Cognitive Manufacturing Enabled by Digital Transformation

Overview

Digital transformation has resulted in an array of emerging technologies and scientific breakthroughs that profoundly change the way products are designed and manufactured. Production systems have steadily evolved from standalone automated processes to highly integrated work cells that fabricate parts, assemble these into products, and move them along the production line.

The control systems and level of automation associated with these complex production systems have also become increasingly complex. Today, we are seeing these automated production systems make the next evolutionary step into cognitive manufacturing based on machine learning, adaptive controls, and analytics that can predict changes in the process, prescribe how to improve the process, and teach the system to run autonomously and heal itself.

We have already seen how cognitive computing takes Big Data analytics to a new level. It allows systems to learn, form hypotheses, and make recommendations much faster than humans. The basic characteristics of a cognitive system are understanding, reasoning, learning, and responding.  Now, we see this approach being applied to manufacturing.

Cognitive machines can utilize large amounts of sensory data to form patterns, reason, and learn over time. Taking this capability to the factory floor will radically change the production process. A cognitive machine doesn’t get tired or take breaks, but works continuously to increase productivity and product quality while relentlessly “learning” how to further improve the production process.

Beyond Smart Manufacturing

Cognitive manufacturing, the next step in computer-enabled production system control, pushes us beyond just “smart” technologies. This is an area in which intelligence and reasoning is retained by the system, and provides the manufacturing system with capabilities for perception and judgment.  These enable autonomous operation of production systems based on embedded cognitive reasoning. Cognitive manufacturing systems will perceive changes in the production process and know how to respond to dynamic fluctuations by adapting the production to stay within target ranges of production cost and rate and, increasingly, sustainability areas such as energy use and carbon footprint. The embedded cognitive capability requires the development of cognitive reasoning engines, or distributed intelligence agents, that are then deployed throughout the production system.

Further, a cognitive manufacturing approach enhances areas like performance analytics, operational intelligence, and closed-loop PLM to help ensure as-built to as-designed production.  Together, these support continuous process improvements and help validate as-built to as-designed processes. The real payback of advanced analytics is going beyond predictive to prescriptive analytics, where we bring together Big Data, statistical sciences, rules-based logic, and machine learning to empirically discover and reveal the origins of complex problems and then evaluate decision-based options to resolve them.  

Intelligent Use of Data Key to Cognitive Manufacturing

According to the US Bureau of Labor Statistics, the manufacturing industries (both discrete and process) have the most stored data (well over 1,500 petabytes) of any industrial or business sector.  Unfortunately, much of this data goes unused, rather than being aggregated, analyzed, and converted into actionable information.  When connected by a “digital thread,” this information becomes a closed-loop mechanism that can support product development, manufacturing processes, and operations and maintenance in the field.

Often, the only effective way to analyze the tremendous volumes of data accumulated by a typical product manufacturer is through large-scale pattern matching technology enabled by machine learning (ML) algorithms.  ML, an element of cognitive manufacturing, excels at processing very large amounts of data and every combination of variables.  Eventually, pattern matches are made and predictable outcomes form that, in the case of product process records, can result in a set of rules and best practices that lead to overall process improvement and better product designs.  The process is straightforward and relatively simple in theory, but can involve some rather complex algorithms and requires access to a lot of data. This is the essence of cognitive manufacturing.

Cognitive Manufacturing Offers Significant Benefits

Cognitive manufacturing transforms the way humans and machines interact. This goes far beyond the traditional human-machine interface (HMI) software generally associated with automation and control systems. While legacy HMIs can provide the operators and factory workers with a real-time view of the current state of machines and production equipment; with cognition, technicians have years of performance, quality assurance records, and repair history at their disposal.

Moreover, cognitive machines can predict breakdowns and errors before they occur and even prescribe preventive fixes. And with the help of virtual and augmented reality technologies, manufacturing engineers can recreate these predicted errors in safe digital environments and simulate fixing them remotely for factories and production lines located globally.

Cognitive machines will improve the overall quality assurance process significantly. Many quality systems are good at detecting defects in the product, but cannot detect the real issues with the process. The basic principle of quality assurance is to improve the overall process incrementally by finding the root causes and deficiencies in the production process, and incorporate the improvements for quality assurance. A cognitive approach using predictive and prescriptive analytics based on machine learning fulfills the fundamental requirements for continuous process improvement and best practices for quality assurance.

Connected and intelligent factory ecosystems will help enable cognitive manufacturing. The next generation of smart sensors and edge devices that not only access data, but aggregate and analyze data at the machine and equipment levels, will be critical components of the overall cognitive manufacturing process.

Recommendations

As manufacturers make the digital transformation and implement advanced analytics and smart connected factories, machines, and production systems; cognitive manufacturing technologies and approaches will both help define and realize this transformation.

In the not-so-distant future, cognitive machines and production systems will not only analyze data and inform the factory floor operators and technicians when there are exceptions and anomalies, but autonomously act on the findings and correct the problems.

To learn what your peer organizations are doing in this exciting area and participate in the discussion, ARC invites you to join us at our upcoming 25th Annual ARC Industry Forum Goes Virtual, Accelerating Digital Transformation in a Post-COVID World, February 8-11, 2021 - Online

If you would like to buy this report or obtain information about how to become a client, please   Contact Us

Keywords: Cognitive Machines, Predictive/Prescriptive Analytics, Smart Connected Factory Ecosystem, Operational Intelligence, Continuous Process Improvement, ARC Advisory Group.