Mechanical testing in the structures lab is probably the most reliable method to fully understand how materials and loading interact in different types of structures such as plates, beams or columns. Physical and theoretical models are very useful towards understanding the basic behaviour, although when dealing with materials such as steel, concrete or glass there is always room for surprises. For example, one might find when testing two identical beams that one beam can resist for example 30% higher load than the other one. Or for instance the failure load of one reinforced concrete beam with high concrete compressive strength can be lower than an identical beam with slightly lower concrete strength. What causes these unexpected results?
There are many parameters which might contribute to these uncertainties in the test results such as scale effects, loading conditions, confinement effects or creep effects of the material. Moreover, the behaviour of a structural member is in many cases governed by crack development. This makes interpretation of test results very challenging since cracks have a random nature which could explain the scatter generally observed in test data. Impact load testing brings new challenges such as instrumentation and capturing failure which takes place within milliseconds. However, tests with sustained loads (i.e. very slow loading rates) are also challenging as the material properties change over time with different levels of loading. All of these structural effects take place in bridges, buildings and infrastructure when they are subjected to different traffic and permanent loads.