Variation in Calculated Pavement Performance Due to Variation in Input Data

Deterioration models are largely dependent on the input data. Data is collected with a varying degree of accuracy, and different variables affect the pavement deterioration to varying degrees. The effect of variation from eleven input variables, with their distribution functions, has been tested in two deterioration models. This sensitivity analysis consists of individual variation and simultaneous variation of input data. The analysis is made with two deterioration models, HDM-4 and MMOPP with two different pavement types. HDM-4 is generally less sensitive to changes in input data than MMOPP. Generally, layer properties, such as thickness and modulus, have the greatest influence on service life in both models. By concentrating the effort on determining those variables that affect pavement performance the most with the highest accuracy, the precision in calculated pavement performance may probably increase considerably. The paper is based on a licentiate thesis (Ekdahl, 1999). For the covering abstract see ITRD E118503.