A comparative study of two correlation functions applying to the quantitative characterization of the pore structure within hardened concrete

Abstract

The paper presents a comparative study between the autocorrelation function (ACF) and the 2-point correlation function (2PCF) for the quantitative analysis of the pore structure within hardened concrete via the image analysis technique. This work was carried out on five cross-sectional images of concrete specimens with varying area porosities ranging from 2.6% to 21.2%. This research indicates that: (i) the ACF provides a comprehensive and effective tool that is capable of capturing the contribution of the area porosity, size, and orientation of the pores in quantifying the pore structure within hardened concrete, while the 2PCF only enables quantifying the size and spacing between the pores; (ii) the ACF requires the cross-sectional images to be square, whereas the 2PCF can analyze images of any size and shape; (iii) for an area porosity below 20%, the characteristic size of the pore structure within hardened concrete, determined by both the ACF and 2PCF approaches, is not significantly different; (iv) on average, the correlation length estimated by 2PCF is more than 1.6 times greater than the corresponding value determined by ACF; and (v) the characteristic size provides a more accurate and comprehensive quantification of the pore structure within hardened concrete specimens than the correlation length.