By Martin Kubie, Proxima Fusion Co-Founder and Head of Engineering

It's hard to miss the impact of computing in our everyday life, but how much have computing advances changed the way we design hardware? Historically, engineering was driven by laborious hand calculations and highly specialized drawing offices. Today, not only can engineers rapidly produce 3D models and evaluate them in multi-physics simulations– they can even automate the generation of those models. It’s possible to evaluate entire families of designs before the first prototype is even built! This shift in approach is especially evident in cutting-edge industries that are working toward deep decarbonization.

At Proxima Fusion, we make extensive use of simulation as a means to accelerate our path to building stellarator fusion power plants. We’re a stellarator company, not a simulation company, but we are big believers that smart software can massively accelerate hardware development.

There are three main reasons for that:

1. Simulation isn’t a substitute for actual testing, but it is an efficient way to speed up that process. We’re iterating on designs faster so that we can build faster, too. There's no more valuable model than a simple experiment, but simulation helps us find the right experiment.
2. Simulation increases confidence and reduces the overhead of systems integration. Through simulation, we’re able to predict the integrated behavior of the many sub-systems of a stellarator. By enabling all our engineers to rapidly observe the impact of design decisions, system engineering becomes embedded in our daily workflows.
3. Our current simulation-driven approach is a stepping stone to the data-driven approach we aspire to. We’re using our simulation pipeline to generate data and train AI models on that data, which will accelerate our path to the most valuable simulations and results going forward. We use simulation to find the right prototypes, and we can use AI to find the right simulations!

The role of simulation in Formula 1 and autonomous aviation

While working as a vehicle dynamics engineer at McLaren Racing, I saw firsthand how simulation enables rapid iteration while optimizing the design of F1 cars. Track and wind tunnel testing time is extremely limited, so leveraging simulation-driven design is a vital way for teams to gain an edge over their competitors. It even plays a crucial role at the track, where teams run extensive simulation studies to optimize performance between racing sessions.

A simulation-driven approach to engineering was even more critical during my time leading simulation development at Wing, a Google[X] spin-off company in California. Reliability is paramount when designing autonomous robotic systems like delivery drones. We developed verification platforms that ran millions of simulated flight hours for every update to our autopilot systems—ensuring smooth operations when making deliveries to real homes.

What simulation-driven engineering means for stellarators

There are several parallels between designing F1 cars and fleets of autonomous aircraft, and designing stellarators. Building a device as highly complex and large-scale as a stellarator requires first gaining significant confidence in performance and reliability.

Proxima’s priority is speed. To put stellarator fusion power plants on the grid in the 2030s, we need to be able to take a bold leap with our first machine, as opposed to incrementally building slightly better devices over a period of decades.

Just as designing an F1 car or an aircraft requires determining the correct trade-off between lift and drag, designing a stellarator demands trade-offs between plasma performance and engineering disciplines—from thermal to structural to electromagnetic. These interconnected design decisions can be made faster and with greater confidence using our integrated simulation tools.

The result? We can move to production with less risk and fewer prototypes.

Building stellarators to power the future

As we work towards the demonstration of our HTS magnet technology, prepare to publish our concept for the world’s first stellarator fusion power plant, and finalize the design for our net energy stellarator, we need a multidisciplinary team with a wide range of expertise. From parametric geometry specialists and numerical analysis experts to fusion design leaders and software engineers, working with a world-class team of innovators will allow us to make commercial fusion a reality.

You can learn more about our approach to simulation-driven engineering in this video: