Back to Basics

Aeronautics & Space Engineering general discussion
, , ,
#1

Don’t know if you folks have had this feeling before. Like some intuition or idea is building in the back of your mind but you can’t yet sit down to figure it out… then you get out of bed one morning and it’s crystal clear. You write it down immediately and, presto, it’s so simple and it fits on 3 pages.

What am I talking about?

For the past year - in fact ever since I started the job I started in June last year, I’ve been managing a 3rd party contractor working on the structural analysis of a new product design. The new product was already sold, delivery due 2026, drawings released to production, just some last details to tidy up with the analysis. The analysis sub-contract was signed just before I arrived. The week of my arrival, they had the initial data load. The week after, the first problems cropped up. The story really hasn’t changed since then.

To be clear, the contractor is doing excellent work. They are diligent Finite Element Model builders, and have insight into joint analysis and material failure that I’ve come to trust in our regular discussions about their progress with the analysis. We handed them the working FEM models of the product that our company had developed over the previous years, and they’ve transformed them into high-fidelity representations of how our system works. Ask a question and they’re very quick with the answers. I will recommend them to people in the future.

To be frank, and I’m glad I have a fairly anonymous handle on a fairly obscure forum here, the problems are primarily of my own company’s making. This “discovery” process, where weekly or monthly this contractor discovers a new error that my predecessor made some time in the past, has profoundly undermined my ability to bond with the engineering team I’m working with. It’s difficult to trust people who make mistakes like ones they have found in our FEMs. Incorrectly applied loads, incorrect safety factor, incorrectly contact stiffness, incorrect mass properties, incorrectly place reactions, incorrect models of the surrounding airframe. Many examples of both oversimplification of some structures (unrealistic behaviour) and excessive elaboration of other parts, bogging down the computing.

Why am I mentioning this now?

This morning I woke up with the whole thing sorted out in my mind. The free-body diagram was finally in focus, the math in all 3 coordinates added up. I think I have finally seen enough of the load cases, structural reactions, and aerodynamics, that even before writing it down I had the first few equations worked out in my head.

2 hours later (yes, a Sunday morning wasted on work) I had my 3 pages and a free-body diagram that illustrates the result. It concisely demonstrates that there was NEVER any possibility that the proposed design would work. The free-body diagram and the sum of forces & moments I used is all first-year engineering school stuff.

The complexity of these sophisticated FEM models with 100,000 elements and 100 load cases has concealed from our designers and engineers the INSIGHT they need to make a sound design.

I feel a little guilty, not focusing myself on this solution months ago, especially when the flaws with the design just wouldn’t go away. What a lot of trouble I could have saved if I’d given myself this insight 6 months ago.

OTOH, most (not all) members of my engineering team also went to engineering school, and my predecessor had a very high standing in the company, and in the local engineering realm beyond our company. For all of them to not attempt this evaluation before embarking on the design… again my ability to trust suffers.

I’m not asking a question, I’m just getting something off my chest. What a way to remind myself that the tried-and-true methods still work, even in a world of fast computers and high-fidelity models. All I needed was a pencil, drawing paper, and pocket calculator.

#2

“The purpose of computing is insight, not numbers.”
Richard Hamming
Numerical Methods for Scientists and Engineers (1962)

#3

Well, good for applying the fundamentals and reaching a conclusion. Best to find an unsatisfactory design before released for construction/use.

Question: if you have applied the fundamentals and reached an unsatisfactory solution, how come the computer models did not?

the referenced textbook is what i had used in my college days.

#4

Actually, the FEM is giving negative results. It’s been a creeping death, the kind that sneaks up on you because the initial design wasn’t evaluated properly. The new product is a derivation from a previous product. The recent work was meant to just extend the capability, but instead the 3rd party contractor revealed all the flaws of the previous product, which have been inherited in the new one.

#5

@pmover,
I almost didn’t reply to your question, not fully.

The “creeping death” has also been misleading. The hints of failure have been there for a long time, but the distractions, such as the faulty model, have prevented me from realizing it. I was starting to get realistic results from the sub-contractor sooner than I was prepared to believe. Negative margins? Tell the sub to check the input data again, and inevitably a flaw would be found. When this process became routine, I lost sight of the underlying message. If the message had been simple, or I could have trusted the source of the message, I would have been more likely to believe it and act on it. Now I’ve wasted a lot of time fixing a very elaborate FEM model, instead of the product, when in hindsight I could have done my own hand-calcs a year ago.

Just being brutally honest about this, and reporting while I’m still “in the thick of it”.

I haven’t exactly figured out how I’m going to solve this problem, either.

#6

yes, you did . . .
While i’m not that familiar with FEM, i certainly “think” i would question negative results. bad data input is bad output data; go no further.

