Siemens is a company that is serious about additive manufacturing (AM). As I mentioned in an article I wrote earlier this year, Siemens is vast—its preliminary 2017 global revenue has just been quoted at 83 billion EUR (97 bn USD), up 4 percent on last year.
Last month I was invited by the Siemens Energy & Power Division (Siemens EPD) to visit its dedicated facility in Finspang, Sweden, to get a firm grasp on some of its dedicated AM activities.
We landed at Linkoping in the East of Sweden and were driven to Norkkoping, home to a respected and digital tech focused university, then shared a transfer with Amy Pempel, senior communications consultant, and Aymeric Sarrazin, senior vice president (SVP) controls and digitalization; both of whom had flown in for the event from Siemens Power Services Division in Orlando, Florida. The following day I visited Finspang, where Siemens has a huge facility of 4 square kilometers, dedicated to the development and production of gas turbines.
Our hosts had organized a digital presentation and dinner reception for us the evening prior to the site tour and it was a rich cultural mix of fine food and fine conversation overlooking the Motala ström river.
From Siemens, I shared a table with Vladimir Navrotsky, AM technology and innovation manager, and Thorbjörn Fors, chief executive officer (CEO) Business Unit Distributed Generation and Compression along with Ana Hernando and Livdia Montes. So, we had a Russian national, living in Sweden; a Swedish national commuting to Germany; two Spanish nationals—one living in Madrid and the other in Barcelona—and a Brit, mid-Brexit.
As well as getting an early snapshot of the real passion Vladimir has for additive tech, I learned a great deal about the region we were in. This province of Sweden has a long and impressive industrial heritage dating back more than 650 years. Textiles was the first sector to make its mark here, and the region was renowned for producing cloth and uniforms for armies all over the world. Paper and pulp factories came later on and then in the 17th century (1627 to be precise) the first canon workshop was established and the production of canons here grew to a peak export capacity of 800 canons per year to the rest of the world, until 1911 when the workshops were closed.
However, the same favorable geographic features that supported the production of canons—namely plentiful forests, water and a deep water port—were also conducive to the production of turbines. In 1913, Staal acquired the empty canon workshops for the production of steam turbines, mostly sold for use in marine environments. Staal became ABB, then later Alstom Power, until the facility was acquired by Siemens in 2003. Steam turbines were produced at the site for 95 years, although gas turbines were introduced in 1945 for aerospace applications.
Today only gas turbines are produced in Finspang, a city with 21,500 inhabitants, but the wider region remains an industrial powerhouse. The largest employer is SAAB with 6,000 employees, followed by Siemens with 3,300. Around one half of Siemens’s Finspang staff are engineers and around 900 work in the service department (aftermarket). The workforce is truly international with 40 different nationalities represented, of whom one fifth are women.
Even the smallest turbine model, the SGT 300, is a vast and complex machine, and this is scaled up in size and complexity for each model in the gas turbine portfolio. To convey some idea of the scale and complexity of these machines, the average number of components that make up a gas turbine is 21,000.
The theme of this event was digitialization, a connected strategy for manufacturers to more functional things in a smarter and faster way. Siemens’s software development heritage is foundational to its digitalization strategy, and intelligent connectivity across the business is the goal, using tools such as virtual reality (VR), augmented reality (AR) and AM.
For Siemens EPD, AM is just one tool within the digitalization strategy but an important one, and it continues to grow in importance year on year. Today the company is using metal AM for rapid prototyping, rapid manufacturing (serial) and rapid repair.
Siemens EPD’s first engagement with metal AM was in 2008, according to Vladimir Navrotsky, and the first machine (from EOS) was brought in-house the following year. The original application development centered around a repair solution for turbine burners. Each of these is about 18 inches tall, maybe 6 inches in diameter, with complex geometry (cooling channels, etc.). They are critical functional components in the operation of the turbine, which means each is subjected to the most hostile and aggressive conditions, including immense temperatures (over 1,000˚C) and corrosive fuels, and they are often in remote locations. The largest SGT 800 model utilizes up to 30 burners at one time.
