07 June 2012
| General, TUD, TD
Robustness – or what golf, soup and structures have in common
COST Action TU0601 provides a philosophical framework for structural robustness.
Robustness is a technical term indicating a system’s sensitivity with respect to changes which we cannot fully control. Let’s draw a comparison with our daily lives: when investing in a pension plan we do not tend to put all ‘eggs into only one basket’ but distribute the investment into at least a couple of different economic activities. This is to avoid unpredictable market developments that may lead to the loss of our pension or, even worse, bankruptcy. This example illustrates how we often make decisions taking into account the unexpected or the undesired. We act according to plan A, and our grand plan actually contains a plan B that will help us should plan A fail. This is how we build robustness into our daily actions.
The same applies when we build structures. In principle, structural systems can be understood as an ensemble of components which are connected by joints into a certain shape with a certain stiffness and strength – as required to fulfil the purpose of the structure.
The structure is ‘good’ if it is reliable in fulfilling its function and if it is safe for the people exposed to it. The structure ‘fails’ if one or more of the components or joints fail during their service life. The process of planning how to build a structure is what is usually called ‘structural design’ – where the engineer is the designer.
The worst type of failure is when the entire structural system fails. Luckily this type of failure is rare... In recent history, the collapse of the Ronan Point, a 22-story tower block in east London, in 1968, was a real eye-opener. Later, shocking examples include the 9/11 collapse of the World Trade Center in New York in 2001.
The design of structures relies on the principles of physics and mathematics, and responds to equations. To use another real life example… compare this to the task of preparing a soup and let the cook be the designer. As a cook, we do not approach this task with the periodic table at hand. Instead, we rely on a recipe or a cook book describing how to mix and cook certain amounts of ingredients, in which order, for how long and at which temperature. By using such a simplified approach, we do not have to worry about reaction equations – which greatly enhances efficiency in the kitchen. Even better, we get a standardised result – a soup whose quality depends on the ingredients and the accuracy we used when following the recipe.
Similar considerations can be made for the design of structures. In practice, structures are comparable in terms of construction and purpose and may therefore be designed in a simplified manner, using ‘recipes’ which are collected in a book called a design code.
Design codes aim to ensure that the safety and the reliability of structural systems are sufficient and adequate by focusing on the performance of the structure from a component perspective: this approach could be plan A. Following this plan, it is ensured that the structures perform satisfactorily with respect to the individual components and joints. However, the possible failure of individual components affecting structural performance – which may possibly result in the collapse of the entire structural system – is not fully accounted for… plan B is missing.
Golfers will compare this approach to planning a shot without accounting for the wind and the unevenness of the green. This linear way of dealing with a non-linear problem will not always be a successful strategy… and the ball may hit the bunker.
Back to structures. If built according to design codes, three principal effects may have negative consequences on their performance: the validity of underlying assumptions; the quality exercised in following the design code and the freedom given to the designer in using the design code.
These three effects represent non-linear problems which may or may not have serious consequences on the performance of the structures… and they require a plan B. If plan A fails and one or more structural components fail due to some unforeseen event like a terrorist attack or a construction error then plan B must automatically set in to limit the consequences for the rest of the structural system. If plan B is successful the structure is robust.
COST Action TU0601 ‘Robustness of Structures’ aims to improve robustness so that it becomes less likely that unforeseen and undesired events such as failures of structural components can lead to failures of entire structural systems. The Action has been facing many challenges. First of all, there is no consensus, even among professionals, about what robustness really means. Secondly - and also due to the lack of consensus about the definition - there does not exist a commonly agreed approach on how to assess structural robustness. Most engineers have a good idea about what contributes to the robustness of structures – but do not have a clear answer for the big question - whether the robustness of a given structure is sufficient.
COST Action TU0601 took the hypothesis that structural robustness must be modelled and embraced a risk-based approach. The risks associated with failure events only involving individual components are considered the direct risks. The risks associated with failure events leading to full collapse of the structural system are considered the indirect risks. The robustness of a structure is then assessed as the ratio between the indirect risk and the total risk i.e. the sum of the direct and the indirect risks. This index shows to what degree the indirect risk is important for a given structure and points to how the robustness might be improved – by design measures reducing the indirect risks.
This Action has not solved all problems relating to structural robustness, but has indeed provided a philosophical framework for its assessment. Moreover, COST Action TU0601 has proposed and developed various model components required for assessing the direct and indirect risks for structural systems. The substantial written documentation resulting from this Action, including a document describing the theoretical framework, a wide range of papers addressing the various model components and, finally, a guideline for practicing engineers on how to improve robustness of structures will act as an important stepping stone in achieving robust structures for the future. And who knows, it may even improve soups and golf swings.