In August, I spent a week at the General Assembly of the International Academy of Production Engineering (commonly known by its French acronym CIRP) in Guimarães, Portugal. The General Assembly is a venue to listen, observe and discuss the advancements in all things manufacturing. Most topics are reports on basic research, but the authors are often leading experts in their fields and most have direct or indirect experience with industry.

Naturally, topics that garnered the most attention revolved around Industry 4.0, smart manufacturing, and the IIoT. But the question remains about what defines, enables or exemplifies all of those things. A committee keynote speech on the topic of process chains set the stage for better understanding a more connected manufacturing environment.

Challenge: How do we move beyond passive or discrete device data (e.g. monitoring) to accomplish a more accurate view of production reality? Is there a viable path that expands data visibility beyond just the equipment?

The Approach: Much research is being conducted on both the source and influence of manufacturing variations. I generalize those variations into two categories: temporal, where variations are observed over time (due to a progression of events or influences of other subjects), or spatial, where variations are observed within different locations of the same subject (an example would be gradients). One of the biggest challenges is in isolating temporal and spatial variations in order to better understand cause and effect. An approach focusing on cause and effect therefore requires extending visibility from the discrete technology devices to also incorporate other devices, processes and externalities.

How it Works: By applying closed-loop controls it’s possible to differentiate between temporal and spatial variations, which allows for focusing on both independently. The key of the differentiation is separating data from the noise. In other words, some things may not change with time and thus not have significant influence over the event as a function of time; in other cases such variations do occur.

“Process chains” describe how cause and effect impacts a subject over time while also indicating changes within a subject regardless of time but accounting for effects from other processes. It is a more manageable way to break down complex or at least complicated systems into elements for further analysis. By studying events as process chains and not as discrete (albeit sometimes dynamic) data points you may better replicate what actually happens in your manufacturing process. Digitization may naturally occur while incorporating the technologies required for process chain visibility. That leads to visibility technologies for process chains becoming exemplary of the “digital thread” and a tenet of things like Industry 4.0, smart manufacturing and the IIoT.

Findings & Applications: By thinking of continuous events instead of only discrete data points, you are able to consider multiple cause-and-effect process chains instead of discrete, static events. Process chains begin to raise the fidelity necessary to better describe what’s actually happening within a manufacturing process.

For more information about this article, please contact Tim Shinbara at or 703-827-5243.