Probability distribution functions for cover used in 3-D model simulating concrete deterioration in port assets

In previous studies, a 1-D numerical predictive tool to simulate the salt induced corrosion of port assets in Australia has been developed into a 2-D and 3-D model based on current predictive probabilistic models. These studies use a probability distribution function based on the mean and standard deviation of the parameters for a structure incorporating surface chloride concentration, diffusion coefficient and cover. In this paper, this previous work is extended through an investigation of the distribution of actual cover by specified cover, element type and method of construction. Significant differences are found for the measured cover within structures, by method of construction, element type and specified cover. The data are not normally distributed and extreme values, usually low, are found in a number of locations. Elements cast insitu are less likely to meet the specified cover and the measured cover is more dispersed than those in elements which are precast. Individual probability distribution functions are available and are tested against the original function. Methods of combining results so that one distribution is available for a structure are formulated and evaluated. The ability to utilise the model for structures where no measurement have been taken is achieved by transposing results based on the specified cover.