Operating in such hostile conditions has traditionally meant that failure rates were high, even with planned maintenance and replacement scheduled on the turbine. Production schedules using traditional manufacturing methods for the burners were quoted at 44 weeks, so if a burner failed before scheduled replacement, there was significant downtime and costs involved for Siemens EPD customers. It was a recurring problem.
Siemens EPD found a solution working in partnership with EOS. When a burner reaches the end of its life or fails, it is only 20 percent of the component that is redundant, but the whole component had to be discarded and a new one produced as a replacement. However, Siemens has developed a repair process whereby the top 20 percent of the worn burner is removed and then replaced using AM.
To achieve this, the EOS machines had to be adapted to accommodate the remaining part of the burner directly below the build plate, precisely aligning it within the machine to add the new tip. Using a nickel-based alloy, the part is reconstructed, with an improved chemical composition and functionality that meets all of the regulatory standards, and can then withstand another 20-30,000 hours of use.
The digital qualification for this comes from specially developed algorithms that monitor every single layer of material and identify any deviation. It takes four weeks, less than 10 percent of the traditional production time. This application of AM is truly remarkable, I have never seen anything like it. And today, all of Siemens EPD burners are repaired this way, on site in Finspang using the adapted EOS M400 platforms. There are four dedicated machines for this application running at full capacity every day.
That’s not quite the full story, though. This is where the concept of digitalization comes into sharper focus. Siemens EPD is applying a full digital strategy for its clients in the way it supplies and services its gas turbines. By exploiting virtual reality (VR) and augmented reality (AR) with its own software applications, it is possible to monitor the burners’ performance and condition in-situ, remotely and in real time to prevent failures altogether. In this way, the four-week repair process can be scheduled to minimize and even eliminate downtime. This is now the only repair and replace service Siemens EPD offers its clients; the old process, which ran in parallel with the AM process for a while, is now completely redundant. And failure rates are now less than 1 percent.
It is hard to overstate what an important milestone I believe this is for AM, not just in terms of what is possible now, but for the vast potential it opens up. I’d put this up with the LEAP engine nozzle in terms of significance.
But there is yet more to tell. Alongside the four adapted metal EOS machines in the AM facility in Finspang stands a fleet of other metal AM machines, also from EOS. Some smaller machines are used for product development, prototyping and functional testing of parts. The bigger ones are being used to produce one-off replacement parts and some serial production parts on a regular basis. Volumes are still in the three-four figures, nothing that could be considered high, and while we were not specifically told there would be high volumes anytime soon, I would be prepared to bet it’s only a matter on time.
The parts that are in production using AM are what you would expect—complex and often redesigned to improve functionality and consolidate the number of components. Jenna Nielsson is the manager of the design team developing new applications specifically for AM. We were not able to take pictures but there were lots of examples of current projects. That said, proportionally, it’s still a tiny number compared with the total number of parts in a complete turbine.
Thus, we return to the fact that always needs to be restated—AM is not a replacement technology, nor will it ever be. It’s a value adding process and often a cost-reducing one— over the lifetime of the product if not in cost per part. This issue raised its head more than once over the Siemens press conference. A few of the attendees did not seem able to differentiate here and persistently asked about the costs, despite patient and consistent explanations from Siemens personnel that the costs (hardware and per part) are secondary to the value AM can bring to its customers over the lifetime of the product—both in time and GBP/USD/EUR.
I came away from this trip truly inspired by what I had seen and heard. It made me really happy to have witnessed a company so full of passion and (well-deserved) pride opening up about its real—and realistic—relationship with AM, within a digital framework, rather than talking endlessly about an AM roadmap/strategy. Siemens is getting on with implementing their roadmap and periodically sharing the results.
Few other company visits have affected me so profoundly—a ‘perfect storm’ I won’t quickly forget.