Reading time ( words)
“Smart” is not simply an “on” or “off” state. Just like people, some solutions are “smarter” than others; as we see from the various ways of measuring intelligence within humans, there are many kinds of “smarts.” When looking to invest in a smart manufacturing strategy, the playing field is more complex than it may appear. We should first understand and define what “smart” means, to what extent it exists, and the other requirements and dependencies that are needed to get the best value from investment.
Perhaps we should start with the IQ test, a measurement of intelligence attributed to people. Could we not have a scale for the intelligence of smart processes? As an adjective, “smart” sets expectations, but there is a huge gamut of actual intelligence or cleverness that can pass for smart. Perhaps we can create an AIQ (artificial intelligence quotient) scale to help us.
At the bottom of the AIQ scale, I would place the smartphone, which does nothing unless you press a button or configure an action to make it do exactly what it was programmed to do. The functions shortcut many otherwise manual actions. Smart functions are normally associated with software, though that is not to say that hardware cannot also be smart. A modern domestic sewing machine, for example, enables a novice operator with little skill or experience to create all manner of amazing stitching patterns. It is “smart” in that it augments human capabilities.
A mechanical manufacturing machine-based process—anything from a simple SMT placer to an assembly robot arm—is likely to be at least as smart as a smartphone. They basically do what they are told to do as a replacement of otherwise manual operations. Results of Six Sigma experiments tell us that there will always be variation in mechanical movements, resulting in slightly different results each time. Sensors are therefore built into machines and linked with the control logic, acting as feedback to make sure that the operational movement meets expectation, avoiding excessive or incorrect actions.
Expanding on that principle of direct feedback in a smart way is machine learning (ML). For example, an inspection machine will modify its assessment of passes and fails as a result of changed settings determined from analysis of prior judgements. This expands out to become “closed-loop” feedback where two or more different machines are involved. For example, deviations in X, Y, and rotation position of placed SMT components are measured by an inspection machine. Then, by an analysis of patterns and trends in the data, parameters are changed to reduce the deviation in placement. Subtle live alterations to the placement process can be done automatically, or the alarm is raised to call for an operator to replace a worn nozzle, for example. Instances of defects are then avoided by not allowing any deviations to go beyond control limits.
These smart technologies, available today, are used to correct and refine actions that are performed, enhancing the effect of the original sensors. It is interesting to note that practical implementation of these steps has been enabled by expanded communication of data between machines that are likely to have come from different vendors. Technologies such as IPC-CFX (Connected Factory Exchange) have revolutionized the ability to share and utilize actionable data in a singularly defined format and meaning, without the need for middleware and IP exchange. This has been smart. We are certainly now higher up the AIQ scale, as demonstrated in this simple example, by the order of magnitude of defects found at test. But the losses are not yet reduced to zero. Our degree of “smart” is not yet perfect.
To read this entire article, which appeared in the March 2021 issue of SMT007 Magazine, click here.