Published 18 May 2022
An exciting week has passed and we’ve been making headlines with the algorithmically engineered 3D-printed Aerospike. Many articles have been written. The one below gives a great summary, thanks to Kerry Stevenson at Fabbaloo for the additional research on the topic:
One question I get asked a lot: “Have you tested it, does it work? Is it just a piece of art because it sure looks like one.”
Here’s my answer: this engine is a valid Aerospike design. It builds upon the latest knowledge of space hardware engineering, and it incorporates the Hyperganic body of knowledge for heat exchanger designs. The Aerospike concept is well-known and understood. The first designs appeared in the 1960s and 70s. As I’ve written in my previous blog post, the Space Shuttle was supposed to have a linear Aerospike, but NASA had to settle on conventional bell nozzles because they could not make the Aerospike work. It was impossible to cool the spike using conventional engineering and manufacturing techniques. In a way, an Aerospike needs to be a giant super-efficient heat exchanger that uses cryogenic liquid oxygen to prevent the spike from melting. The designs we present are technically sound and ready to be tested.
More importantly, these are not just two separate designs. They are representations of an almost infinite number of variants that can be created from the same algorithm. If you look closely, you can see that both designs are quite different, even though they look similar at first glance.
They were reengineered from the ground up using the Hyperganic Core platform for different usage scenarios. The beauty of it is we can incorporate any insights gained from simulation, expert feedback or physical testing, and generate a new variant within minutes. Even if a particular specimen has issues, we can easily generate a new one by either varying the parameters or modifying the algorithm.
What we are doing at Hyperganic is moving engineering to a software paradigm. In software, you are concerned with building a process to create a valid output. In engineering today, what the engineer does is the process, and the result, after months of work, is merely one output. If this sole result is wrong, then you, the engineer, are in trouble because a redesign may take months again. In software, you can run it again, test your code, and modify the algorithm and parameters until you are satisfied.
In conventional engineering, you are obsessed with getting it right the first time. In Algorithmic Engineering, you focus on opening the right parameter space and then iterate until you converge on the perfect solution.
With this project — one of many — we are building new base capabilities for NewSpace companies. Our goal is to lift you up to a higher starting point from which you can converge on a design that works for you, and which incorporates your own proprietary IP. We are not a space company, even though we employ many aerospace engineers — we provide you with a code base as a starting point to build your own technology stack.
Once you move to Algorithmic Engineering, there’s no looking back because any competitor that still relies on manual design and lengthy iteration cycles will have a hard time keeping up.
If you are at RAPID/TCT in Detroit this week, please visit us at the EOS booth.