Digitalized help on the way for manufacturing challenges

There are always concerns that automation is going to take away jobs. However, in light of today’s rapidly shifting technologies, supply chain disruptions—and millions of already unfilled jobs—it’s more than likely there will be plenty work for everyone, even if job descriptions change and retraining is necessary.

“Rockwell Automation wants to expand human possibilities. And when we talk about the future and things like robots on the plant floor, we’re not talking about removing people from the equation,” explained Cyril Perducat, chief technology officer, Rockwell Automation. “We want to augment existing people and their capabilities. There are a lot of buzzwords today, including artificial intelligence (AI), but they aren’t just about technology. They’re about creating the right experiences for people. There’s no way to invest in the future without making tools usable for people.”

Perducat presented during the closing keynote address at Automation Fair 2023, held last week at the Boston Convention and Exposition Center.

Challenges and strategies

Perducat reported it’s important to start with the individual combination of challenges that each industry and company is facing. “Manufacturers want stable production, but many demands are quickly evolving,” added Perducat. “Many cars are built as almost unique models, and many pharmaceuticals are produced using increasingly individual specifications.”

Perducat added the second big challenge is supply-chain volatility and shortages of raw and finished materials, such as semiconductors, which ramped up during the COVID-19 pandemic, and have continued more recently due to the ongoing war in Ukraine and other geopolitical turmoil. These upheavals and other uncertainties have been accompanied by increasing cybersecurity risks. “Cyber probes, intrusions and attacks typically come from malicious actors targeting manufacturing sites and infrastructure,” explained Perducat. “These challenges are why users need to stay agile, optimized, and maintain high productivity.”

To help users develop useful responses to these multiple challenges, Perducat outlined present and future conditions from four angles: production that is both agile and optimized as well as resilient and sustainable.

Agile optimization

First, flexible machines that enable fast changeovers will give way to reconfigurable manufacturing systems that also enable simplified updates. Plus, digital twins for machine and production line design will evolve into digital product lifecycle integration that enables constant evolution.

“This is about making data available and networks plug-and- play, so users can get increasingly granular views of what’s happening in their processes,” said Perducat. “This can also enable development of digital twins for specific processes or pieces of equipment.”

For example, if a user can understand the six or seven primary variables that drive production as well as the primary elements that characterize its physical and digital reality in real time, then production can be continually revised and optimized. “This intersection between digital twins of products and their production lines is important,” added Perducat. “This is because optimization now also provides a foundation for re-optimizing processes later or customizing more ‘batches of one’ product.”

Resources and sustainability

Second, materials considerations in the present will become product lifecycle designs in the future. Renewable energy sources become fully integrated along with carbon capture and other viable energy models. And, machine learning (ML) and analytics will be modernized for closed-loop optimization.

“All consumable raw materials and other resources can be optimized in the same ways as the processes that use them,” explained Perducat. “Consequently, these considerations will also become part of product designs. This means they’ll need more data for closed-loop systems, so product designs can help reduce consumption and contribute to sustainability, too. Likewise, we can look at scheduling energy intensive production tasks when solar is readily available.”

Perducat added that Rockwell Automation wants to use AI for closed-loop optimization of complete production lines and facilities. “Even though many of today’s tools aren’t fast enough, users can still look to make operational adaptations,” said Perducat. “AI allows users to not focus so much on the data, and instead concentrate on opportunities for optimization. This is because its algorithms are better than traditional methods at predicting variations and optimizing for better outcomes. AI can work at all levels, so we’re developing AI tools and building blocks. The programmable logic controllers of the future will all have native AI capabilities to simplify operations and data analytics in devices.”

Resilience gets a revamp

Third, supply chains will be strengthened by being made more adaptive and part of integrated ecosystems, which will be able to automate and integrate all material movement. Meanwhile, present cybersecurity directives to identify, protect, detect, respond and recover will be joined by zero-trust architectures and enhanced threat detection.

“The plant-floor must adopt the same cybersecurity practices for its people and equipment as information technology (IT), including zero-trust,” added Perducat. “Rockwell Automation is doing this and looking to do more.”

People-pleasing user experiences

Fourth, to further simplify things on the human front, Perducat reported that predictive maintenance systems will add natural-language interrogation to help users. Similarly, assistive features in design tools will offer AI-enabled code sharing, collaborative generation and verification. Plus, autonomous mobile robots (AMR) and intelligent conveyance will be advanced by robots that can perform all types of material movement.

“Why look at multiple screens and trends if you can just ask the system what changes have happened as a production system shifted from point A to point B? Experienced users know where to look, but the large-language models (LLM) underlying ChatGPT mean rookies no longer need to learn as much about what data to capture and use,” explained Perducat.

“ChatGPT and LLM can do these tasks and enable optimization in much less time. This gives users the copilot they need to drive production in the right direction. It also lets everyone share the code and knowledge of their best developer and verify that they’re following best practices and company standards, which is a huge aid to user experience. This is also how AI and generative AI can enable robots to act as moving sensors, improve movement of materials and systems, and further augment people.”