How I took SPC and TOC to War

If I were about to start flying the F-16 again, a single engine fighter, and I didn’t want it to become a giant lawn dart, I would want to know what has worked in the past to keep it flying.

As it happens, a very long time ago, I ran, among many shops (all were part of what is called a CRS or Component Repair Squadron), an engine overhaul shop.? At the time, this shop only processed GE F110 engines.? And we were getting ready to go to war, the first Gulf War.?

Now, a normal Tactical Fighter Wing will need to keep between 2 and 3 spare engines on the line.? The shop will normally produce perhaps one a month, if all goes well.? All was not going well when I arrived.

For the uninitiated, a Jet Engine, once overhauled, has to pass through the Engine Test Cell successfully.? If everything goes exactly perfect, it will pass on the first try.? Ours were taking in the neighborhood of five passes before they would pass.

BTW - If you ever get invited to witness a test cell run, wear ear protection.? You’ll need it.

We were also having personnel problems, both a shortage of and a skills problem.? I had just arrived from the depot, where I had been involved in rolling out the ancient art of SPC (Statistical Process Control) as part of a TQM (Total Quality Management) effort going on service wide.? I had also been using TOC (Theory of Constraints), to radically improve throughput on the B-52.

What do the two concepts, SPC and TOC have to do with each other?? A lot, it turns out.? You see, during an initial evaluation of our existing production processes back in the depot, I had found a 62% failure rate (not a typo). ??Using SPC and all the tools of TQM, we had steadily worked it down to almost zero failure – at least on some processes.? That’s when the concepts of TOC came into play.? It’s hard to say what the real constraint in a process is, if the quality of the process is not there.? Fix one issue (quality) and the constraint frequently moves.

So now back to my engine shop.? The big boss had called me in early on, and told me one, he had heard about what I had been up to back in the depot (that’s never a good thing to hear), and two, he needed a plan from me in a week to fix the engine shop.? Oh, by the way, we would be forward deploying to two geographically dispersed locations and would keep flying here at the home base.

Doing the math, taking into consideration the 5,000 mile long legs of the triangle that now made up the deployed base, the high engine test-cell failure rate, and the other issues, I came up with a plan, which would ultimately yield 84 spare engines being produced during the conflict.?

You might ask yourself, why did you need so many??

Quick lesson on fighter jets.? During peacetime, you’ll have a Fighter Wing, which will have between 70 and 80 jets, (about 76), and it will usually be divided up into 3 squadrons (I’ve seen bigger and smaller wings), and those squadrons will be composed of +_ 12 jets per flight.? They will typically fly what is called a 4-turn-4-turn-2 shift.? That’s four of them taking off, flying about 1.4. to 1.6 hours, landing, refueling/reloading, then 4 more, then land, then 2 more.?

As a maintenance officer, you try to make the jets fly out their hours so that if you look at when the engines will need maintenance, it will look like a waterfall chart.? Same goes for when they are scheduled into depot maintenance.?

Little did I know, the sortie lengths during the war would change dramatically.? We suddenly started seeing 8 hour long sorties.? Even before the fireworks started.? That meant the demand pattern, i.e., spare engine consumption rate, was suddenly looking like a rapidly approaching cliff that we would fall off of.?

That’s why we needed so many spare engines.

It’s been many decades since all this took place, but I think the lessons still apply.?

What did we do to make this happen:

First off, we identified the actual processes that engines went through in the shop.? Turns out, there were about 13 distinct processes, which we just labelled A process, B process, etc. Then we figured out how those processes were performing. ?That allowed us to use SPC (actually, several different chart types applicable to a job shop) to identify where the process variation was happening.? Some was due to equipment issues, i.e., jigs and fixtures not properly aligned or calibrated, some were personnel related.? If you have any experience doing this, you know you find a lot of issues before you get your processes in statistical control.? But we did.? Once we did, then the constraints in the system revealed themselves.?

The math also told us we needed to buy 3 new engines and we needed some additional engine mechanics and we needed to grab just about every engine carcass (cores, usually) out there, if we were going to make this happen. We also needed it fast.

But once we got the place running a little better, and isolated the causes of repeated test stand failures, we eventually got it to where it was a one and done pass every time.? Fun times.

?

?

?

?

?

?

?