Modern Manufacturing Analytics: The Perks and Pitfalls

Challenges: Process, Technology, Read

Posted by Christine Schmitz

Estimated Reading Time : 5 min.

artificial-intelligence-data-centerI, Machine, take you, Data, to be my wedded partner, ‘til death do us part. 

 

"In my experience with discrete part manufacturers, making parts is, not surprisingly, viewed as the primary revenue-producing activity and, therefore, those activities that directly affect part making are supported. These activities could include, for example, selecting the right endmill and associated machining parameters for an Inconel milling application, choosing the best coolant and application method (mist, flood, high pressure?) for a new titanium part, purchasing a new work holding system, or updating CAM software to enable constant feed machining. 

 

Activities that are not perceived to affect today's parts make it more difficult to establish a business case and, correspondingly, to justify resource investment (whether personnel or capital expenditures, or both). Historically, part inspection might have fit this category, but today's accuracy requirements have made in- and post-process inspection the norm. However, the emerging Industry 4.0 and associated tools, such as data analytics, is still finding its way into the "must invest to make parts" category. I am convinced that it will, but a critical mass of case studies and vendor support for convenient integration in the shop floor ecosystem will be required."  -- Tony Schmitz, Professor, University of Tennessee, Knoxville and ORNL Joint Faculty

 

 

I have a feeling Nick Goellner is going to love this story. 

 

My husband, Tony, and I have known Nick for a while and we both work with him, though separately. I collaborate with Nick and the team at MakingChips. Tony, on the other hand, consults with Nick on machine tool dynamics issues at AME. We are older than Nick, and I fear we may represent for him an illusory ‘manufacturing dream couple’ -- that is to say, a husband and wife who sit around all day and night discussing the industry from every angle, and being completely fulfilled by this interaction.

 

Sadly, this is an inaccurate portrait of me and Tony who, on some occasions, do not even read each other’s weekly posts, let alone discuss them over dinner.

 

This morning, however, Nick’s dream scenario was fulfilled. 

 

As Tony and I set about our Saturday morning routine of sipping through a pot’s worth of coffee, answering emails, paying bills, etc. I mentioned a figure from this week’s manufacturing news article which showed descriptive versus predictive versus prescriptive analytics. I asked Tony what he thought about manufacturing leaders struggling to know what to do with all the data they’re collecting and how to fully utilize IIoT and AI on the machine shop floor.

 

Tony gave me a great answer - something along the lines of business leaders being so focused on the immediate issue of production and traditional means of quality control, supply chain management, and the like, that it is difficult to allot the necessary resources toward fully transforming asset and facilities management with data analytics. 

 

I was thrilled. So thrilled, in fact, I actually asked: “Would you go on record saying that?” Of course, Tony said he would be happy to contribute, and provided the quote at the beginning of this article.

 

While I know this true story somehow fulfills a fantastical vision of Nick’s, the fact is my world and Tony’s world rarely collide so easily. This got me thinking about how our marriage, while not nearly as perfect as Nick might imagine, seems a perfect metaphor for the metalworking industry as a whole. 

 

Tony is at the top of the academic manufacturing and machine tool research field, with all the advancement and possibility theoretically available to alleviate nearly every single issue within the entire global manufacturing industry, and here I am, just trying my best to keep up to date with industry advancements week by week and make some sense of it all.

 

I am the blueprint; Tony is the 3D model. I am the hand and the ear; Tony is the accelerometer and the computer. I’m just a small-town girl living in a lonely world; Tony’s just a city boy born and raised in South Detroit ... not really, but I hope I’ve got you singing, “Don’t Stop Believing”.

 

I came across an incredibly good blog from Sight Machine this week covering all aspects of digital manufacturing. There are too many worthwhile articles to reference, and honestly, they are so well written, I couldn’t do any better than just to say - go read them for yourself. However, one particular article seemed to address this industry disconnect which Tony and I discussed. 

 

Here’s a quote:

"In the old days, optimization efforts were largely based on hunches and guesswork, because manufacturers didn’t have a lot of experiential data to work with. IoT and other innovations have changed that. Today’s shop floors are teeming with sensors and other devices that acquire and transmit real-time readings on every aspect of production: temperatures, pressures, tolerances, ingredient mixes, the arrival of new parts, and so on. Input is abundant. But unfortunately, answers remain few and far between. Why is that still so?"

 

This takes me back to the original figure I mentioned to Tony, showing the three categories of data: prescriptive, predictive and descriptive. According to this week’s manufacturing news article from American Machinist, AI technologies are increasingly being embedded within every major enterprise software offering for manufacturers. 

 

The trouble is, manufacturing leaders also lack a unified system to orchestrate enormous information inflows. Traditional business software is structured around workflows, not data. This type of software is great at automating business processes, but not for analyzing factory data.

 

According to the Sight Machine article, software structured around data rather than workflows -- similar to the platforms pioneered by Google, Netflix, and Amazon -- can help create a digital twin of a shop’s entire production process: an end-to-end representation of all machines, materials, environmental inputs, and the complex relationships among them.  

 

Going back to me and Tony and our metaphorical representation of the manufacturing industry, after nearly 25 years of marriage, we know a lack of communication is at the heart of almost every one of our conflicts. Similarly, poor communication continues to be one of the biggest issues within our industry as well. 

 

While manufacturers continue to stockpile data for problem-solving and analytic decision making through IIoT and other means, unfortunately even when the data is applicable to a given problem, it’s often not in a form that makes it usable for analysis.

 

If only this problem could be solved with a fancy date night or weekend getaway. Sadly, there’s not much room for romance when we are talking about the manufacturing issues of our time.  

 

Sorry, Nick. 

 

However, don’t call the divorce lawyers either. This is a communications issue I believe will be solved with time. While there’s still some ground to cover, there’s equal commitment on both sides and progress is being made with each passing day. 

 

Don’t Stop Believing.

 

BAM!

Subscribe to MakingChips and get notified on what's new or text CHIPS to 38470!