Application of experimental design and optimization to PFC model calibration in uniaxial compression simulation

Jeoungseok Yoon

DOI: https://doi.org/10.1016/j.ijrmms.2007.01.004

IF: 6.849

2007-09-01

International Journal of Rock Mechanics and Mining Sciences

Abstract:A new approach has been devised for calibrating contact-bonded particle models using ‘experimental design’ and ‘optimization’ in uniaxial compression simulation. These are applied to calculate an optimum set of microparameters used in generation of models to be tested in uniaxial compression simulations. Sensitivities of microparameters with respect to uniaxial compressive strength (UCS), Young's modulus, and Poisson's ratio (i.e. macroscopic responses of model) were tested by the Plackett–Burman (PB) design method. Then, for each macroscopic response, two microparameters having the largest impacts were singled out and their non-linear relations to macroscopic responses were estimated by statistical Central Composite Design (CCD) method. Using the results from PB design and CCD method, the problem of finding a set of microparameters was solved. Using an optimization method, the optimum set is obtained that gives the best agreement either in quantitative or in qualitative ways between the results both from the bonded particle model simulations and laboratory. The overall procedure was applied to calculate optimum sets of microparameters for generation of bonded particle models for uniaxial compression simulations on different rock types. Results from both simulations and laboratory tests gave fair agreements. The method currently provides adequate solutions and shows relatively fair applicability to simulation of rock materials with their physical properties falling within the following ranges: UCS (40–170MPa), Young's modulus (20–50GPa), and Poisson's ratio (0.19–0.25).

engineering, geological,mining & mineral processing