Concrete is a quasi-brittle material with a material morphology that is heterogeneous across multiple scales. Failure of concrete due to external loads is characterized by distributed microcracking followed by damage localization eventually leading to loss of structural integrity. Distributed microcracking is a failure precursor whose identification and characterization can be used in conjunction with structural health monitoring using diffuse ultrasonic waves (CODA waves) to develop an early warning system for preventing sudden failure of concrete structures. To this end, a multiscale model for concrete considering crack closure and microcracking, is proposed. The mesostructure of concrete is generated using Random Sequential Adsorption (RSA) of aggregates based on concave slicing of randomly cut polyhedrons. While the coarse aggregates are spatially resolved, microcracking in the mortar matrix is modelled using continuum micromechanics and Linear Elastic Fracture Mechanics (LEFM). The effective stiffness of the synthetic concrete microstructure is computed using a combination of numerical and analytical homogenization schemes. The high-resolution mesoscale model for concrete is reduced using K-means clustering to improve the computational efficiency. Implication of the influence of the microcrack morphology at the microscale on the diffuse microcracking and overall concrete stiffness degradation is presented.