ESTECO modeFRONTIER UM shows how CAE fosters creativity & innovation
I had the great fortune to spend some of last week in Trieste, Italy, with an amazing group of modeFRONTIER users. We’ll get to them, and what they had to say, in a second but, first, you need to know that Trieste is gorgeous and ESTECO knows how to throw a party that celebrates its hometown as well as its user community.
ESTECO organized a scavenger hunt that sent teams of user meeting attendees around the city. This photo is team AFILTERSQP with James Joyce, the Irish writer, who seems to have spent a significant chunk of his life in Trieste. Each team had a list of sights to see and photograph –from the Roman amphitheater and arch to the massive central square with its impressive fountain– and extras to find, such as a woman in a hat and kissing lovers. AFILTERSQP, alas, didn’t win a prize but that’s OK — we had caffè in B along the way, made by a very amused barista. Trieste was, for a time, the coffee capital of the world and takes its coffee very seriously; caffè in B is espresso served the Triestine way, in a small glass, not a cup. The things one learns while traveling …
ESTECO was founded in 1999 as the result of a European Union-funded research project on design optimization. The EU’s Frontier project included a “who’s who” roster of partners: British Aerospace, Daimler-Benz Aerospace, Electrolux-Zanussi and other commercial enterprises, the universities of Trieste and Newcastle and the Defence Evaluation and Research Agency (of the UK), among others. The project’s aim was to develop a technology able to drive industrial competition throughout Europe. The end-result was a commercial product for multi-objective optimization called modeFRONTIER and the formation of ESTECO to keep on developing it.
What is multi-objective design optimization? I’m going to borrow the example from the modeFRONTIER class I attended last week. Let’s say you need to find the best possible steel I-beam for your needs (I-beams are also called H- or W-beams in other parts of the world; I’m going to stick to I for simplicity). For our application, it needs to be as light as possible while supporting the planned load with the smallest possible deflection.
The I-beam is specified by the section thickness, a; its height, c and width, b. What combination of these three dimensions can best meet our criteria?
Let’s say that you can get or make an i-beam where a can be between 0.001m and 0.01 m; b can be 0.05m to 0.15m and c, 0.05 to 0.25m. You’ll also need to know the Young’s modulus and beam density (material properties of the steel) and the proposed length and load on the I-beam.
You could crunch on this for days, and come up with a number of alternatives, or you could use an optimization tool like modeFRONTIER to do the heavy lifting for you.
In modeFRONTIER, you define the problem’s inputs and outputs, what you want to hold constant and what to optimize, how you want to calculate, and what your constraints are. Setting this all up in modeFRONTIER takes a few minutes; you wind up with a workflow that describes what you want to do:
Here, a, b and c (at the top) are the inputs; stress, deformation and weight are the outputs (near the middle). We’re constraining the optimization with a maximum stress of 100 Mpa and a max weight of 20 kg (at the bottom left and right). Lastly, the optimization objectives are to minimize the deflection and weight (center bottom).
The center row of the workspace, the DOE to checkmark at the Exit, define the “black box” that computes the outputs based on inputs. These could be CAE solvers that numerically model and solve the design problem or a branch that goes to CAD or another program.
In this case, the math is pretty simple and we don’t need to plug into an external CAE tool. We type the formulas into modeFRONTIER’s calculator (the center of the diagram above) and tell it that we want to try 20 different “experiments” or combinations of inputs to see what we find.
These experiments make up a design space. Our problem’s results are below; a couple of alternatives in the 20 that exactly meet all of our criteria, a few meet some criteria and others that simply won’t work.
Tables are great, but limited, so most design space tools have lots of visualizations, like the Pareto frontier in the lower middle.
Outside the training classroom, problems are much more complex. At the modeFRONTIER user meeting we heard about refining the shape of a car to minimize drag, how to design the acoustics of the interior of a car for passenger comfort, what it takes to develop motorcycle racing engines, best practices in automotive light weighting, and so much more — but all highlighted the use of simulation for innovation, not necessarily to replace bend-and-break testing.
This is huge. CAE has traditionally been done to avoid or reduce the need for physical prototypes. Those are expensive, and testing often takes too long to be part of the mainstream design process. Removing some of that time and cost enables products to get to market more quickly and perhaps cheaply —but this use of simulation doesn’t help in innovation. That’s problem avoidance, not enabling creativity.
