Quantitative Mesostructural Investigation Of Rock Based On Digital Image Processing Technique

Rock is a multi-scale material, which used to be researched at macroscopical scale. Because the fracture and failure of rock under loading is caused by microcracks' growing and propagating at meso-scale, it is helpful to understand the mechanical behavior of rock by quantitatively investigating mesostructure of rock.First, based on damage mechanics and stereology theory, the test of quantification of rock meso-structure is designed. Some relating issues such as specimen preparation, experiment equipment and experiment routines are settled down. The test takes scanning electron microscopy(SEM) as the observer, and thousands of meso-structural images of marble of Jinping in Sichuan province are obtained.Second, based on image processing technique, the images of meso-structural images of marble were processed by regional growing theory. The whole process is programmed by Matlab and the microcracks' meso-structural information such as length, angle, width, area, perimeter, etc are also obtained by Matlab according to stereology theory.Third, the microcracks' meso-structural information is processed by statistical theory. Based on the statistical law of the microcracks' meso-structural information, the Representative Element Volume(REV) of rock at meso-scale is decided by Monte Carlo theory.Finally, based on geometrical damage mechanics, the initial damage tensor of REV of rock at meso-scale is obtained, and then introduces it into G. Swoboda damage model to obtain the stress-strain relationships under uniaxial compress and triaxial compress conditions. The method is verified by comparing the stress-strain relationships of theory and of practical experiment.