In early October we connect with Inductive Automation’s Kevin McClusky to discuss during using artificial intelligence and machine learning to excel in the post-pandemic manufacturing world. (Access his on-demand Base Camp Digital webinar here.)
Today we chat with Kevin about specific techniques with these technologies, the feasibility of lights-out factories, and the correlation between piano-playing and smart manufacturing. Take a look…
Smart Industry: How is AI and ML changing amid this pandemic? Any industries adopting these approaches particularly aggressively?
Kevin: A lot of companies have started exploring how to do more with less, and how to take advantage of all the data that new technologies and architectures provide. AI and ML can be an important part of this exploration. One note of caution: it's important to define the goals of an exploration, rather than just trying to apply AI and ML to all problems. We'll talk more about that during the webinar. As far as industries, we've seen quite a few oil & gas companies using it for modeling, but it really applies to every industry.
Smart Industry: Define process tuning and explain how this is changing in this digital/data era.
Kevin: Every process is configured by an engineer with a variety of settings when first installed. During start-up, these settings may be changed to dial in the process. That is process tuning. Over time, equipment wears, and processes need to be re-tuned in order to achieve the original performance. One of the interesting developments in process tuning is the application of machine learning to determine better settings for optimizing processes. This can apply to systems that need to be re-tuned, but can also sometimes apply to new systems that have just been installed, if the algorithms are able to identify patterns better than a human could.
Smart Industry: What's an example of a commonly used metric to shift business strategy?
Kevin: Cost-per-unit is an obvious one for manufacturing companies. One nice shift recently is the ability for a lot of companies to get this nearly instantaneously, with data from payroll systems, utilities, operating overhead, and raw-materials prices taken into account. Sometimes the shift to getting this instantaneously, rather than daily or weekly, can help with insights as to why the metric might be going up or down. If a company has multiple facilities, this can also help focus CapEx or focus performance improvements for facilities by capturing best of breed SOPs from top-performing locations. For companies with a single facility, this can still be beneficial when comparing line-to-line or shift-to-shift.
Smart Industry: Is the lights-out model truly achievable for everyone? And what will this look like in the coming years?
Kevin: It depends on the industry. In some industries, this is already happening. In others, it's years away, if it ever is to happen. Systems that have continuous processes and those that already have heavy automation are the most likely candidates for achieving this goal. The shift to having remote accessibility for workers also has opened up the possibility to have staffed lights-out operations, with no one at a facility but with humans still monitoring and making operating decisions.
Smart Industry: You’re a pianist. Any correlation between tickling the ivories and working in this space?
Kevin: Great question! I play a bit of Jazz, and I find improvisation to apply equally to music and in our industry. I try to keep an open mind toward new technologies and creative solutions to problems, so when I hit a "wrong note" it just becomes part of the flow with some well-chosen following notes. Those next few notes can change the direction of a song or of an approach to a problem, and sometimes that's exactly what's needed.
