• By William Pollock
  • Channel Chat

By William Pollock

The Internet of Things (IoT) is revolutionizing all areas of our lives. In manufacturing, Industry 4.0 is as transformative as the first Industrial Revolution 300 years ago. Advancements in manufacturing analytics are coming fast and furious. Not all manufacturing analytics are totally new, however.

I recently had a conversation with Bob Hopkins, senior system architect at Eastman Kodak Company. We talked about Process Monitor, a software system developed at Kodak and used effectively over the past quarter century to improve the quality and yield of the company's highly sophisticated chemical processes.

A quarter century ago, at the height of the silver halide dominance in imaging, systems engineers at Kodak developed a process monitoring analytics system that was decades ahead of its time. This toolset was one way that Kodak manufactured high-quality and consistently uniform chemicals, dyes, and emulsions in plants operating in five different continents.

Hopkins said that the original mission was to achieve Six Sigma performance in manufacturing, but the end result has provided benefits far greater than originally anticipated. Hopkins explained that the company has used this system across its entire global film manufacturing supply chain and across all levels from research and development to large-scale manufacturing. Using the analytical capability of its proprietary product, Kodak found process malfunctions before they occurred, predicted product response performance, and summarized massive and complex data to an actionable data subset-and as a result it could reduce the manufacture of off-specification product.

This monitoring approach is rooted in some simple ideas, but using these simple ideas brings sophisticated results. Hopkins gave a few examples. The system:

  • increased Kodak's understanding of its existing processes, so employees could take proper steps to improve reliability and performance
  • improved key performance indicators, such as mean time between failure and first pass yield metrics, of established processes after years of stagnant performance on these measures
  • gave clues to allow the company to solve production issues that had (in many cases) been unresolved for years
  • transformed maintenance practices from preventive to predictive for pumps, valves, and other components
  • provided the answers to after-the-fact diagnosis of product failures
  • answered questions about why certain product recipes did not scale
  • detected faults in time to prevent further product losses downstream in the supply chain

Technology available to record and store data has never been easier to use. Today we have multiple ways to store manufacturing data either in our own facilities or in the cloud. But the age-old question is how to best digest this recorded data and make sense of it in the context of manufacturing operations?

Kodak Process Monitor recognizes that a complete picture of manufacturing encompasses "data," which does not all have a common basis. Time-stamped sensor data is collected from machines at different rates; on-line product testing results become available at different times, usually after the fact; and production order and maintenance records have their own unique system of keys. For flexible manufacturing, the process itself is not fixed, but changes every time the recipe changes. This data cannot be combined without much manual synchronization, filtering, and effort. It is a level of effort most engineers will go through only during a crisis when their process is producing unacceptable product for unknown reasons!

The Process Monitor automated process takes these disparate inputs, synchronizes them, and outputs a much smaller set of meaningful parameters produced at the right time for timely actions. A very simple example from one of its sensitizing processes is a set of parameters computed routinely that indicates the wearing of metering pump gears. An early warning signal is sent in plenty of time for maintenance to swap out parts when the process is down for product changeover. This eliminates a "loss of flow" failure due to gear wear, a fault that could cause lost product and waste during a production run.

Kodak Process Monitor can improve almost any manufacturing process. Manufacturers in chemical, agrochemicals, petrochemicals, and cosmetics all share the same manufacturing challenges as Kodak. They are concerned about yield optimization, energy consumption, product quality, and reduction of run-to-run variability across different process plants. Dealing with multiple plant operations is challenging, because of differences in systems, processes, business definitions, and equipment. Collecting data from the automation control system (typically operating in a programmable logic controller), the manufacturing execution system, laboratory information systems, and data historians provides enough information to carry out the necessary analytics. Process Monitor, originally developed for use only by Kodak, is now available for any manufacturer to use.

Reader Feedback

We want to hear from you! Please send us your comments and questions about this topic to InTechmagazine@isa.org.

Like This Article?

Subscribe Now!

About The Authors

William Pollock is president and CEO of Optimation Technology, Inc. Before founding Optimation, Pollock was an associate professor of systems engineering at the State University of New York. He also spent 10 years as a control systems engineer at Eastman Kodak. He has a BS and an ME from Rensselaer Polytechnic Institute. Optimation (http://optimation.us) is a member of the Control System Integrators Association (CSIA) (www.controlsys.org).