Multiplying problems require multiplying solutions—or at least one that’s super-capable.
Experts from McKinsey & Co. set up all the challenges arrayed against industrial and life sciences manufacturers at the opening of the Process Solutions User Group (PSUG) Summit, held this week in conjunction with Automation Fair 2023 this week in Boston. Then, experts from Rockwell Automation showed how the company’s PlantPAx distributed control system (DCS) is adding even more muscle and speed to knock these challenges down.
Host of hassles? Digitalization and AI
McKinsey & Co. reported the key difficulties facing the process and other industries include:
• 25% increase in raw material costs over the past three years;
• Labor shortage highlighted by a doubling of unfilled manufacturing jobs during 2010-23;
• Increasing environmental sustainability goals that only 29% of manufacturing believe they’re on track to meet; and
• Plateauing performance improvements characterized by labor productivity growth that decreased by 38% over the past 15 years.
“Based on what McKinsey has learned, process industry leaders are concentrating on five priorities to overcome these challenges: continued focus on quality compliance and excellence; at-scale technology adoption (Industry 4.0); reimagining supply chains to build agility and resilience; focusing on environment, sustainability and governance (ESG) in the medium term; and succeeding in obtaining talent and capabilities,” said Vivek Arora, partner and industry leader 4.0 for North American life sciences at McKinsey & Co. “And, because Industry 4.0 means digitalization, generative AI will open up even more opportunities and potential gains.”
For instance, Arora reported that cumulative productivity gains attributable to digitalization have included:
• 15% increase by adopting tools like supervisory control and data acquisition (SCADA), laboratory information management systems (LIMS) and enterprise resource planning (ERP) in the 2000s;
• 20% increase by adding digitally enabled performance management and advanced analytics in the 2010s;
• 30% increase by adding touchless solutions, such as integrating operational technology (OT) and information technology (IT) and selectively converting batch operations to continuous manufacturing; and
• 50% increase by implementing interactive AI to automate machine learning (ML), produce prescribed insights with generative AI, and allow on-the-go decision-making.
“Generative AI is really the next frontier because it can extract insights in seconds—if it’s attached to the right data sources,” said Andy Luse, partner and industry leader 4.0 for North American industrials at McKinsey & Co. “It can also generate content tailored to specific contexts and interact with users through human-like conversations.”
Luse added that digitalization and AI can obviously help with many of the essential process industry functions that he and Arora outlined, including supplier connectivity, workforce management, plant-floor production, quality control and facility support. However, he reported that McKinsey also discovered that companies who maximize the value they get from digitalization and AI are those that apply it end-to-end in their operations to integrate workflows and other overall tasks. These breakthrough gains include 20-40% increases in the capacity of formerly bottlenecked equipment, 20-35% reductions in conversion cost, 30-50% reductions in deviations, and even 10-20% decreases in greenhouse gas (GHG) emissions.
Scaling overcomes obstacles
Despite the potential gains, Arora reported that only 11% of companies have deployed Industry 4.0 digitalization in their production networks. He stated this is due to several headwinds, including:
• Large and complex manufacturing footprints of varying digital maturity—often with hundreds of sites and suppliers on a network;
• Thousands of people are needed to participate in digital transformation, so more training and capability-building are needed;
• Many use cases aren’t standardized and their asset performance can’t be assessed yet; and
• Technology is evolving faster than many users can keep up.
“We believe there are six strategies that successful scalers use to digitalize their processes and organizations, and all six must be employed together,” says Arora. “They include having a strategic roadmap and project pipeline to drive value; attracting and retaining expert talent to carry out projects; using an iterative model to support teams and provide agile delivery; possessing enough plant-floor technology like sensors to support production; maintaining complete and accessible datasets; and adopting and running model transitions across sites to capture value.”
For instance, Arora reported that one life sciences manufacturer recently reimagined its plant-floor operations, applied a holistic set of Industry 4.0 technologies, including digital automation, digital production management (DPM), AI and generative AI. It focused on asset and labor productivity, capital expenditure (CapEx) avoidance, and reducing the cost of poor quality. The manufacturer deployed digital and advanced analytics (AA) at more than 10 sites, increased its overall equipment effectiveness (OEE) by 15%, reduced its cost of poor quality (COPQ) by more than 30%, and benefitted more than 200 plant-floor staffers. In fact, DPM was added to 3.6 million datapoints, which enabled daily data extractions, and allowed real-time production transparency.
“This pointed out high variability in process parameters, shifted bottlenecks, identified 15 alarms driving losses, allowed root-cause analyses, and improved OEE by 10%,” said Arora.
Likewise, Luse reported that a specialty chemical company started digitalizing three years ago to transform its networks in 15-20 of its more than 100 plants. Through employee training in digital methodologies and deployment of multiple digital and AI solutions at scale, the company increased throughput by 15-20%, trained 120 people as digital practitioners, and created AI solutions that could improve the run rates of more than $100 million in production.
“The keys to scaling are standardized solutions, modularization (including code for reuse), collaborating early with IT, and working with ecosystem partners up and down the technology stack. Because no one can do these projects by themselves,” added Luse.
Arora added that, “Scaling champions must master three must-have skills: building a clear vision and strategy, investing in people, and setting up the right governance and ways of working.”
Scheduled for release in mid-2024, Rockwell Automation’s PlantPAx 5.3 will feature enhanced scalability, scope and security, said Arvind Rao, vice president of industrial solutions, Rockwell Automation. It will include a firewall by Fortinet, industrial demilitarized zone (iDMZ), real-time threat detection from Claroty, and endpoint protection via OPC UA communication protocol and some added policy-management software.
“We just launched FLEXHA 5000 Universal I/O, which can work with PlantPAx via FactoryTalk Optix edge connectivity to enable closed-loop prescriptive analytics,” said Rao. “We also launched the FactoryTalk DataMosaix data platform in July for integrating information from Rockwell Automation’s devices and also from components from other suppliers. This gives users multiple improved opportunities for designing, simulating, optimizing and maintaining their process operation from end-to-end and at every layer up to the cloud.”
Beyond open data integration, Rao reported that Rockwell Automation and PlantPAx are enabling AI at the edge-computing level with the recent launch of FactoryTalk Analytics LogixAI, which can provide setpoint or closed-loop recommendations for process optimization. Likewise, asset optimization is further aided by out-of-the box AI for anomaly detection.
“More data is generated now than ever, but more is also getting stuck in silos. Plus, 70% of the effort in building analytics and other software applications is the data engineering,” explained Rao. “We’re aiming to have PlantPAx and FactoryTalk DataMosaix take away that burden, so users can get away from single-point solutions, and build their applications quickly and at scale.”