“I haven’t exactly figured out how I’m going to solve this problem, either.” when i read this statement, i immediately recalled a situation i encountered 25-years ago.

while at an EPC firm in Houston and when construction was nearing completion for a gas plant, the process engineer was completing his work. he came to my office first thing friday morning with a heat exchanger data sheet to question/review/confirm a statement on the supplier quoted/accepted data sheet. I contacted the supplier and yes, a major “oversight” occurred. the supplier mis-interpreted the IFP datasheet and built 2 exchangers that were 50% of required duty. the process engr & I promptly informed the PM. Yes, before i departed at noon that day:

  • heat transfer specialists & process engrs analyzed the situation,
  • VP called a meeting and wanted to know what happened and made it clear this situation should not happen again; a very serious oversight,
  • properly sized exchangers were ordered w/ an exceptional delivery, 2-3 weeks.

in many ways, i felt terrible, but also relieved in that we, individuals & company, ended up doing the right corrective action. There was no ass-covering or otherwise regarding this matter. we promptly dealt with it in a very professional manner. This was definitely a lesson learned and one to remember. Yes, i may have had a moment that i would be gone, but this matter was a shared responsibility as several engr individuals reviewed the purchase package, which included that exchanger datasheet.

you actually wrote the answer: “Just being brutally honest about this”!

Actually, all individuals involved should get some gratification knowing “someone” made a rather “simple” analysis indicating negative results.

Hang in there!

1 Like
#7

Yup! That’s the next step on my list of things to do. Since I can see it coming (and already notified the execs of the problem) I’m planning for the solution. Nothing but that minor detail of “what” to do left to sort out. A trifle.

#8

I’ve not had a third party hand me a sack of stuff like that, but have had the privilege of finding the sack on my own, and having to try hard to push it back up the corporate delivery chute. It’s tough when the powers that be have already invested heavily in the idea. Tougher yet when the flawed idea is sold against your advice, and delivered. It’s rewarding in one sense - when the idea gets proven faulty in the field, and the powers that be come back from that asking you to fix it…and you do.

I have every expectation that you will prevail in fixing it Spar, and that you realize that in my example the pain was much worse for the company because you have the opportunity to fix it when doing so is still not too expensive.

#9

@btrueblood
Fortunately, I have the executive on my side right now. They support my investigation and are encouraging me to learn more, dig deeper, and keep challenging. I’m lucky to be enjoying this much trust.

#10

Since fist posting about this, I’ve witnessed the goals of my executives diverge, pull me in opposite directions, and then re-converge on a common goal. It was uncomfortable for a time. These decisions aren’t easy so I understand it was difficult to find the right direction.

I’m making my way toward solutions, but the fact remains that in some ways, my predecessor was incompetent and made promises he couldn’t keep. The decision to bring me on was a decision by the executive to act on their suspicions of his incompetence. My arrival and investigations have only confirmed they were right in this regard. They got nothing but a fleeting satisfaction from being right - the bitter aftertaste is a line of products that don’t perform as well as they should.

#11

A different way to express my original point. The phrase “Back to basics” can be plotted in time.

Modern methods to solve engineering problems are based on computation. It was obviously born in the 20th century. The mathematics typically used by computers is based on boolean algebra and matrices and the like, hence that math was born in the 19th century.

Other methods used in engineering are based on Newton’s laws, such as equilibrium of forces and the principle of inertia (his 1st and 3rd laws). These date back to the 17th century. So by comparison, any solution to an engineering problem that can use 17th century mathematics is more “basic” than a solution based on 20th century methods.

I tend to like basic solutions. It’s easier to digest the problem and its solution, and I feel there are many advantages to keeping it simple. Once the math is modeled, variations and developments can follow quickly.

It’s not the prevailing opinion these days, it seems. I’m working with many engineers who use and prefer FEM analysis. They all have their favourite software solutions to use. Many of them are very skilled at using the software, which impresses me, and they can gain some very interesting insights into the character of a structure with its use.

While I can use FEM, and I’m trained on some of it, I usually don’t choose it. At times I use FEM to solve tricky problems that defy simple “Basic” solutions (usually because of redundancy or stiffness characteristics). Even then, only to the limited extent that a certain problem needs, and no more.

In the past, I’ve done many projects and solved many problems with simple “Basic” math alone. Thinking back now ( and this is what’s motivating my update today ) I notice that many of my past solutions have been optimized to allow for the “Basic” solution method.

Need a truss? Use only enough members to allow for a static solution.
Need a beam? Use an example that’s available in Roark.
Need a plate? Make it thick enough for a linear buckling solution.

There may be a harmonious relationship that joins simple design, simple math, and simple manufacturing. As one allows any of these 3 things to become sophisticated, it makes the other 2 more complicated. The advantage gained by using elaborate FEM models is eroded by the tendency to make the design, and thus the fabrication, more complex.

Sorry - not actually asking a question, am I? This look more like a blog entry!

#12

I’ve had the luxury/privelege in the last dozen or so years to help my company go from garage shop to a middling-sized shop. Production runs of 10’s have grown to runs of several thousand now. And so, we’ve gone from barstock and weldments to castings and forgings. Designing parts for efficient material use in castings and forgings can be fun, if challenging. I certainly made use of FEM when available, as it can help you fine tune things for weight savings…though you need to have the ideas for the design in the first place (or you end up with hallucinations from AI).