Supply Chain Crisis - 3D Printing to the Rescue!

First there was Covid, now Suez. Following the flurry of posts advertising '3D printing' as the solution to every supply chain crisis, I felt compelled to share some cautionary points. Not to discourage, but rather to help people think critically about the steps necessary to really use 3D Printing / Additive Manufacturing (AM) as a tool for mitigating supply chain disruption risk.

We know that one common difference of most AM processes versus many manufacturing technologies is the relatively short lead time to produce a part. This is primarily because the process itself is digitally driven and doesn't require part-specific tooling. This naturally leads people to think about on-demand production in situations where missing parts are causing disproportionate consequences for their business.

Perhaps it goes something like this:

You operate Just-In-Time stock management and your stock level is approaching zero for some parts. You have no idea how long it will take for your new stock to arrive as it is stuck in the log jam of ships outside the Suez canal. Your production will grind to a halt in a few days due to the missing parts. You speak to the supplier but they have nothing on the shelf and the lead time for producing another batch to send by air-freight is 12 weeks. Production needs the missing parts ASAP, and once production stops the company loses $1m/day due to delivery delays, all for the sake of a few missing brackets.

But it's ok, you saw a post from some guy on LinkedIn who said he could 3D Print the parts for you. So you send him the CAD models and get the parts printed in only a couple of days. Production starts to flow again. Everyone is happy. You are a hero. Perhaps you get a promotion, hell - maybe they give you a medal. And then you wake up...

Back to reality...

In most industries the chicken and egg of qualification is a huge barrier to using AM/3D Printing in this way. Contrary to (apparently) popular belief most companies do follow and enforce some kind of quality management standard, especially those companies that produce physical hardware. The idea that you can just swap out one manufacturing process for another on a whim is sadly not a realistic one. Aerospace may have higher standards than many industries for what they expect in terms of a qualification, but they are not alone in expecting documented proof that a new material and process is fit for the purpose intended. Almost all industries have standards and approvals, product safety certifications, etc.

Quick fix?

The timescales for qualification will normally obliterate any chance of solving such a missing part crisis in an opportunistic way. You normally need that missing part in a matter of days, not months. In general AM can only be successfully applied in this way within an environment where the application of AM is already qualified and accepted for production parts.

If you want the flexibility to use AM for urgent parts then you will need a linked up strategy which includes using AM for production first.

But what if you want to treat AM like an insurance policy?

Indeed you could also do the qualification as a sort of 'insurance policy', but you would need to identify which materials and families of parts to qualify in order to protect yourself from such unpredictable events. Most importantly you would need to convince someone to pay for it. Even if you were very succesful in convincing your management of the risk of missing parts, the obvious alternative they have available might be to just increase minimum stock levels for those parts and carry more inventory.

The trade-off between increasing stock levels versus qualification of AM as an insurance policy is of course an economic one. The question is how much risk you can mitigate for a given cost of qualification, compared to the cost of the alternative option of increasing stock levels to achieve the same goal.

A key lever in this trade-off is how broad scope of parts you can cover with a given scope of qualification.

From a material qualification perspective, if you can qualify one material which is superior to 5 different alloys which are used in your conventional bill-of-material, then that is better to do that than to qualify 5 different materials to cover the same scope, for reasons that should be obvious (although considering how often people seem to want to do that... I do wonder).

Another important factor is the breadth of the scope of application which can be qualified for. If you have a broad scope which can be qualified together because they have a similar functional requirements envelope (e.g. electrical harness brackets) and you can cover 1000's of small variations in part geometry with a single 'family' qualification. That is far better than qualifying each single design of part. The obvious reason being that you can now be ready to print whichever one of those thousands of part variants might go missing in the future, rather than keeping extra parts of every variant on the shelf constantly 'just-in-case'.

Another key factor to the trade-off is how long any alternative mitigation plan would take to implement. If the leadtime from an alternative source could be managed in a few days (e.g. buy a standard part from an alternative supplier) then you don't need to increase the stock level much to mitigate the risk that your primary supply chain is disrupted, even if it is for a long period of time. However, if the lead time to enable an alternative source is 18 months (e.g. develop and qualify a new forging process & supplier), then you need keep a lot of additional parts on the shelf to mitigate a disruption. In the latter case a precautionary qualification of an AM process 'might' make sense when compared to the alternative of increasing stock levels.

Once AM is already being applied in a field of application, then using it for risk mitigation of similar missing parts is of course much more straightforward, because everything is in place to enable it. In fact for AM parts used in production you will probably just reduce the amount of stock held as risk mitigation because the lead time is shorter and the economical minimum order quantities are smaller. There are equations for this stuff which are used by businesses every day to decide how much inventory they should hold - those rules don't fundamentally change.

Considering the disproportionate impact to the revenue of a business whcih can be caused by missing parts, even when they are relatively low value, there is clearly a desire to find solutions. AM can be an important tool, but we also need to be realistic.

The elephant in the room

Even though the AM process itself doesn't require part specific tooling, that doesn't mean the process chain needed to produce a functional part is quite so simple. Machining fixtures, inspection fixtures, post-treatment processes, galvanic treatments, painting, etc. All of these things are still necessary and (at least today) they are managed the old fashioned way. So for most parts, using AM will only solve a part of the problem, and there will still be a lot of other problems to solve to get your missing part to the shop floor. It will get better, but it isn't there yet.

The other obvious gotcha is that the future demand for parts in these scenarios will by nature be unpredictable and require very short lead times. If a crisis affects a whole supply chain and everyone wants to lean on AM to solve it that creates a bottleneck. Either you need to have lots of spare capacity most of the time so as to protect lead times during peak demand (very expensive) or else something has to be delayed to allow the expedited job to take precedent.

As Columbo would say "...just one more thing"

If you do qualify AM as a risk mitigation strategy, you can't be sure that the equipment & processes you qualify today will still be supported when the crisis comes, and so any qualification might be obsolete by the time it is needed. For that reason I still don't believe that it makes sense to dedicate a production facility to 'risk mitigation' only.

If you did dedicate a facility for such purposes then the average utilization of equipment would generally be very low in order to maintain a short lead time during peaks of demand. If you consider the high fixed costs of running an AM facility (including the cost of qualification) then this would make very poor business sense. It therefore seems much more sensible for such parts to be produced as a relatively small proportion of the production workload of a facility which is otherwise dedicated to more predictable series production using the same qualified materials and processes, not to mention all the post processes and QA facilities you will need in-house in order to deliver finished parts while keeping the lead time reasonable...

This might all seem a little bit negative, but that really isn't my intention. It's just that I don't think it is worth chasing rainbows. The economy doesn't reward ideas that sound nice, it rewards ideas which really work and solve real problems that people face. The prerequisite to solving real problems is fully understanding them, and also understanding the many alternative solutions to the one you have in mind. I would be the first to admit that I probably don't understand this problem fully - but I am trying to.

I really do believe that AM will play a role in changing how companies will handle their supply chains in the future, and that it will reduce how much inventory they will need to hold, and it will get them out of trouble when supply chain crises hit. However I don't believe that it will happen overnight, and I don't believe it will be applied wholesale across the board and replace every warehouse with a 3D-Printing hub. I also don't think it will be easy.

As with all such visionary ideas, there is a kernel of something there, but the path to implementation will be a long and winding one and nowhere near as obvious as some would like to portray...