Performance Evaluation of Automated Machines for Measuring Gradation of Aggregates

Several automated devices are commercially available for measuring the gradation of stone aggregates. These computerized machines, which provide a rapid alternative to manual sieving, capture and process two-dimensional digital images of aggregate particles to determine grain size distribution. Five of these automated gradation devices were evaluated for accuracy using fifteen aggregate test samples. To quantify how well the machine results compare with data from standard sieve analyses, the CANWE (Cumulative And Normalized Weighted Error) statistic was developed and used. While the machine data did not match the sieve data exactly, the evaluated devices were found to provide good measures of particle gradation for most samples. These tests also indicate that some machines will give more repeatable results in multiple tests of a given material, while others yield better results when testing different materials. The methodology used in this study is suitable for objectively evaluating the accuracy of other rapid gradation machines for various applications.