The Lean Digital Factory (LDF) program stands for me for excellence, innovation, collaboration and empowerment. Within the LDF approach our DI factories around the globe pursue jointly a common purpose: to shape the future of industrial automation and to make the 4th Industrial Revolution come true. Experts all over the world collaborate across Business Units and functions in agile networks and make things happen. Within this network our lighthouse factories show the potential and value of our Siemens automation and digitalization portfolio, which creates value to our customers and supports the transformation to lean digital manufacturing. I am proud to be an integral part of this great program with our MC factories.Achim Peltz, CEO of the Business Unit DI MC
The Workstream “Big Data and Analytics”
The third workstream deals with the objective of a comprehensive Manufacturing Data Ecosystem, using artificial intelligence to increase quality and efficiency in production while laying the foundation for scalability.
Via applications running on edge and cloud environment, value creation is done by prescriptive decisions and self-learning systems enabled by artificial intelligence (AI). Advanced analytics, machine learning, and deep learning optimize product and production system efficiency and quality. Due to the amount of data, the model training is mostly done in the manufacturing data platform. Solutions can be scaled within the ecosystem to other factories.
All kind of data is seen as analyzable, as a complete Manufacturing Data Ecosystem (MDE) is in place. The MDE consists out of an “Industrial Edge” platform, a state-of-the-art manufacturing data lake concept by MindSphere (MDP) and data source connectivity to feed data into a raw data store and an open analytic platform.
The ecosystem enables data citizens and analysts to dashboard information or to get deep process insights to improve, predict or stabilize manufacturing processes, efficiently without manual data preparation. For this, data and algorithms can be stored, provided, transformed, and trained on cloud and/or edge environment, enabling closed loop manufacturing, machine learning and comprehensive cross-factory reporting and decision-making without compromising IT security.
A total and comprehensive traceability and tracking concept is enabled for quality, warranty, engineering, and sustainability requirements, also due to the embedded data persistence in the Manufacturing Data Platform. All relevant information correlated in all dimensions is available to allow holistic tracing along the whole product lifecycle.
Let me now introduce you an exemplary edge & AI use case in one of our electronic factories. Very often you hear something about future IT/OT Integration, Predictive Maintenance, Artificial intelligence, and the combination out of it. For us, at Siemens this is not a vision anymore, we are already rolling out such technologies in our manufacturing sites and are able to help our customers doing the same increasing the machine availability and achieve even higher productivity. The tools are already there, one of them is edge computing with which we bring IT/OT together and enable today’s automation customers to leverage the newest technology like AI in automation. In our full automated assembly lines for our PLCs of our well-known SIMATIC products we need PCB cutting machines, who are cutting the printed circuit boards accordingly to our needs.
The challenge here is, the spindle of such cutting machine gets stuck time after time due to aggressive and corrosive dust, mostly in a period of one to six months. If this happens, the machines stand still for up to 1-2 shifts because we need to change the drive system of the whole spindle. In the maintenance process we also need to fetch new adjustments parameters for the drive system by the machine producer, and if the machine producer only is available at the normal shift system, we cannot get the information at night shift, for example. Such unplanned downtime significantly reduce machine up times and leads to productivity losses at workers!
So, what we did was developing a smart algorithm on the edge computing to be able to predict the machine breakdown up to 36h before it occurs, giving sufficient time to make an unplanned downtime a planned and prepared maintenance event. To train this artificial intelligence algorithm we first collected data on the Edge Layer and transferred it altogether in the cloud to use the available computing power in Siemens MindSphere to train the artificial intelligence algorithm. Once the AI algorithm was trained, we put him in an Edge Application and downloaded it onto the Edge device near to the machine.
The Edge Device then collects the machine data like spindle speed and drive current demand, analyzes it with machine learning and detects anomalies for the spindle. Based on the detected anomaly the algorithm then predicts the failure rate of the spindle, reducing this kind of availability loss by 100% per machine. After it has been equipped to the total of 18 spindles in the manufacturing site, it results in total savings of at least 120k€ p.a.
Similar applications running on edge environment can be found in various application fields. Due to the amount of data, the model training is mostly done in the manufacturing data platform. Solutions can be scaled within the ecosystem to other factories.
In general, the five main topics for advanced data analytics describes best the trends we see:
- How to improve quality level while reducing test effort?
