The pitfalls of “black box” software & design

The pitfalls of “black box” software & design

Black box design?  Black box software?  Let’s explore and explain what it’s all about.  We’ll focus chiefly on concrete analysis/design in most of our discussion below, but the same principles and pitfalls around black box design apply to all materials and analysis software…

When it comes to designing reinforced concrete, most structural engineers have a “love it or hate it” relationship. Of all the materials we build with (and hence, all the materials engineers have to be familiar with), concrete is arguably the most complex and requires the most inherent understanding to truly “engineer” it.

Other materials have a bit more going for them, chiefly because they’re usually “off the shelf” products with fixed properties. Steel is a great example: Choose a 250UB31 and you can instantly dial up all its properties from the book. There’s no end of published capacity tables to tell you if it’s the right beam for the job. And, if it doesn’t cut the mustard, you simply move to the next size up.  Timber is similar – choose a 190×45 piece of F7 pine and it’s not difficult to analyse what it can and can’t do.

Reinforced concrete, on the other hand, is a completely different beast, chiefly due to two significant features: (1) It’s a composite material, consisting of the reinforcing steel and the concrete matrix – which is, itself, a composite of cement, aggregate, and water. (2) In contrast to an “off the shelf” product, nearly every parameter of a concrete member is variable and can be tweaked. When designing a concrete beam, the engineer can tinker with (i) the concrete’s compressive strength and mix design, (ii) its dimensions, i.e. depth and width, and (iii) the amount and configuration of the reinforcement. Even this last item is full of different permutations – the bar diameter, the number of bars, the bars’ yield stress, and how/where they sit within the beam’s profile can all be manipulated to achieve different outcomes.

The science and theory behind it all is admittedly complex – so much so that design codes around the globe all have different interpretations, theory, and limitations. The astute and diligent engineer has to be across concrete’s concepts of stress, strain, ductility, rectangular stress blocks, lever arms, neutral axis, tensile steel, compressive steel, ultimate behaviour, service behaviour, shrinkage, creep, variable reduction factors, flexural strength, both transverse and longitudinal shear, torsion, positive and negative bending, transformed section analyses, uncracked behaviour, cracked behaviour, and more. And that’s just to design the beam’s capacity and performance on paper!  That’s all before we come to the issues and difficulties of actually achieving it on site: There’s correct cover to the reinforcement, workability, vibration, drying time, curing, plastic settlement, drying shrinkage, and durability….all the time hoping that the formworkers, steelfixers, pump operators, and screeders/finishers all execute precisely what your calculations assumed.  Oh, and also that the weather holds on the day of the pour!

Concrete design: The rectangular stress block diagram

As such, the design of reinforced concrete and the process involved remains one of the last bastions of “classic” engineering: You trial a section; see if it works; if it doesn’t, you then change one or more parameters and trial it again until you’ve achieved the requisite strength and optimised the section. To be honest, it ain’t easy, and many engineers find it daunting, preferring to steer their work in other directions. Or, more commonly, they’re increasingly relying on software to do most of the heavy lifting.

The structural engineers of even just 20 years ago (let alone 40, 60, 80 years ago) would look on with green envy at what today’s engineers can utilise on their computers to analyse and design concrete. There are even apps for your phone that can give you a concrete beam’s strength in seconds – something an engineer in the 1970’s would have taken an hour to do by hand and with a slide rule, only to get a less accurate answer.

But – at the risk of perpetuating a lament that’s been voiced now for two decades – is this increasing reliance on technology coming at the expense of genuine knowledge, understanding, and experience? It’s what some engineers refer to as “black box engineering” or black box design:  Putting some parameters into a device (the black box) and blindly relying on whatever output comes out the other end.  By becoming reliant and dependent on the machinations inside the device and the process involved – none of which you can actually see or interrogate –  users start to lose the feel and understanding of what’s actually going on. Content that the black box will do their work for them and think on their behalf, engineers are no longer investing in understanding how reinforced concrete actually works. There are parallels in other aspects of 21st century life – it’s the same attitude as “Why should I learn how to spell when my word processor or phone will auto-correct all my errors for me?” But what happens when you’re offline and have to physically write something with a pen on a piece of paper? No red line will magically appear under your words to rescue you…

Concrete design : Knowledge versus wisdom, and black box design

Black box design won’t always lead to the smartest or most appropriate solution

 

What happens in the long run is that becoming over-reliant on engineering software drives a wedge between knowledge and wisdom. There’s an old line, “Knowledge is knowing that tomatoes are a fruit and not a vegetable. Wisdom is knowing not to put them in a fruit salad.”  Similarly, knowledge is knowing that six 20mm diameter rebars gives exactly the same area of steel as three 28mm diameter bars, while wisdom is knowing that six 20mm diameter bars won’t actually fit in your 350mm wide beam! (Unless you put them in two layers, thereby reducing your lever arm and thus the beam’s strength.  Many black boxes won’t catch such oversights).

(An engineer with wisdom might also consider that those six N20’s may reduce the amount of stress in the bars at the service state and thus reduce the risk of flexural cracking in the soffit! Illustrating, again, the many complexities with designing reinforced concrete! Black box design and most software packages won’t capture or consider these considerations, focussing only on strength.)

Over-reliance and an unhealthy dependence on software is also driving a skill and knowledge shortage through today’s typical design offices. I have endeavoured to get this far without bringing age or “generation” into the discussion, but there are some truisms I’ve observed. When faced with a concrete design exercise, many younger engineers who’ve only known a digital process seem incapable of going about the task without software.  With many engineering software tools requiring subscription or expensive licences, most medium-to-large consulting firms battle with the problem of having more employees than available licences. Watch what happens when five engineers need to design their own concrete slabs and there are only two licences available for the FEA software! Suffice it to say, the three unlucky ones don’t turn to alternative methods or find other ways of going about the task….they’ll sit and wait it out until a licence becomes available.  Not exactly good for employee utilisation and efficiency!

Black box design : Screwdriver versus a cordless drill

Make no mistake, we get the convenience. Why would you choose to use a manual screwdriver when you’ve got a cordless power drill at your disposal? But there’s a reason why tradies don’t throw out their trusty manual tools…..one day, the battery will be flat, or the cordless drill won’t fit in the operable space. Good tradies don’t lose their ability to think and work “old school” when the moment requires it.  Neither should good engineers. That is the challenge for the modern consulting firm, and I urge younger engineers to get out their concrete textbooks and acquaint themselves with the finer points of design.

The efficient and economic design of reinforced concrete relies on a comprehensive and holistic understanding of concrete. As we explored above, there are so many variables an engineer can tweak and manipulate to drive a member’s properties and behaviour. You need to know which levers to pull, and going round and round inside a black box where the levers are neither obvious nor (sometimes) accessible will have its shortfalls. It’s only through consistent application of the theory – and that means doing hand-calcs, hand-sketches, getting your hands dirty – that you truly become conversant with the material. Black box engineering – as wonderful, fast, and powerful as it is – is robbing some engineers of reaching their true potential and delivering the best solutions for their clients.

Cheers,
AD

PS:  If you enjoyed this piece, you can subscribe to The Working Engineer for free and be notified each time a new article gets posted.

Did you like this article?  Why not subscribe for more?

Subscribe for free and receive an email notification when we publish new articles...

Invalid email address
We promise not to spam you. And you can unsubscribe at any time.

The Working Engineer logo

Leave a Reply

Your email address will not be published. Required fields are marked *

Back to top