Abstract: Three CNAS-accredited laboratories were invited to determine the grain size G by using the ASTM E112 standard for 14 identical grain size pictures that were taken on the same aluminum alloy (AlSi1) sample, and the results were quantitatively evaluated. The results from three laboratories show that the statistical counts of intersections or grains are the main parameters for determining G and it is the key point of quality control of the effectiveness of determining G. By using the Intercept Procedure, the range of the intersection counts is 3 to 11, the relative standard deviation is 2.5% to 10.1%, and the range of G is 0.09 to 0.53. While by the Planimetric Procedure, the range of grain counts is 4.0 to 16.5, the relative standard deviation is 2.9% to 15.7%, and the range of G is 0.06 to 0.36. The results also show that laboratory - to - laboratory comparison activities should not focus only on the final results, but also on the comprehensive evaluation of the original observation results, intermediate results, and final results, so as to obtain opportunities for improvement and control the testing risk.

Key words: grain size, ASTM E112, interlaboratory comparison, aluminum alloy (AlSi1)