Lightning Strikes Twice: Revisiting Generative Design for Mass Production

RICHARD HATFIELD: I'm Richard Hatfield, founder and CEO of Lightning Motorcycles. I'm a lifelong motorcycle rider. And engineering excellence has always been my passion.

PETER SIMPSON: Hi, my name is Peter Simpson. And I work for Autodesk as a technical consultant. I work on the process specialist team as part of the Fusion 360 customer engagement organization. My main focus is generative design and finding new ways to harness the power of the software.

Because of this, I was tasked with undertaking the main generative design for this part, and also any of the editing and processing that we needed.

NICK MARKOVIC: Hi, my name is Nick. I am a research engineer at Autodesk Research. My main focus areas are developing new digital twin workflows and developing new manufacturing technologies. In this project, my responsibility was ensuring that the new Swing On prototype treated from the generative design technology was compliant and safe to operate by using Fusion 360 simulation product.

This is the LS218. This is the fastest production motorcycle in the world. It's won numerous prestigious races and set a series of land speed records and it's electric. The story of Lightning really began in 1990, when I was invited to drive and participate in the development of an electric Porsche. And in the process I developed a deep understanding and expertise around electric drive systems and wanted to bring that to motorcycling.

In 2006 I built the first lithium battery sport bike. And in the first test ride up the first hill, became convinced that electric motorcycles were the future. And this was a great opportunity to build a company.

Recently, for the first time in human history, CO2 levels exceeded 415 parts per million in the atmosphere. The USA has the second highest CO2 emissions of all the countries in the world. Transportation is the largest source of carbon emissions in the United States.

And worldwide, motorcycles have higher emissions than cars. And there are significantly more motorcycles in the world than cars. So we believe that building great electric motorcycles, which motivate consumers to convert to electric motorcycles, can contribute in a significant way to reduction of global warming.

Lightning's mission is to build world-class two wheel electric transportation with superior efficiency, performance, and affordability than current gasoline alternatives. We believe two wheel electric vehicles are a great solution for daily transportation and an important solution for reducing traffic congestion and parking issues. We dream that electric motorcycles will replace gas motorcycles in all the major markets in the next 5 to 10 years.

Our goal is to accelerate that Shift to the next generation of personal transportation by offering great products with competitive prices for riders to enjoy around the world. And we believe deeply in this mission and we build it into every bike. When people ride our bikes, whether they're world champions or first time riders, they get it.

The electric motorcycles that have been available for consumers have not met the goals that consumers are looking for. So the range was inadequate, the charging time too long, the performance inadequate, price too high for the build quality. So as a result, electric motorcycles so far have not been able to win the hearts and minds of motorcycle riders over the gasoline alternatives.

So we've really focused at Lightning on developing technologies to address these objections. So we build our products to create excitement and to bring new riders into electric motorcycling. Lightning seeks out competition to drive innovation in our company and then to use that innovation that we develop to create exciting products for the marketplace.

So early on, to drive electric vehicle adoption, we set two milestones for ourselves. The first was to build electric motorcycles that could outperform the best gas motorcycles in the world. And then the second was to build a supply chain that would allow us to provide superior products at competitive pricing to our customers.

So the first milestone was achieved at the Pikes Peak International Hill Climb where we became the first electric vehicle to race against the top gas racing motorcycles in the world and win by a decisive margin of over 20 seconds. So the key to how we've been able to achieve our second milestone, this combination of price and performance, is that our approach from the beginning has been fundamentally different.

We began experimenting with lithium batteries for vehicles in 2006. And the process of developing competition bikes, doing EV engineering contracts, have created a deep fluency in these EV technologies and have created an understanding of how really to optimize the trade offs between power, and handling, and heat, and weight, torque, charge time, and price. And then take that and roll it all into products that have superior overall performance.

So this has allowed us also to develop a network of relationships of engineers and other companies that also are passionate about the technology around electric vehicles. And then secondly, one of the design principles we developed through the engineering contracts is the importance of using a modular platform in EV technology so that we have the flexibility to use similar components in a widely disparate group of applications. So the same type of components can be used in an 80 volt ATV as are use in an 800 volt fuel cell hybrid bus or in an electric airplane. And this allows us to develop quickly and cost effectively.

So there have been a whole series of challenges in building the technology in our motorcycles. So some of the challenges are building the level of performance that we need on something as small and as compact as a motorcycle. Thermal management is a critical issue, achieving this level of performance and not having temperature issues on the components.

Weight is the enemy of performance motorcycles. So the ability to use the tools that we have from Autodesk to minimize the weight of our components, while still achieving the strength required and being able to package this in the very limited space of a motorcycle.

