Characterization of plastic shrinkage cracking in fiber reinforced concrete using image analysis and a modified Weibull function

Numerous strategies have been advocated to reduce the potential for plastic shrinkage cracking in concrete through mixture proportioning, curing methods, or the use of fiber reinforcement. The effectiveness of each approach must be adequately quantified to determine whether the additional initial cost of each strategy is justified. The majority of current research to characterize plastic shrinkage cracking in concrete relies on manual crack observation and measurement that is typically only performed at select locations. These manual measurements of crack width may provide only limited information and may be subject to operator bias. This paper describes a systematic methodology that uses a semi-automated image analysis approach to accurately quantify the salient features of the plastic shrinkage crack patterns. A restrained slab specimen with a stress riser was used to amplify the potential for plastic shrinkage cracking. The specimens described in this paper were exposed to an accelerated drying environment for the first 6 hours after casting. At an age of 24 hours, the crack was assessed using a series of thirteen gray level images that were acquired along the crack path. Crack contours were extracted from these images using gray level intensity thresholding to create binary images. Subtracting the binary image from a uniformly spaced grid enabled the crack width to be determined at numerous locations. This image analysis technique permits the crack widths to be assessed rapidly at numerous locations without operator bias. The measured crack widths were statistically analyzed using a modified Weibull distribution function. This enabled the crack width distribution to be fully described using only two or three parameters. It is anticipated that this approach can be used to better quantify the effects of short randomly distributed fiber reinforcement on plastic shrinkage crack formation.