Control Design's "Let the Machines Talk" (www.ControlDesign.com/talk) article forced me to ponder this notion. Given the very attractive price points of industrial-grade networking hardware, the range of selection, and the ease with which many things now connect—think wireless here—why aren't more machines "aware" of their own surroundings? The hardware to do it is available, and the majority of it works very well. From sensors to devices to controllers to safety components, the networking problems associated with exchanging data and sharing information are largely solved—that is, at least from a hardware perspective.
Isn't it better for upstream equipment to know that the downstream palletizer has a pusher jam, rather than just knowing that the out-feed conveyor is backed up? Armed with knowledge of the pusher jam, the other equipment can make better decisions on what to do or not to do. This decision could be as simple as each part of the line using all of the conveyor buffer space in order to continue producing. Or, based on data about units left to produce, and current raw material remaining, it might be best to immediately send a text message to maintenance personnel so they arrive prepared. Armed with good data, it might be the machine that tells the operator or service folks what to do next, rather than the other way around.
I think that's the Holy Grail: interconnecting machines and sharing data such that the entire production line is aware of the status of each of its dependents, only alerting personnel to exceptions requiring manual intervention. All other events are handled by sound local decisions based on awareness of the rest of the line. Note: The machine can be constantly aware; the operator needs to do research to have current information at his or her disposal.
Certainly, the reasons behind events that occur during a production run would be nice to see in OEE dashboards, or in key performance metrics (KPMs), and that information certainly would help with predictive maintenance. For example, knowing that the incidence of that pusher jam correlates tightly to a certain group of products—or a set of products running at a particular rate—could be something that's not intuitively obvious. It's hidden from casual observation, but perhaps is the single key to solve a nagging problem. A machine aware of its own surroundings has that information available. The analyses can happen on their own; no need to manually overlay data, parse reports or evaluate evidence.
So why is it that this intelligence isn't the norm?
[pullquote]A process implies a pre-determined set of steps to accomplish something. That "something" is defined, known and readily available. "Machine," in contrast, implies certain specificity. Often, we think of it from the perspective of the dedicated task it serves. It is, however, commonplace for a machine to exist as part of a larger process, while the controls remain targeted to the machine, rather than to production capability. A machine usually only has certain input expectations, rather than the smarts to deal with things that fall outside them.
It's that mindset—failing to focus on the production capability—that needs to change. Capability is what a machine provides, but that capability is only realized when inputs are available for use and outputs can be received. Knowledge of both is very important, but we usually think of them as discrete, rather than continuous. Process automation sees things as continuous, which is the way machines should, too.
Machines need to exchange data to do it. In today's world, a PC that is placed on the plant floor is almost always connected to the corporate intranet, thereby possessing the ability to get and give data. Many expensive production machines in those same areas are connected to each other with only a conveyor. Very few exchange live data. It simply doesn't make sense.
When we talk about getting rid of the lights, it forces us to view the machines and the roles they serve differently. It also tends to push us in a direction that leads toward smart, highly flexible, productive facilities. Remember, it's not about how fast we produce; it's about how effectively we produce.