When we began winning on the racetrack, people sought us out to work with us and tap into our expertise. This gave us a chance to work with electric vehicle innovators, entrepreneurs, inventors, engineers who are all leaders in their field. Over the years, we've done a series of engineering projects from electric cars, electric airplanes, electric boats, hybrid buses, ATVs, scooters, and material applications.

And a couple examples of these were, one, a cutting edge battery technology company that contacted us to build a battery, the first battery pack capable of charging in five minutes. And to use it for a demonstration run between San Francisco and Los Angeles. And then secondly, collaborating with Autodesk in the development of their generative design software where we were able to work with them, utilize their software to design a swing arm for a bike that was both lighter and stronger. And this was featured in Autodesk rollout of the software. And we've learned a lot from each of these projects.

So the tools that we have access to from our relationship with Autodesk have been a real key to achieving the level of performance and the maturity of products that we built. On a daily basis, we use Autodesk products like Fusion, Inventor, Alias to do the surfacing, the CFD and FDA analysis to achieve the aerodynamics, and the strength, and the light weight that we need for our products.

So let's hear from the team from Autodesk who've been a major factor in our success.

PETER SIMPSON: Thanks, Richard. Many of you may have seen from the title of our talk and previous AU presentations this is not the first time that we've worked with Lightning Motorcycles. We actually did a similar project back in 2018. And we are revisiting that project and looking at the very same part.

In the original iteration, we set out to see how far we could really push the performance of this part. This was in order to allow Lightning to just continue pushing the performance of their motorcycles as a leader in the field.

This project showed promising signs. We saved a huge amount of weight on the part. And we add it to some part consolidation efforts from their side.

Unfortunately there were some shortcomings with this. And I'll go on to talk about them now. Due to the nature of generative design at the time, these shortcomings were somewhat unavoidable. There was a lack of manufacturing bias within the software in that we couldn't define a manufacturing method as well as we can today.

The other issue was to do with the load case setup. It was much harder to actually translate those real world loads into the generative design software. To try and combat these shortcomings, we decided to leverage the advancements in generative design in Fusion 360 to focus on a more manufacturable outcome.

We focused on utilizing the milling constraints with generative design to reduce the cost and complexity of the part while still allowing valid performance benefits. This would be traded off against some of the performance gains that we saw in the first iteration. But ultimately, it was in an attempt to find a commercially viable part that Lightning could sell on their bikes worldwide.

The advancements made in the past three years in generative design have really enabled us to revisit this project. When we have such a clear trade off of manufacturability versus performance, the progression of the manufacturing controls within generative design is invaluable. It really allows us to pinpoint specific methods that we're interested in and have expertise in.

This allows us to tailor the final outcomes , and ultimately make all outcomes relative to a chosen method. Milling, and specifically three axis milling, was deemed to be an area of expertise that we have within the company. Not only this, but we thought that with three axis milling, you can gain a complex enough geometry while still maintaining some speed and cost relativity.

So the other key advancement that really allowed us to succeed in this project is the advancements in terms of solar sophistication and even the addition of some new solvers within generative design. These allow us better load case attributes and allow us to then translate that into better representation of the real world loading. Understandably, if we have a better representation of the real world loading, we can then be confident that our outcomes are going to be better set up to cope with that once they're produced.

The benefits of the solver are also that we could set this up as a three part solution, keeping it as an assembly. Instead of going for the monolithic approach like we did in 2018, where we could do part consolidation alongside light weighting, we kept this as three specific parts. This allowed the manufacturing to be a bit more simple, cost effective, and time effective.

An often overlooked advancement in the generative design world has actually been the advancements to editability of the outcomes. Because we can now take all of our outcomes out and edit them within Fusion, it makes it a lot easier to make small changes to the designs. This is really valuable in terms of manufacturability, but also when you relate it to the simulation outcomes that we're going to talk about in a little bit.

So now it's time to actually look at the outcomes from this project. In the next section, we're going to be basically comparing the original part that Lightning are currently using, the 2018 generative design iteration, and our current generative design prototype. As you can see, the existing part is simple, but effective. It's manufactured by dye casting.

Due to the mass production of this part, that makes it both cost and time effective. The first generative design iteration from 2018 is highly optimized. We're saving about 40% of the mass of the original part. However, it's really costly to make and almost impossible, in some regions, to really ensure a reliable, repeatable manufacturing process.

Personally, I believe that the final iteration really bridges the gap between the two prior iterations. We're saving about 10% of the weight, yet we're still actually managing to keep manufacturability and cost and time effectiveness at the forefront. Through further development and some further manual edits, I really believe that we can better represent our trade offs and ultimately get a production part on Lightning bikes soon.

