11 Nov 2011

Engineering a safe existence

UPDATED 12/11/11 - Links checked etc.

I have decided to dedicate this weeks post to the unfortunate victims of the earthquakes which have plagued Turkey recently. My initial response as an Engineer is that some of these losses were avoidable. Massive condolences to those more personally affected than I. A gut wrenching loss of life.

"earthquakes don't kill people, buildings do

There are organisations such as 'Build Change' who are proactive in their approach and will save many lives due to their efforts. A completely awesome engineering resource and charity [cap doffing]. For this post, I do not wish to debate whether there are adequate amounts of volunteered expertise ready and waiting to help mitigate these kinds of problems. For example, the ones faced by developed countries sitting on a convergence of destructive fault lines. Like Turkey.

Neither am I planning to dive into technical requirements for buildings in earthquake zones, or teaching locals how to build and engineer a safer home. However. I would like to put forward a suggestion as to why we engineers are the way we are. I hope that this will help you understand the thoughts and processes which have contributed to an engineers 'misconceived' conservativeness.

As Structural Engineers, we [predominantly] exist in a state of knowledge limbo. We aim to utilise our experience and carry out our designs as economically as possible, but what does that really mean? It means that we attempt to be as confident as we can, and by aiming at a 90-95% utilisation of the strength of the materials that we design with. This figure does not include for the mandatory factors of safety which we are obligated to adhere to either. These essential factors of safety average out another 30-35% of under utilisation due to inconsistencies in material strength, and inaccuracies in construction and loading assessments. If we consider that their ultimate load bearing capacities are only using 60% of their strength... then this results in an over design of approximately two thirds! Guilty as charged. In our defence though, the above calculations are rather more complicated in reality and I have neglected to mention permissible stresses, load combinations and various clauses and sub-clauses, all of which are used to help us economise on design. Trust me, the above statement has been extensively rationalised so that I could attempt to come up with figure for you.

No wonder we are viewed with suspicion when from time to time we are able to 'magic' up additional capacity where once there was none, using nothing but our Engineering judgement.

I was out running a few weeks ago, and I run through quite a busy golf course. At times I have had to take refuge behind trees or protect my head as I sprinted past a golfers tee. This one time, I ran over the crown of some dense grassland and happened upon an excellent view of a 90 yard par 3 hole. This part of the course looked like it had been designed as a training hole, to allow golfers to hit an entire bucket of balls undisturbed.

The green was peppered with tiny white balls. They were also everywhere else too! On the fringe, in the rough, in the bunker, right next to the flag... and my immediate thought was, WOW. This guy is destined to do well with all that practice. He will try and try and try, and eventually he will stumble upon that perfect swing, which will then go on to become his most consistent. For golfers, the amount of practice is directly proportionate to their level of consistency and success... as with most things in life.

Not in engineering though.

Mistakes, failures and testing costs time and money, sometimes it costs lives. When we step up to the tee, as an engineer - we only have one ball, one hole. So we have to settle ourselves that our judgement is correct. Safety takes presidence.

OK, so we test in laboratories, but these are laboratories - and not building sites We also proverbially increase the diameter of the hole somewhat, because we are better equipped to break down and therefore understand the problems facing us. Most of the 'edge' which we have comes from the pervasive application of factors of safety... driven by the fact that the repercussions of our failure are horrendous.

I hope that I have given you a better idea as to why we are keen to maintain control of what is carried out during construction, and are infact best placed to aide project managers and contractors build with more efficiency and confidence than anyone other single member of the design team.

We have pride, safety, connectivity, and are educated, motivated, hard working, creative.... and altruistic. Every construction project needs an Engineer on board*.


Engine[er]

*I would however draw a line at asking an Engineer to say design your garden decking or specify the foundations of a dwarf garden wall which passes near tree roots. You may not like the answers that you get. Certain small building projects perhaps need the touch of an experienced builder rather than an Engineer who invokes the protection from their indemnity insurances...

4 comments:

  1. Thank you just for this great blog post. It absolutely was extensively informative and to the point. I am delighted I found this website through google.

    ReplyDelete
  2. Hi Devis! Very kind words thank you! Follow us on Twitter and spread the word if you wish. I tend to upload my weekly blog post every Friday. Sometimes more often if I have been busy. If you have any other comments on my other posts then please feel free to do so!

    ReplyDelete
  3. Over simplified and misleading. One of the problems of permissible/allowable stress design which hides variance, probability and reliability studies behind the scenes. Limit state design, attempts to bring this to the fore front of the designers mind. There is no reserve: for there to be reserve we would have to know the loads and resistances exactly: we don't. Manufacturing and construction always involve variability away from a nominal or mean value. Quality Assurance (QA), and statistical process control (SPC) aims to control the standard deviation, keeping variability tightly centred on the mean. The strength of materials varies, and the dimensions of structural sections and assemblies vary. The building code of Australia (BCA) requires we use 5th percentile characteristic strengths. That is there is a 5% probability that the strength will turn out to be less than what we use in design. Allowing for variability in dimensions when calculating areas, section modulus etc.. is complicated so capacity reduction factors are used to simplify allowance for such variability. So that when we have calculated the strength of a member (beam/column) from the material strength and section properties, its resistance is hopefully the 5th percentile characteristic value.

    Then there is the applied load, once again this is uncertain especially if due to earthquake, or hurricane. If not otherwise defined, then the 95th percentile load is used: that is a load which has a 5% probability of being exceeded during the lifetime of the structure. So the result is a load with 5% probaility of exceedance and a member with 5% probability of not being strong enough: no reserve, just a low probability of the structure being under strength and over loaded at the same time.

    Cannot take the permissible stress equations, for example 1/0.6=1.67, leads to 67% stronger than needs to be: because we absolutely do not know how strong it needs to be. The design load chosen always has the probability of being exceeded.

    The latest and greatest earthquake design failed at Kobe, then at North Ridge. The real issue is not keep magnifying the design load, but controlling the mode of failure when the design load is eventually exceeded. Choice between concrete apartment block falling on you versus felt blanket of a Mongolian yurt. It is not quality robust design to specify reinforced concrete buildings in regions where steel is scarce: no matter how much inspection its not going to be installed.

    If you really like mathematics then check the joint committee for structural safety (jcss):

    http://www.jcss.byg.dtu.dk/Publications/Probabilistic_Model_Code.aspx

    And engineering of a deck, depends how high it is and whether call a balcony:

    http://www.brisbanetimes.com.au/articles/2008/11/20/1226770617762.html

    Its not the size of the structure that is important but the risk and associated hazard to life, and economics of replacement or consequences of loss.

    ReplyDelete
  4. Great comments Conrad! Really enjoyed reading this!

    I do agree with what you have to say, but this post was meant to be overly simplified. I am trying to interest and teach non-engineering types about are crazy existence as well as show graduates that there is a lot more to our job than statistical analysis. We as engineers very rarely step into the pure mathematics which we were trained in during uni. Most of our engineering qualities are taken up with appreciating the 'bigger' picture and applying it. I'm careful not to cross over to the idea that we infact all just technicians...

    ReplyDelete

Starting up an Engine[er]

Starting up an Engine[er]
Click here to go to the all NEW blog site!