- How to achieve one piece flow while maximizing utilization?
- How to increase up time while minimizing maintenance costs?
- How to stabilize process while lowering control effort?
- How to improve collaboration while reducing communication efforts?
The Workstream “Robotics”
Now coming to our next workstream “Robotics”, which has the focus of cooperation of digitally guided workers and interlinked autonomous production systems to enable efficient, flexible, and easy to scale-up manufacturing.
Even for low volume products and formally typical manual work, the automation level is increasing by modular autonomous production systems and robotics. Such systems are consisting of reusable elements, are easy to scale-up and manufacture products with high efficiency, quality, and speed. Respective machine programs are automatically generated for example with a robotic agnostic language, enabling a high flexibility and significantly reducing manual programming efforts. Higher operator efficiency and flexibility are facilitated by augmented reality applications, which provide workers with customized product and process information, depending on their needs and competence profiles. At the same time process quality is raised. Due to permanent exchange of information and status quo between production system agents, a flexible and efficient material and production flow is autonomously organized. But what does this mean in practice?
A common challenge in one of our electronic factories is empty box handling, which normally is very labor intense. Even if we have a full automated intralogistics system, the boxes, who are coming back from our warehouse must be fed back into our logistic system. In former times, the process was manual and very irregular which made the planning of the workers very difficult. So, the use case is about box handling with a low-cost automation solution consisting out of a lightweight robot and an autonomous guided vehicle. The light-weight robot is feeding returned pendulum boxes from pallets back to the fully automated material distribution system of the plant. Therefore, the robot recognizes the missing boxes at the empty pallet and communicates directly via “Profinet” with an autonomic guided vehicle (AGV), requesting a replenishment.
The AGV travels via an elevator to the robotic station and on its way, it scans the palette store which it passes, to capture possible storage places with full palettes. Once the palette at the robotic cell is replaced, a confirmation message is sent back to the AGV, again via “Profinet”. For security reason a radar surveillance is installed at the robotic cell, to capture workers or any other object and in case of any movement, the operation will be stopped automatically in the cell.
For LDF, a digital factory is building on employees with a digital mindset and the right competencies, acquired by live-long learning. Co-working and agile collaboration supported by the required working infrastructure is leading to flexibility and efficiency.
The cross-functional collaboration in a digital factory is based on trust, reliability and openly shared information in communities and networks. In addition, the workers are facilitated by digital office and shop floor management and encouraged to experiment by using agile methods to learn fast.
Our employees in our factories have a growth mindset for dynamic self-organized learning, to secure own employability. Transparent, future oriented competence overviews give orientation for live-long learning. Skill transparency is a conception of oneself, beneficial for all people, organizations, and processes.
Having the courage to experiment and growing on the results, employees drive digitalization and eliminate waste in processes and energy consumption in own environment, using for example low-coding tools especially for process automation. In the recent build Digitalization Center “The Impulse” we also have now an Open Space Lab exactly for this purpose.
Therefore, at LDF, we analyzed in a structured and systematic way all our job profiles and tried to identify together with our employees the required competencies and qualification. For all our employees on the different levels, we are offering learning programs on the shop floor with short videos, like learning nuggets, up to profound and professional trainings to self-learning and online classes on a learning platform called „My learning world„ at a desktop.
Driven by the purpose to shape the vision of our future factories, a worldwide agile network of >200 experts (MF & IT) collaborate, based on trust, reliability, and openly shared information. This team is encouraged by the management to experiment & learn fast. While using our products, LDF proves their benefits to customers, and collaborates with R&D to take them to the next level, if new functionalities are needed.
This innovation power is honored by winning several industrial awards, supporting to position Siemens as digital thought leader in the market (I4.0, Diamond Star, World Economic Digital Lighthouse).
With future oriented competence overviews & trainings, LDF provides orientation for live-long learning to SAG employees via an established LDF learning channel.
LDF is the blueprint for the Siemens-wide FD program, and actively participates in it. The FD program mission is to push digitalization, to scale digital solutions across Siemens, and finally generate additional customer business.
What means “Lean Digital Factory” to us? In one word, it means to us “Together”, we work and create the future of digitalization “Together” with all our people and the flamingos are for us the perfect mascot.
Why, because Flamingos always start together, and the pull of the ones who are starting in front of you helps the birds next in line. Like we do in LDF.