NICK MARKOVIC: Undertaking validation studies in a virtual environment is essential in design and make workflows. Numerical validation tools are used to understand how a part or assembly performs under certain conditions. For example, simulation tools are used to calculate how loads lead to deformation or catastrophic failures. This can save time to manufacture.

The user can experiment with virtual design variations or changing the design requirements. This can save costs by minimizing physical prototyping and physical experiments. Fusion 360 has a large simulation portfolio, and it's growing.

All simulation studies can be run in the cloud at an affordable rate. Also, this is a great opportunity to share these results to the Lightning folks on the cloud without the need of manually sending gigabytes of result files. In this discussion, I will perform structural stress analysis using Inventor Nastran to understand if the new generative design swing arm that Peter has created is structurally sound.

The swing arm assembly can be tricky to simulate, as there are plenty of moving and rotating parts. However, Lightning have simplified the load cases for us to perform rapid prototyping and validation. The rear wheel assembly was idealized as a rigid body element. Also we simplified the rear shock assembly by using a combination of simplified solid bodies and rigid body elements.

Lightening has also provided three load cases. The first load case represents the vertical load acting on the swing arm during acceleration and cruise condition. The second load case represents an oblique load acting on the swing arm during the quiet cornering at an angle of 40 degrees. The final load case represents a torsion load acting on the swing arm.

The end plates and center components were bonded using automatic contact function. And these bodies were adequately meshed using parabolic solid elements to enhance the result accuracy.

The final element analysis works by breaking down a real life object into a large number of elements. And mathematical equations are used to predict the behavior of each element and grid points.

The computer combines all of the visual element behaviors to predict the behavior of the global assembly. Now the simulation [INAUDIBLE] has been completed. The final step is to run a series of linear static simulations to calculate the maximum deflections and the maximum stresses that each load bears.

Also, a factor of safety is calculated. This is denoted as FoS. Mechanical engineers commonly use this factor to assess if the stresses are above or below the material strength limits. The main objective is to ensure the factor of safety is greater or equal to unity. Once the cloud solves are finished, the results are automatically uploaded to the simulation of for post-processing.

This graphical user interface is what we use in the simulation results environments. For multiple lood cases, the user can open different window layouts where the result comparisons are easily made. This [INAUDIBLE] shows the factor of safety distribution. In simple terms, blue collar means the region's low stress and the warmer colors indicate the part is well stressed. The red color indicates the stress is close to the design limits, and this is what we need to avoid.

The next step is to find the critical factor of safety. And understand what is real and what is an artifact created from the simulation. You can see that load case three is the bounded load case because it has the lowest factor of safety value.

We can also view the global deformations when the swing arm is loaded. The displacements were exaggerated to understand the global structural behavior. I always use this as a sanity check to see if the model is behaving correctly.

This is the most exciting bit. I have summarized the key results as a radar diagram for each design variant, where the model setups are identical. This analytical shows the maximum displacements for each load case.

Ideally, we would like the radar to be as small as possible, as this will create the stiffest swing arm and thus increasing the overall ride performance. We can see the latest generative design model is much safer than the previous generative design model. This demonstrates how much generative design technology has improved since 2018.

Last but not least, this radar shows the minimum factual safety for each load cased. The red dotted triangle region indicates the factor of safety is less than 1. And therefore, this means the design fails to meet the criteria.

However, the latest generative design triangle points are outside the red perimeter. And therefore this meets the design criteria. However, the 2018 version fails to meet the latest requirements for load cases one and three.

This concludes the simulation chapter. And I'll pass it back to Peter and Richard to close the show.

PETER SIMPSON: Thanks for that Nick. Ultimately, I think we can really take a lot from this project. And I think a lot of these can be categorized into three key takeaways. Takeaway number one is that the constant advancements within generative design mean that we can really reimagine what is possible with this software. The second takeaway is all to do with the benefits of collaboration within Fusion 360.

Due to it being cloud based, there's no longer a need to send documents back and forth. This really allows people to work in real time collaboratively in between different teams and even between different companies. The final takeaway is to do with Fusion 360 once again.

The ability for us to complete the whole design and make workflow within one system was really invaluable for this project. Next, I'm going to hand it over to Richard to let their next steps in order for Lightning to take this part to production.

RICHARD HATFIELD: Thanks, Peter. Our current priority at Lightning is to scale our production to meet the demand for electric motorcycles. And our goal was to see hundreds and thousands of Lightning bikes on the street everywhere around the world. We really want to develop the products, make the products that push the transition from fossil fuels to renewable energy. And our goal at Lightning is contributing to making the world a better place. And doing that one electric motorcycle at a time.