Quantitative aspects of grain size measurement

Abstract

Grain size measurement is directly dependent on the ability of the microstructure to be revealed in a form that is representative of the material. A single phase, equiaxed ferritic steel was used throughout the present investigative work, this material being chosen because of the apparent simplicity of the microstructure. The lineal intercept, circular intercept, and planimetric measurement techniques were used. All the results are reported using the ASTM grain size number, G. Two aspects of grain size measurement are reported in the present paper. The first is the impact of missing boundaries on grain size measurements. The etching techniques established within industry to reveal microstructures often only partially reveal grain boundaries. An experiment is reported where the impact of missing grain boundaries on grain size measurements is assessed and hence the importance of revealing all grain boundaries is determined. An image analysis system was used to completely reconstruct the microstructure in a binary form, then to remove a known percentage of the boundaries, followed by measuring the grain size using the different techniques. The selection of the boundaries to be removed was done randomly to allow for any bias. The results reported show that, even with up to 20% missing boundaries, the impact on the grain size measurement was not significant, giving a difference of ∼0.5 grain size units. Sampling is the second factor studied. In order for measurements to be representative the number of grains within a field of view from each specimen, the number of fields of view per specimen, and the number of specimens have to be considered. From an analysis of the results of the characterisation of the ferritic steel it was clear that the number of specimens used for measurement was the most important factor regarding microstructural representation.