A sensitivity study of parameters used in shrinkage and creep prediction models

Abstract

Over the last 15 years, there have been numerous models put forward for the prediction of the time-dependent behaviour of concrete and consequently concrete structures. The development of these models from a number of different sources has meant that the engineer now has a choice to make when undertaking creep and shrinkage prediction and it is difficult to know which is best suited for this purpose. Some are considerably more complex than others requiring greater amounts of input focusing on material data, concrete properties, environmental conditions, specimen geometry and loading conditions. The study reported here focuses on the sensitivity to change, in both the short term and over time, of the individual parameters that comprise the input and whether one model is more appropriate in any given situation. It has been found that for each model certain parameters are more sensitive than others. It is suggested that for any given model, the parameters that do not reflect the expected behaviour when changed counterbalance each other, cancelling out any errors. This further suggests that when deciding on which model to use when predicting shrinkage and creep strains it is prudent to look at the specific conditions that prevail, assess the relevant input parameters for which data are available, assess the sensitivity level of each of these parameters and then make a decision as to the most appropriate model to use.