The Statistical Properties of ( Asphalt ) Mixtures December 2016

The Statistical Properties of (Asphalt) Mixtures December 2016 Saeed Maghsoodloo and Randy West National Center for Asphalt Technology (NCAT), Auburn University, AL 36849 Abstract In the design of asphalt paving mixtures, aggregates from n stockpiles (generally n = 2, 3, 4, ..., 10) are blended at selected proportions to satisfy specified gradation ranges and other criteria. The mean and variance of any characteristic (such as %-passing through a given sieve) of each aggregate is generally known. In a laboratory environment, the proportion wi (i = 2, 3, ..., 10) from each stockpile is controlled by the mix designer so that the n proportions (or weights) can be assumed fixed. This is unlike a plant operation where aggregates are fed from n bins into a mixing-drum so that the proportion of each aggregate, Wi, from each bin is a random variable. The statistical analyses of the former case is well-know and is first repeated herein in sections 1 and 2, while less accurate information is available about the latter. So, the main objective of this article is to provide more accurate statistical information for the case when aggregate proportions are random variables.