Artificial Intelligence in a sustainable Digital Enterprise

The course of history has been shaped by innovative thinkers who could conceive of novel approaches to the challenges they faced. Indeed, the stone age did not end because we ran out of stones but because of the development of new and more effective solutions to the problems of the time.

The future is being shaped in much the same way. Multi-dimensional trends that stretch across business, society, and technology are driving new challenges for modern companies. Overcoming these challenges will require companies to innovate and adopt new methods of managing their businesses. Specifically, companies must transform into sustainable Digital Enterprises to manage the product and production lifecycles in both the real and digital worlds. The key feature of a Digital Enterprise is its ability to unite the real and digital worlds, gathering data from real world operations and turning it into insights that drive improvements in the real world. Bringing these two worlds together enables the flow of data between all stakeholders involved in the product, production, and service lifecycle. This free flow of data supercharges the ability of a Digital Enterprise to adapt to market demands, innovate faster, and enhance quality, all while reducing carbon emissions and resource use.

The digital transformation journey of Siemens Electronics Factory Erlangen

Siemens’ state-of-the-art electronics factory in Erlangen, Germany, faced the same challenges that many manufacturers are attempting to overcome. The factory needed to find ways of producing products while prioritizing decarbonization, speed, quality, and cost-effectiveness. The Siemens Electronics Factory Erlangen, which produces SINAMICS frequency converters and SINUMERIK CNC controllers for Siemens and its customers, also had to navigate the integration of smart, digital solutions into legacy infrastructure and systems within the factory.?

The factory recognized that these challenges could not be overcome without embracing digitalization to continue an ongoing transformation into a sustainable Digital Enterprise that began decades earlier. Digital solutions for engineering, shop floor management, product lifecycle management, and more provide the foundation. With these in place, more advanced technologies and functions can be built to further enhance the operations at Erlangen.

Industrial artificial intelligence (AI) is one of the advanced technologies in use at the site. Where a consumer-focused AI may help generate text for a wedding speech or images for your social media profile, the industrial AI systems deployed in the factory control machinery, manage operations on the shop floor, and analyze data to indicate performance and guide decision-making. These industrial AI systems must be robust, reliable, and proven safe even when working in close collaboration with human workers.

Let’s explore how the Siemens Electronics Factory at Erlangen is making use of industrial-grade AI to support greater decarbonization, quality, and time-to-market.

Industrial AI: Efficiency, speed, quality

Before processes, machines, and systems can be improved, they must be understood through the collection and analysis of data. A Digital Enterprise generates immense amounts of data during everyday operations. In a factory environment this may include information on energy consumption of various systems and the entire plant, throughput data, real-time operational data coming from connected machines, and more. Collecting and understanding this data is critical to the management of a modern digital factory, but the sheer amount of data makes its aggregation and analysis a challenging task.

Using data to reduce energy consumption and downtime

Fortunately, the connected data flows of a Digital Enterprise present a golden opportunity for the application of AI to accelerate the analysis of these immense datasets. This will drive optimizations to various aspects of the factory both on the shop floor and in other systems much faster than before. The Siemens Electronics Factory Erlangen, for example, has leveraged data from throughout the factory to implement intelligent energy efficiency measures, reducing its energy consumption by 25%, and its net carbon footprint by 50%. Furthermore, targeted improvements to production efficiency have helped reduce the energy used to produce each product by 50%.

AI also enables true predictive maintenance schemes to ensure that machine downtime never comes as a surprise to factory operators. Machine and maintenance data are analyzed and compared to past cases to identify patterns and potential solutions. The factory makes use of predictive maintenance in a milling process that occurs as part of printed circuit board production. The milling process produces a fine dust that accumulates on the milling spindles and can hinder the rotation of the spindles or, with enough buildup, cause unplanned downtime. To prevent such costly delays, the predictive maintenance solution monitors spindle current and speed to watch for anomalies and even predict future critical states.

Intelligent processes are better processes

AI can be transformative for individual processes and machines as well, even enabling closer collaboration between humans and robotics on the shop floor in an efficient and safe manner. The Siemens Electronics Factory Erlangen employs AI and computer vision to enable robotic arms to pick and place parts with the same flexibility and dexterity as a human operator. Traditional robotic arms have no ability to distinguish between different parts, requiring parts to be pre-sorted and organized. The incorporation of AI into the control systems for robots enables them to identify and grab various parts out of an unsorted box and place them precisely where they belong.

By making the robot smarter, these types of tedious pick-and-place operations can be accomplished completely automatically by cost-effective robots. Of course, before these intelligent robotic arms can be unleashed, they must be trained. As a Digital Enterprise, the factory can employ physics-based simulations and the Digital Twin to virtually train the algorithms on part recognition, picking, and placement. The synthetic training data is generated and labeled automatically, increasing the speed and reducing the effort required to train the robotic arms.

Using AI to enhance precision and efficiency

More sensitive processes can also be automated by giving the robotics manual dexterity like that of a human hand. Integrating force and torque sensors into the end effector of an AI-controlled robotic arm enables it to precisely sense and adjust the force it is using to manipulate an object. This is crucial for processes involving delicate and small parts, such as components of a printed circuit board (PCB). In fact, the robotic arms handling such delicate tasks may require the same “Fingerspitzengefühl” as that of a surgeon stitching up a wound!

At the Siemens Electronics Factory Erlangen, the production of PCBs involves the fitting of wired electronic components through tiny holes in the substrate, called through hole technology (THT). THT involves very sensitive and delicate parts being plugged into very small holes in the PCB, often mere tenths of a millimeter in diameter. AI enables robotics to handle the components gently, ensuring they are placed accurately and secured without incurring damage. All in all, the automation of such a delicate task increases the quality of the process and frees human workers from the tedium and poor ergonomics associated with such a task.

In addition to the immediate benefits of AI, such as increased production quality and reduced cost, the addition of AI into the shop floor environment has also contributed to the sustainability of the factory. The increased precision and accuracy of the automated THT fitting process reduces scrap and thus wasted material and energy, making the factory more efficient overall. Since the intelligent robotics no longer require pre-sorted parts, the plastic inlays that were once required to organize sorted parts have become obsolete. The result is the elimination of thousands of plastic parts that ultimately end up as waste.

The future for sustainable Digital Enterprises

A Digital Enterprise can leverage AI and the vast quantities of data generated every day to identify and act on opportunities for decarbonization, reducing resource usage, recycling, and more, across both internal processes and global supply chains. The decisions made in product design, for instance, account for 80% of the environmental impact of a product that makes it to the real world – in other words, waste is no more than a design flaw. A sustainable Digital Enterprise can use a combination of the comprehensive Digital Twin, data, and AI to understand the relative sustainability costs of various design decisions, optimizing for efficient performance, material usage, and recyclability.

In production, AI may help optimize production schedules to match demand, identify opportunities for energy efficiency, and like in Erlangen, vastly improve production quality and reduce scrap and material waste. And, perhaps most importantly, AI can become a powerful tool in the management of elaborate global supply chains, helping companies select suppliers and build logistics systems based on cost, quality, and sustainability. As we continue to evolve and expand the capabilities of a sustainable Digital Enterprise, facilities like the Siemens Electronics Factory Erlangen offer a crucial proving ground for technologies and solutions that can help customers transform to overcome the challenges of today and tomorrow.

Discover more about the transformation of:

Siemens Electronics Factory Erlangen - Siemens Global

?