.png%3Ftable%3Dblock%26id%3D1902599d-e9d4-4633-a58a-0182495851f9%26cache%3Dv2&w=3840&q=75)
For sustainability reasons, and for the betterment of society, it would be beneficial to be able to design structures that are leaner than are currently constructed. Structures that use less material, are optimised or are even eliminated altogether. Are there ways to revise limit states design to give better outcomes?
Generally with Limit State Design, there is a 5% chance that, while using the appropriate factors, the structure will fail. This probability is very small. And, sure, we don’t want structures to fail. Our primary goal is to protect citizens and provide them a safe structure. However, materials exist in such a way that failures are not catastrophic. Most structures provide warnings that they are over stressed or going to fail well before failure. The warnings are excessive deflections, cracking or other signs of distress. If this occurs, we can observe it and undertake measures to rectify it. I believe that in certain circumstances, we can increase the likelihood that a structure will fail.
Engineers use Limit State Design to design buildings that do not yet exist, using materials that have not yet been made, and workmanship that has not yet been determined. LSD does not specifically consider that actual structure that is in place. It is based on a theoretical building. If we could precisely determine the capacity and performance of the actual building we could reduce the uncertainty.
Lets look for a moment beyond LSD. Is there a way to make leaner designs even beyond the probabilities that LSD captures? Aleatory and epistemic uncertainty are two different types of uncertainty. Aleatory uncertainty is uncertainty that exists in the randomness of nature. While epistemic uncertainty is uncertainty based on our lack of knowledge. LSD deals with Aleatory uncertainty only, or randomness that exists in nature. Epistemic uncertainty is not tackled within LSD. However, Epistemic uncertainty permeates throughout everything we do as structural engineers. It is almost impossible to produce a comprehensive list of these uncertainties but they include things like:
- Simplification of the structural engineering modelling process,
- Codes and regulations that attempt to change structures to deterministic and predictable structures using formulas,
- Structural engineers ability to analyse and interpret results from complex computer modelling procedures.
Along with this, there is also the idea that engineers carry their own biases and prejudices. These may be based on knowledge, experience, environment and the ecosystem that exists at a project level. And there is of course, ‘noise’ that exists with the engineer including general tolerance for uncertainty and tolerance for risk, situational noise and occasion noise relating to generally the mood they are in at the time when they do the work.
If we could incorporate all these uncertainties into the design, we could provide better, more environmentally sustainable solutions.
No wonder people complain that we over-engineer "everything". Its true!
LSD probabilistic. So it is all about randomness and uncertainty. It tries to reconcile specific known types of uncertainty (not Black Swan events). Failures don't occur (to my knowledge) as a result of deficiencies in Limit States Design. So one might conclude that it is doing its job with overwhelming success! But that is exactly what we DONT want. I am arguing that it is not working exactly because it is doing its job.
If I know that a 250UB25.4 works (100% utilised) as 300MPa steel. The steel is coming from a reputable manufacturer - likely 99% assured. And the steel is in use for an office building ( most offices are typically 0.5 kPa ) with a design load of 3.0 kPa (I have no idea why this is so high - remember people aren't heavy, the other stuff is heavy apparently). And I know the dead load because I can measure it on site. Why do I need to factor the material down by 0.9 and the live load up by 1.5 (resultant 1.67)? My solution is going to have 67% more embodied carbon than I believe is truly necessary (assuming strength governs - I am happy to convince people a bit of movement is tolerable). This seems crazy and ridiculous and, well, unsustainable.
I don’t think we are teaching the skill of problem solving, not just in engineering but across the board. We need to focus on the “why” not rote learning. We are teaching as if there is one right answer. Have we lost our curiosity?
I don’t think we are teaching the skill of problem solving, not just in engineering but across the board. We need to focus on the “why” not rote learning. We are teaching as if there is one right answer. Have we lost our curiosity?
The challenge is the asymmetry of structural design. You need to get away from failure, because flirting with it is dangerous.
