Is the goal of smart manufacturing to integrate with machines and sensors to get data? Or is the goal to solve problems?
I recently attended a smart manufacturing workshop in Milpitas, CA hosted by SEMI, an industry association representing electronics manufacturing, where the goal was to identify smart manufacturing opportunities for the industry.
According to SEMI.org, “Smart manufacturing is defined as the use of production and sensor data with manufacturing technologies to enable adaptability in process.”
You can see it visualized in this infographic:
The keyword in my mind being "actionable."
Speakers at the smart manufacturing workshop included experts from companies like Intel, Siemens, Cisco, and IBM. Kirk Wheeler, R&D Engineer/Technologist at Intel, kicked off the event by pulling back the curtains on what Intel is doing to achieve smart manufacturing.
Smart Manufacturing Is About Saving Time
“Smart manufacturing is all about time,” Wheeler stated, and summed up the key objectives in the following way:
- Time to make a decision (ideally close to zero, or instantly)
- Time to raise an alert (ideally proactive, so before the issue occurs)
- Time to produce a high-yield product (requires less and less time)
So how do we get closer to these objectives?
Wheeler shared that you have to “collect a lot of data, then see what parts of it matter to you.”
Collecting data is important, but one challenge manufacturers face is knowing how much to invest in sensors and other technologies for data gathering. After all, spending dollars to get data that will save pennies is difficult to justify. Wheeler candidly stated that in order to deploy sensors throughout a factory “the infrastructure needed is ridiculous.” He added, "I'm not the ROI guy, but I assume they've run their numbers."
Manuel Aybar, Manager Industrial Engineering at Qorvo, Inc, another semiconductor manufacturer, spoke of another technology issue: The buy vs. build question of software. In talking about Qorvo's experience building applications from scratch, he recounted that some of the applications they’ve invested in building themselves have taken tens of thousands of development hours over the course of several years.
In that environment, adapting to a rapidly changing market and manufacturing environment becomes very difficult. By the time your application is completed, the marketplace has innovated and moved on. That's not to mention achieving a ROI.
A recent survey conducted by Putman Media supported Aybar's experience. Of manufacturing enterprise software deployments, respondents indicated that 73% only sometimes to never were able to able to achieve a measurable ROI (Return On Investment).
If you compare this reality alongside the goals of faster decision making, faster communication, and more efficient production, then the gap becomes glaring.
When and Where to Invest In Smart Manufacturing Data
So, how do we evaluate what data gathering technologies and approaches are worth the investment? Here are a few guidelines:
- The data should increase your visibility to hidden problems
- The data triggers an important human process
- The data is visualized so that it can be understood and acted upon
In reality, the goal of smart manufacturing data should be increased visibility. Visibility to issues that are happening or may be on the horizon and visibility to opportunities for improvement.
Intel's Wheeler shared an often overlooked truth: “Once you have ‘big data,’ the human component comes into play.”
Data and the visibility it can provide is worthless unless it can be acted upon by people to fuel improvement and innovation. Which leads us to the answer to our original questions: Is the goal to integrate with machines and sensors to get data? Or is the goal to solve problems?
The answer is clear. For most manufacturers, they could spend years solving the problems they already know about before needing to uncover more. Solving problems every day is the great differentiator of manufacturers that will thrive in the new smart manufacturing environment.