is problem-focused, with real-world loading and geometry inputs driving the design iterations, and allows multiple viable designs to be created simultaneously. Figure 1 shows the standard design-to-production workflow, in comparison with the generative design workflow. The grey workflow shows a few ideas being created, some key ones being chosen to explore further, and then a complex loop of validation via design for manufacturing, simulation, etc. The generative design workflow can be seen in color. Large quantities of designs are created and iterated upon within the software itself. These all have a problem defined via loads and geometry, as well as manufacturing bias within the setup of the problem. This can allow for vast productivity increases, as well as creating new designs that the human mind would not be able to produce in such a time, making it an incredibly powerful tool.
- Understand generative design advancements allowing the revisiting of the project to be successful
- Understand the generative design setup
- Evaluate relevant trade-offs associated with generative design
Generative Design Advancements
One perceived shortcoming of generative design in years gone by is that it creates excessively complex geometry. Our initial work with Lightning Motorcycles shows a good example of this. The software at the time was very new, yet still incredibly powerful. It allowed all these complex geometries to be created yet there was a lot of refinement needed for the workflow to succeed from the very early outcomes.
Figure 2 shows this initial work with Lightning Motorcycles. You can see that the part has a high level of detail, and the original assembly was solved as one part. This was down to the limitations of the software at the time. There were far fewer solver options, which meant that it was far harder to set up the real-world loading of the part within the generative design software. As a result, the loading would be approximated. Nowadays, we have more complex solvers and a greater variety of load case attributes that we can fully use to better represent the real-world loading. Subsequently we were able to split our design into the same three parts as the original design, two end pieces, and a central component acting almost as a torsional brace.
The other shortcoming of generative design at the time was the fact that it was strongly focused on additive manufacturing. This was an area that was always going to be developed upon as it is key that the software allows for a wide variety of design problems to be applicable. Within generative design you can now include 2.5, 3, or 5-axis milling, 2-axis cutting and die casting, along with the traditional additive and unrestricted manufacturing methods. This allows the complexity of the part to be controlled. The inbuilt cost estimation, powered by aPriori, gives even more knowledge to the user about the cost effectiveness of their designs, leading to a new wave of generative design outcomes to be possible.
Generative Design Setup
For any problem-based solution to succeed, you need to be able to accurately represent the problem in order to make sure all outcomes are matching the real-world conditions that the part is going to be under. As you will see in the simulation section, we had a very well-defined problem due to the simulation criteria Lightning Motorcycles provided.
For both of the side components, the setup was very similar. Figure 4 shows the geometry used in the setup on the rider’s left-hand side part. (‘Rider’s left’ dictated as if rider was seated on bike facing forward.) The geometry can be seen, with the red obstacle regions set to 40% opacity for ease of viewing. The obstacle regions for generative design represent areas that the solver will not add material in. These may be due to other parts of a wider assembly, or potential ranges of motion. It is worth noting that the left-hand side features the obstacle for the sprocket. This was key to represent well as Lightning vary the sprocket size on their bikes. As a result of this, we oversized our obstacle from our original model to fit all sprocket sizes. The other obstacles are dictated from parts of the bike assembly, such as the wheel and motor.
The preserves are shown in Figure 3 in the characteristic green, as found within generative design. These are taken from the original design to maintain a similar assembly line to reduce the need to change the product line that Lightning already have in place. These are chosen by evaluating the critical fixture points, as well as the loading regions that allow the part to complete its purpose. It is possible that we could look at these in more depth, rearranging fixturing, etc., but this may cause a large increase in the assembly time and may not provide much benefit.
Figure 4 shows the actual setup of the generative design. To represent the part as a subset of the assembly, we fixed certain geometries that were not loaded to represent the fixturing to the other assembly parts. We then extracted the relevant loads from our simulation model and applied these to the loaded bodies. This allows each part of the assembly to be relevant to the overall assembly model whilst still maintaining the 3-part assembly overall. This was the same for both side parts of the assembly, with the only difference being slight changes in the obstacle geometry. As a result, we will not show the setup of the rider’s right-hand side.
The geometry and setup for the central component were slightly more complex that the side parts. The geometry can be seen in Figure 5. This geometry was very complex due to the location of the component, and specifically its proximity to other components. As a result, there was a very well-defined build volume to avoid the tire, motor, shock absorber, etc. that is represented by all of the obstacle geometry, again shown in 40% opacity red coloration.
