Henry Ford was fond of saying that “nothing was particularly hard if you divided it into small jobs.” He followed his own advice, built the world’s first large-scale assembly lines that cranked out millions of Model Ts every year, and left his competitors in the dust. Engineers are now taking Ford’s advice to the extreme and breaking down the factory into even smaller pieces: bits and bytes.
“Software, data and analytics are changing what we can make in ways I couldn’t imagine when I left school,” says Christine Furstoss, global technology director at GE Global Research. “It’s more than 3-D printing jet engine parts from a digital file, which we already do. We can build a factory that can make itself better. We call it the Brilliant Factory.”
Furstoss recently talked to GE Reports editor, Tomas Kellner, about the concept.
Tomas Kellner: What is the Brilliant Factory?
Christine Furstoss: The Brilliant Factory is a commitment by GE to use the latest technologies to consistently, and in real time, optimize our operations. It’s really a mindset. We can now collect more data than ever before because of better sensors and better control systems, and we’re able to analyze that data better than ever before because of advanced software and analytics. Finally, we are better able to take action, to understand how well our processes are working inside the factory. It allows us to understand and handle things when they don’t go right, and do that in real time, maintain the highest quality and on-time delivery.
TK: What do you mean by “mindset”?
CF: The reason I say that it’s a mindset is because GE has 400 factories around the world and these factories conduct many different types of operations and make many different types of product, from lighting systems to power generation turbomachinery and healthcare equipment. As you can imagine, no two factories look identical. Some make physical parts and some assemble parts, some repair, remanufacture or service parts. They all have needs that are a little bit different.
With the Brilliant Factory, we’re creating a toolkit with different solutions. A solution can be a sensor that collects data or a type of program that analyzes the data, or it may be a control system to change the machine and adapt in real time if we see something happening. Our new software platform, Predix, is a very important part of our solutions.
We’re learning how to apply these different tools at the right places and at the right time in our factories. Just like no two factories are identical, no two brilliant factory solutions will be identical.
TK: How did you come up with the name?
CF: The reason we call the solutions “brilliant” is because it really is about leveraging the best we have, and really be able to look at a problem in a unique way and supply great technology for the solution.
The brilliant factory is a commitment to technology and continuously changing the way we work, and putting the right solutions at the right places at the right time.
TK: Why are we talking about the concept now? Could you build a Brilliant Factory, say, five years ago?
CF: The opportunity to apply new types of thoughts to our factories really arrived over the last three or four years, especially with the emergence of the Industrial Internetand the ability to handle data sets that are very, very large.
As you can imagine, when we’re machining a part, drilling a hole or putting new materials together, we can collect lots and lots of important data that was previously out of reach. Being able to transmit that data, store that data, figure out which data points are important and then do the analysis; that has been a huge step-change in enabling the Brilliant Factory.
TK: What type of data are you talking about?
CF: It could be anything; how well the machines on the factory floor behave, how well and repeatedly they react to the materials that we’re putting through the factory, such as how do they respond to the cutting tool. The types of ways that we now take those measurements, the things that are in the machine that sense and store those measurements, and the ability to quickly change the configuration, those have also changed.
I think that we’re at a very unique point in time where we have this convergence of the hardware - the actual physical machinery on the floor, the sensors, the control systems - with IT infrastructure and software. They are both now converging on this nexus we call advanced manufacturing that is enabling the Brilliant Factory.
So, to answer your previous question, could we have done this five years ago? You bet, but we wouldn’t be able to leverage the concept from one factory to another as fast and as easily.
TK: Can you extend the process and create a feedback loop where manufacturing informs product design in real time?
CF: Absolutely. That’s why we call the process “brilliant.” This is about a mindset, as I said, that we’ll continue to optimize and continue to get better and, most importantly, continue to learn.
These feedback loops are at the core of the Brilliant Factory, whether they come from our plants, or remanufacturing and service shops fixing parts after they’ve been used by our customers for a long time.The feedback helps us understand whether the parts we designed could be manufactured with the features and materials we specified, within acceptable cycle time, cost and yield. Equally important, we can find out after being in service how the parts looked and behaved. We can send that information to our design teams and to their own software-enabled design tools, so we can validate and learn.
I’ve already mentioned Predix, and these types of activities show the true power of the platform. It’s allowed us to interface with lots of different tools and components. It helps us take data from factories, which literally can generate hundreds of gigabytes per hour, depending how large the factory is, and sort it, structure it and feed it back into the system, including the designers.
TK: Can also make supply chain and distribution “brilliant”?
CF: Exactly. We are already working with a lot of key partners and suppliers. As you know, we ship products all over the world, and customers expect parts on time, they want them reliable, they want to be able to plan. The same sets of tools that we are using on our factory floor apply to our suppliers, the whole distribution network and our inventories, so we can effectively manage the whole value chain. In allows us to be responsive to our customers in a very cost-effective and value-added way.
TK: What savings can a Brilliant Factory deliver?
CF: There’s no one blueprint or pattern. We don’t want to invest in sensors and tools if a shop doesn’t need it. The Brilliant Factory is all about productivity, gross margins and all the things that we are dedicated to in order to be best-in-class for our great products, not just on performance and reliability, but on operations and responsiveness . We’ve taken a look and said that if we could get across the whole of GE - not just one factory or product, but the whole company - if we could just get one percent productivity savings in our total supply chain, that could be up to $500 million for GE.
TK: GE recently opened a factory in Pune, India, which applied some of the Brilliant Factory principles you’ve mentioned.
CF: That’s right. We’ve put in the infrastructure in Pune and we’re just starting to flex it and understand it. But, remember, the Brilliant Factory isn’t just something that’s for a new factory. In fact, we are seeing the biggest impact in term of saving cycle time and improving quality in our existing factories. Many of those factories have older infrastructure. By putting in these new pieces of software and hardware right into the fabric of what we already have, we’re getting huge cycle time reductions, higher quality and more repeatability. That’s why I’m excited about the Brilliant Factory, because it isn’t for one subset of GE, it’s not just for new factories. It really can have tremendous impact inside a factory where we perhaps haven’t had as much investment over recent years.
TK: What do you see on the horizon?
CF: We’re just starting to scratch the surface in what I’ll call adaptive technologies. We have so much that’s becoming digital now.
These feedback loops that I mentioned are very important. We haven’t fully connected all of the digital threads. We now have many of our design rules - the handbook we follow to design our products - all digitized. Our material properties are digitized. We can gather information on how a machine is working and whether a cutting tool is in the right place.
But what if we make a mistake and something doesn’t go right? What if instead of scrapping the part, we could immediately review the design intent, the rules and the path, and understand how to adapt the next manufacturing process to give the assurance that the part can meet all of the intent of design and expectations of the customer if we make an adaptation.
We could do this in seconds, instead of taking days or weeks to analyze that data. Such truly adaptive and optimized decision-making could apply on the discrete process level, not just to factories. I think that’s the vision.
TK: When are we going to get there?
CF: I don’t think it’s going to take us 10 years to get there. We are putting the pieces in place.
It sounds somewhat trivial, but when we get parts back for service from the field where they worked for six or seven or even more years, they don’t always look like they did when we first shipped them.
But what if we’re able to uniquely look at every part quickly as it comes in the shop, assess what needs to be done, and put it on the right track, right away. Just like there is personalized medicine, we will have personalized manufacturing. That’s the power of the Brilliant Factory, and we’re off and running on our journey!
For more stories like this, visit GE Reports.