Presenters at the modeFRONTIER user meeting spoke about discovering new design alternatives that weren’t visible to even the most experienced creator before they did their design space exploration — humans tend to bog down in what’s worked before and what we think can’t work; computers have no such limitations and, if the problem is set up to not limit the possibilities, can expose innovative alternatives that should at least be explored.
I am used to automotive, aerospace, and other big industrial examples. One of the highlights, for me, of the modeFRONTIER user meeting was hearing how illycaffè uses design space exploration and optimization to make better espresso. Here’s what I learned from Luciano Navarini of illycaffè:
- Over 50 million cups of espresso are consumed every day, worldwide. The planet runs on caffeine.
- illycaffè founder Francesco Illy developed the modern espresso machine, a gleaming thing of dials, nozzles and tubes, in 1933
- That same year, Alfonso Bialetti used innovations in aluminum forming to come up with the moka pot, the two-chambered stove-top espresso maker we’re all familiar with. The outside of the pot has changed with design fashion over the yeas, but the inner chambers are still largely as they were originally designed 80 years ago
- Aficionados often detect bitterness, acidity, sweetness, chocolate, and other positive notes, but also burnt, smoky, medicinal, metallic and astringent flavors when the espresso is brewed in a moka. Mr. Navarini went on a mission to try to figure out how to make a moka that created none of the off-notes while preserving the positive attributes. (Starting to sound like an optimization already, no? Maximize sweetness, minimize medicinal …)
- The moka uses steam pressure, created when the water in the bottom of the pot is heated. The steam forces water through the coffee grounds and upwards into the top chamber
- Controlling that pressure is difficult on a typical stove top. Mr. Navarini compared it to a volcano (Stromboli), where very intense evaporation leads to the extraction of many of the undesirable compounds in the coffee. One goal for the perfect moka, then, is avoiding the strombolian phase by carefully crafting a pot that better controls temperature and pressure on a normal kitchen stove
- Mr. Navarini and his team at illycaffè and the University of Trieste made a CAD model of a moka for a multi-objective optimization to try out different geometries until they hit their targets for the various flavor components. Sounds simple but it wasn’t. Their exploration found that coffee extraction starts at low temperatures, lower than most people are aware, and that flavor depends on the initial amount of dry air in the kettle — not water
- In the end, the result was a lovely design for the interior of the pot that has much more curvature than the traditional moka. To go with the sexier interior, illycaffè worked with Italian design firm Alessi to create a beautiful new moka, the Pulcina. You can read more about Mr. Navarini’s research here and the Pulcina here. Amazon had my Pulcina waiting for me when I got home from Trieste and, while I’m no expert, it does make delicious espresso with no volcanic notes.
Other speakers highlighted how they want to use simulation as a communication vehicle, to work with colleagues on complex designs. ESTECO’s new SOMO is an enterprise collaboration and distributed execution framework that aims to help groups deal with the complexity of running multidisciplinary design projects. Connecting NVH with CFD, for example, and ensuring that decisions made in one group do not cause undisclosed problems for another will be a huge step forward in many organizations.
Mr. Navarini’s example is easy to relate to and points out the most compelling use case for simulation: to create a revolutionary product. It starts with the question: how can I make better espresso on a stovetop? modeFRONTIER enabled illycaffè to determine which were the most important design variables in a moka and the relationship between them. That, in turn, led to a design space with a few alternatives to further explore and, ultimately, the Pulcina.
After the conference ended, a journalist asked me about the future of design. I think this is it: simulation-driven design. If we do it right, it isn’t constraining creativity but nudging it in new directions. If a design space exploration turns up even one alternative we hadn’t thought of or exposes an interdependence we weren’t aware of, this product will be vastly superior to the last. Putting simulation early in the design cycle, making it an integral business practice, can make design fun again. it creates a sandbox in which we can play with designs and reframes the design process from “not disappointing the customer” to creating delight.
Note: ESTECO graciously covered some of the expenses associated with my participation in the event but did not in any way influence the content of this post. The title image is of Magazzino 42, a recently renovated former dock warehouse that was the venue for the event. That image and those of modeFRONTIER in action are courtesy of ESTECO. The photo of AFILTERSQP was taken by a friend of one of the team members; I apologize for not getting your name.