The preserve regions were also slightly complex on this part. The part consisted of a series of through holes or capped holes that allowed the two side parts to be connected, via the central component. Due to our decision to keep the fixturing the same, these were all taken from the original geometry and used within our setup. The only non-fixture-driven preserves are the lugs at which the shock absorber mounts to the swing arm assembly. This is very important to the loading of the swing arm as it provided the location of the main reaction force to the loading in the overall assembly. (The approximation of the shock absorber will be discussed later.)
The loading of the central component was dictated by a combination of the forces on either side relating to the forces of the side components, and the reaction force from the shock absorber. Figure 6 shows a crude representation of this. The setup of this central component had to be created due to a lack of loading within the assembly definition of the real-world loading. We worked hard to extract different forces and apply them to the remaining preserves.
Generative Design Trade-Offs
As is the case with any design process, generative design is full of trade-offs. For the most part, these are trade-offs of performance (mass, factor of safety, displacements) against ease/cost of manufacture. Due to the nature of our previous work being mostly focused on the mass savings possible, Lightning were very keen to see if we could harness that whilst focusing more on manufacturability and cost. Figure 7 illustrates how this trade-off was performed in a simple yet effective matrix. We compare the original design to our generative design outcomes from 2018 and 2021. Although there was an extra 30% mass savings in the 2018 project, we believe that the focus on manufacturability will allow this part to be built upon, and a production part is coming soon.
Fusion 360 Simulation
Fusion 360 Simulation is a validation tool to help you understand how a design performs under certain conditions. A highly trained specialist could spend much time performing a detailed analysis to obtain the exact results of real-world conditions. However, you can often predict and improve a design based on the trending and behavioral information you obtain from fundamental analysis. If you perform this basic analysis early in the design phase, you can substantially improve the overall engineering process.
Use the analyses in the Simulation workspace to determine how loads lead to deformation and failure, so you can understand if and how a part will fail. Or you can determine natural vibration frequencies to avoid resonance. You can identify temperature distributions and thermally induced stresses.
Save time-to-manufacture, in the Simulation workspace, as you experiment with virtual design variations or adapt your model to changing design requirements. Use the tools in the Simulation workspace to minimize physical prototyping and destructive testing requirements. Fusion 360 Simulation studies that run in the cloud rely on cloud computational services.
- Simulation selection
- Simplifying geometry
- Load cases and boundary conditions
- Results exploration
Static stress analyses are one of the most common types of finite element structural analyses. The component or assembly is subjected to a range of load conditions and the resultant stress, strain, and deformation results are analyzed to determine the likelihood of failure of the design.
Linear static stress analyses assume the following:
- The structure returns to its original form
- There are no changes in loading direction or magnitude
- The material properties do not change
- Deformation and strain are small
Want more? Download the full class handout to read on.
Peter attended the University of Birmingham, graduating with a master’s degree in Mechanical Engineering. He began his career with Autodesk during a summer internship and has since rejoined Autodesk as a graduate technical consultant working in the Birmingham office, recently taking a full-time role in the Process Specialist Team. He has worked on a variety of projects, often focusing on the utilization of generative design within different industries, helping to drive the adoption of the platform and further develop the software. In his spare time, Peter is a keen sportsman, playing football, rugby, and golf on a regular basis.
Richard is the founder and CEO of Lightning Motorcycles. He founded the company after being invited to work on some leading projects within the field of electric vehicles. As a life-long motorcycle rider, he spotted a gap in the market that would allow the fields of electric vehicles and high-performance motorbikes to push the possible performance of these bikes. Lightning has since gone on to hold multiple speed records and race victories, often being the first electric powered bikes to even compete, let alone win these prestigious events.
Nick is a well-rounded individual with 11 years’ experience as an aerospace engineer coupled with a strong academic background. Nick has also gained key exposure in the oil and gas, wind, nuclear, and manufacturing, engineering, and automotive markets. He is a specialist in computer-aided engineering, with an emphasis on multiphysics optimization. Nick has joined Autodesk Research as a senior researcher in Manufacturing Industry Futures focusing research on manufacturing digital twin, creating novel generative design workflows, and smart design. Nick is currently managing Project MOnACO which is funded by the CleanSky program aiming to manufacture the world’s biggest laser powder bed fusion jet engine part.