Consider an ordinary staircase with a banister rail supported by a wall at one end and a steel post of constant profile at the other end. The base of the post is embedded rigidly in the fabric of whatever structure the stairs are part of. The post must not bend beyond a set (small) amount, permanently deform, or fail to support the guard rail, in the event that someone falls downstairs or against it, and the maximum deflection of the post in the event of a fall is in any case to be small compared to its length (rough indication: 0.5–4.0% or say ~5–40 mm vs ~1000 mm).

The rail and post, and any fixings, are each thick enough and well enough secured, that they will not shear or become loosened by any plausible impact of a person (or persons, including anything they carry) falling down the stairs and against them at any point. Within these criteria, it's desired to make the post with a narrow profile, and therefore minimally thick.

From a pure engineering perspective (ignoring any building codes for now) is it sufficient to reason that any collision/impact will be elastic and therefore the minimum profile for the post can be derived by:

  1. Estimating a maximum KE and PE for the falling persons,
  2. Setting a maximum desired deflection of the post.
  3. Reasoning on a pure energy basis that the post can be treated as a simple/ideal deflected cantilever beam, and its minimum profile is that which satisfies strain energy of the steel at maximum desired deflection >= maximum KE+PE of likely falling objects?

It seems to make sense and yet I am not too familiar with impulsive impact (more familiar with statics), and I'd like to check if there's a reason this would be an underestimate or must be treated with caution. (For example maybe a falling object would have CoG slightly higher than the actual rail height? Or other reasons might exist that this would lead to an insufficient result?)

  • 1
    $\begingroup$ It all seems to make sense, except you have little detail on fixings, and I suspect they'll be the weakest point. Also, you don't mention the required longevity, environmental conditions, or consider fatigue from repeated stress on components. Finally, regarding CoG, AFAIK that's barely an engineering aspect; rather a requirement spec, depending on regs and/or client requirements $\endgroup$
    – Jodes
    Commented May 17, 2016 at 8:15
  • $\begingroup$ You also need to consider the physical dimensions of the restraint system. If you design something that is strong enough to survive the impact but acts like razor wire, that's probably not a good idea. $\endgroup$
    – alephzero
    Commented May 17, 2016 at 11:58
  • $\begingroup$ Both good points, thanks! Not sure how I will fix it but the fixing will be visible and checked to relevant code. Physical dimensions not an issue - the core will probably be 25mm steel cladded to 40-50mm with timber but any severe impact will almost certainly be to the side (which isn't size-limited and will be broad) not to the narrow edge. The issue is available space in one dimension not both, so I'm limited on one side only (thickness of balustrade rail as you'd look at it) not both dimensions $\endgroup$
    – Stilez
    Commented May 17, 2016 at 13:19

1 Answer 1


Using a purely energy-based approach, while theoretically possible, is not recommended. Not because it wouldn't work, but simply because most codes don't use such methods.

The reason is that railings, as well as any structural elements, are ruled by codes. In Brazil, the relevant code is the NBR 14718, which defines the static and dynamic horizontal and vertical forces that a railing must resist. Could you transform these forces into energy? Sure, probably. But you can't simply "estimate a maximum kinetic and potential energy for a falling person" and use that, because you won't know if your values are sufficient to satisfy the codes. Well, unless you do your estimates, convert them into forces and check that they are sufficient... at which point, why not just use the forces?

The codes also define factors of safety, which a purely theoretical method such as the one you're suggesting can't generate. The Brazilian code at least also defines the allowable deflections at $\dfrac{L}{250}$, where $L$ is the railing's span.

And then there's the issue of dynamic impact factors. From a purely energy perspective I don't believe this can be easily solved. After all, a load applied instantaneously will generate a temporary deflection which is twice as large as the permanent (elastic) deflection.

  • $\begingroup$ Can you edit to say more about the last part ("dynamic impact factors... twice as large")? In particular, I don't understand why a temporary deflection would be larger, or different from a spring (within elastic limits) when as far as I know, it's basic physics that a spring's maximum extension for a dropped mass can be calculated on energy principles? $\endgroup$
    – Stilez
    Commented May 18, 2016 at 9:11

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.