Heat exchangers are at the heart of heating and cooling, which accounts for 40% of energy-related CO2 emissions. As global temperatures continue to rise, we are against the clock to heat and cool our world more efficiently.
With Additive Manufacturing, the complexities of objects and machines can now be independent of manufacturing costs, allowing for more design freedom for us to create structures that transfer thermal energy more efficiently. However, the laborious, human-driven process of creating such structures becomes the bottleneck.
Together with TRUMPF, the German industrial machine manufacturing giant, Fabio and his team tapped into the power of algorithmic design to create a series of heat exchangers that are as functional and elegant as nature.
The heat exchanger has been the go-to showcase trophy for players within additive manufacturing for a while now, but we haven’t really seen any radical innovation in the design and engineering paradigm of these objects.
Now, imagine that you have just optimized the heat exchangers within an air-conditioning system. Can the same heat exchangers be adapted for radiators in cars? Maybe, but that may require us to go back to the drawing board. That is why, going forward, the engineering approach with which we create objects will become more important than the blueprints of the objects themselves.
Many inspirations for the problems of heat exchange can be found in nature, where cross pollination and tight integration of functionality govern the way structures are grown. What I find extremely interesting is how components can be made “interdisciplinary”, which not only blurs the boundaries of that component’s functionality, but may also give it more than one key function to play.
Through a series of heat exchangers, we wanted to show how such inspirations from nature can be incorporated into objects created by man. The results also speak for themselves — our shell-and-tube heat exchangers have surface-area densities that are 14 times more than cylinders of the same diameter, with minimum pressure drop.
How do you think Algorithmic Engineering accelerates innovation?
Algorithmic Engineering is about framing design goals as mathematical models using Hyperganic Core, our software platform. It marks a significant shift in my task as an engineer by moving away from a practical, manual engagement with a part’s geometry to a higher-level approach. This involves operating on a more abstract level with an emphasis on solution principles. That way, my engineering workflows become flexible recipes, or input-output functions, that are agnostic to the values of case-specific constraints, but are capable of solving all possible instances of a given type of generic problem.
Once the coding part is in place, we basically have a black box that uses Artificial Intelligence strategies to automatically execute the geometry generation process and create multiple design variations.
I give some input parameters and all I have to do next is to wait for an output, which it always produces successfully thanks to Hyperganic Core and it’s incredibly robust proprietary geometry kernel. This then allows me to automate design and testing to create a closed iterative loop, which is not unlike evolution, only much faster. This will lead to the perfect heat exchanger for the specific use case that was modelled as input.
Further, the principles and learnings that I gathered from this project can be applied almost instantly to another. Engineers starting to solve new challenges are now tapping into this vast pool of shared knowledge, exponentially accelerating innovation.
What are some challenges in creating heat exchangers algorithmically?
An interesting challenge we faced was to ensure that the total cross section of the fluid domain is always constant at each stage of the heat exchanger. This was needed to keep the mass-flow constant.
To address that, a mass-conservation law was quickly embedded in the algorithm, applying the constraint to the whole branching structure. Another challenge was posed by the presence of local minima in the branching areas that caused interesting challenges in terms of printability. We addressed this once again by updating our algorithms.
Finally, the heat exchanger must be post processed on a traditional CNC machine. Two lateral rings were created for the heat exchanger to withstand the pressure of the clamps holding it in place during that process.
What do the results of this project mean for our future?
Our algorithmically engineered heat exchanger has a surface area 14 times higher than a cylindrical tube of the same dimensions. The pressure drop goals were reached after just a couple of design iterations and the overall heat exchange properties exceeded initial targets. When we look at key challenges we must solve in the coming years, efficient heating and cooling systems play a key role in a sustainable future.
I also find the idea of self-replicating machines quite fascinating. This heat exchanger is, in a way, an example of that because it will be used as a component in the same metal printer that was used to print it. I’m sure we will see more of such examples in the future as Additive Manufacturing finds its way into more applications.
Over time, we will see entire machines being additively manufactured, removing limitations of integrating 3D-printed parts back into traditionally manufactured machines. We are certainly pulling the future closer to the present. The sooner we start building a universal library of algorithms that describe objects, the faster we will reach a point where our combined effort will see exponential growth and bear fruit. The flexibility of a software-based approach, describing objects using computer code, like we are using in Hyperganic Core, will allow these building blocks of algorithms to recombine and morph into objects that are ever more complex, sustainable and functional. I look forward to the endless possibilities that the future holds.
Born in Germany, raised in Italy in a family of musicians, Fabio studied aerospace engineering at Politecnico di Torino, always seeking to bring the creative and aesthetic part into engineering.
During his master‘s degree at the Technical University of Munich, he specialized in fluid flow simulation being intrigued by the inner workings of turbulence. As a strong observer, Fabio always looked into natural patterns to find answers and get inspiration.
Starting at Hyperganic straight after graduation empowered Fabio to use his passion of cross-polinating different domains to build completely new products.
With the use of Industrial 3D Printing and Hyperganic’s new paradigm Fabio created automated workflows and objects in the areas of Heat Exchange, Protective Gear and Footwear, pushing the term Algorithmic Design to new heights.