In the northern German city of Bremen, a unique factory filled with scientists is churning out new materials, not in ones or twos, but by the millions. These are the materials on which we’ll build new aircraft, new spaceships, new medical implants: anything new that needs something stronger, lighter, tougher or more durable than before.
The Center for Materials Science in Bremen has industrialised the science and business of invention. Like any factory, it needs its own raw materials. Alongside the contents of the periodic table and the intelligence of its researchers, the Center has one key component: data.
By generating and consuming data at a truly industrial scale, each of its researchers becomes thousands of times more productive than the usual materials science PhD. The Center's approach is an instruction manual for the enterprise on how to use data to meld scientific thinking into process.
Dr Nils Ellendt, CEO of the Center, says:
Every advance in materials science has been made by accident, and we don’t have the time for that anymore.
The answer is an industrialization of the inventive process, powered by a factory line of creation and testing. This is underpinned by the center’s new approach to collecting, collating and using data.
Creating new ideas in the enterprise is mostly the result of the human process. Analysing markets and seeing opportunities to do something different, better or quicker than what currently exists. The idea of automating innovation is alien. However, the Center has proven that it can be done by creating and testing millions of new compounds. It doesn’t remove human creativity, it elevates it.
The researchers create new materials by taking a basic idea – say, a droplet of a new steel formulation – and putting it through thousands of automatic variations in heating, cooling, compression and so on. Every step, every variation and every test is captured as data. As each new material progresses through creation, test and categorization, this data follows it. The end result is stored in a huge dataset that can be searched, sorted and categorized
Dr Ellendt explains:
We store every bit of information about the production and testing process, as well as the results of the tests themselves, and this is new in materials science.
In science, all results are valuable. Even if it isn’t immediately obvious. Take the laser, for example, which when first created was described as ‘an answer in search of a problem’. However, the laser eventually went on to spark a huge part of our information revolution. And this was only possible because it already existed and could be recreated at will.
When creating a new service, an enterprise will likely be interested in: how much revenue did it generate? Or, how many people used it? But this is a tiny part of the story. There are thousands of questions you can ask of a new idea as it goes through the pipeline and into the real world. How long did it take to develop? And with how much resource? What didn’t get created because we worked on this? But this information is rarely kept, let alone collated and made searchable.
Dr Ellendt, explains:
Each innovation is its own case study, one with lessons for every department, but only when the big picture is available to all.
One of the greatest barriers to building the data-illuminated enterprise has been the inflexibility of relational databases, which force data into patterns before they can even begin to be analyzed. That was the big issue facing the researchers, who were creating data in a bewildering variety of formats, from single numbers to tables, graphs, images and multi-dimensional scans.
Dr Ellen explains that while thoughts initially centered around structured databases, “we quickly realised that unstructured data was the way to go.”
We didn’t have the luxury of data in standard formats. We had to store it in very accessible ways no matter what it was, and efficiently work with it.
The center’s research uses small steel droplets or microsamples, which are put through a series of tests and evaluations. By using these microsamples researchers are able to learn more about the properties of the sample with less material, less cost and much less time.
Once the data is stored and managed, its usefulness increases exponentially. The data can then be processed in ways that reveal what it’s saying, in ways that can be searched and compared. Dr Ellendt says: “We move the data itself through a pipeline into documents, making it available for automated processes. MongoDB is uniquely suited to our purpose.”
An enterprise that takes a scientific approach to all its data, and considers its innovation process as a rich source of information for the future, is refining and recycling its most precious resources for its own use.
You can read the full University of Bremen data story on MongoDB.com: Creating the material world through data, one million inventions at a time