Data

Putting data to work to solve industry-wide challenges

594 0

Is your company hoarding data from additive manufacturing processes in case of a rainy day? It’s not uncommon, writes our columnist SJ, but we need to put that data to work on solving industry-wide quality challenges


When I was a kid, I was the child of a hoarder. But my mother was not a hoarder of just anything. The habit was more specific for her. She was notorious for hoarding… wait for it… napkins.

If we went to get takeout, it was, “Can I get extra napkins, please?”. At a restaurant, “Excuse me, can we have more napkins for the table?”

They would be bursting out of her purse, out of the glove compartment in the car, from the drawers of our kitchen.

One day, I made the unfortunate mistake of opening a cabinet in the bathroom and a huge wad of them fell forward and upended the entire contents of that storage space upon my head.

Frustrated, I stomped my little 8-year-old feet to my mother’s office to demand an explanation. In my most adult voice, I said rather petulantly, “Why do you need all these napkins, woman?”

And she looked at me, my hair covered in napkins, and said gently, “Because, silly, you never know when you may need them for a rainy day.”

Advertisement
Advertisement

Coping with variance

Twenty years later, and I still think back to that conversation quite often as I navigate the difficult landscape of certifying parts built using additive manufacturing – a very new process compared to its traditional predecessors.

As a sweeping generalisation, most companies have an established path to part certification and part qualification for additive that is understood internally at their organisation.

But as yet, there isn’t an agreed-upon-by-all-parties industry standard that service bureaus and other advanced manufacturing firms are held to. Which means that for one part, there can be a variance in material characteristics, depending on who printed it and where.

These material variances can affect design parameters, which in turn, impact the end performance of the part.

To customers, this variance comes across as the manufacturing process being unstable (or even worse – uncontrolled), so it’s no surprise to me when customers start asking for extra napkins.

Customers amass titanic data piles relating to the AM process, but then don’t do anything productive with them once a project is over

Data stockpiles

In this context, a typical stack of ‘napkins’ consists of data. This data is elicited by customers asking the following kinds of questions: “How many test bars should I print? Can I see X-Ray or CT results post-print like I can with a casting? How many part cut-ups from each batch do you think we need to be safe? What mechanical testing do you recommend for this material? Are you sure we can’t fit more test bars on the plate? Do you think you could send me all the photos for each layer of the build, so we can check for anomalies? Do you mind sending me the oxygen report post print, so we can have it for our records?”

From my experience, most of my customers are amassing titanic data piles in order to try and understand the additive manufacturing process – but they’re neither sharing their results, because of intellectual property issues, nor doing anything productive with that data once a project is over.

They just lock it away in a drawer for safekeeping, ‘in case of a rainy day’. Which makes me stomp my boots in frustration (yet again!), because there’s so much more we could be doing with it.

Putting data to work

First, we could be investing a whole lot more in big data for advanced manufacturing. This would allow us to collectively analyse and extract the information we need to build better certifications for parts made with non-traditional methods.

Second, we could take evidence from the results that we get and align it with the data that we’re seeing from in-process monitoring systems.

With the advent of in-process monitoring, we now have the ability to compare building conditions and their variability from one part to the next before having to pay for downstream post processes or mechanical testing. That would effectively allow us to iterate the process faster with lower costs.

Once our variability window is established, we could tie it to a widely accepted standard (for example, Six Sigma, or Taguchi, or whatever floats your napkin).

Quality standards improve our repeatability, which in turn, improves reliability and gives our customers more confidence. So, stop hoarding your napkins, would ya? And let’s take our industry — and quality — to the next level.


Get in touch:

SJ is a metal additive engineer aka THEE Hot Girl of Metal Printing. She currently works as a metal additive applications engineer providing AM solutions and #3dprinting of metal parts to help create a decarbonised world. Get in touch at @inconelle on Twitter