This study presents a series of sensitivity analysis on parameters of lab-scale bonded particle model using Plackett-Burman experimental design method. Sensitiveness of model parameters on the several selected macroscopic responses of a particle assembly model are quantified, e.g. uniaxial compressive strength, Young's modulus, Poisson's ratio, ratio of compressive to tensile strength, and friction angle. Results indicate that among the 10 model parameters, parallel-bond radius multiplier which is related to – in a petrophysical sense – cementation between rock grains exhibits significantly positive effect on strength and Young's modulus, whereas negative effect on Poisson's ratio and the ratio of compressive to tensile strength. Results presented provide information on how one could calibrate the model parameter with minimum number of trial. Furthermore, results are to be later used in optimization of the model parameters such as presented in the earlier study by Yoon (2007).
This study presents a series of sensitivity analysis on parameters of lab-scale bonded particle model using Plackett-Burman experimental design method. Similar approach has already been developed in an earlier study (Yoon 2007). The present study applies similar approach to a more complicated contact model for applications with PFC modelling. The parallel-bond model has been, up to now, known as a better contact model for simulation of rock behaviors. Its distinct advantage over the contact-bond model investigated in the earlier study is that cementation effect between the particles is considered, which better mimics the physical rock structures where grains are cemented by the diagenesis. Simulations of rock behaviors by Particle Flow Code 2D and 3D (Potyondy & Cundall 2004) with parallel-bond model exhibited better match to soft and sedimentary rocks (Jong 2005) and crystalline hard rock like granite (Park & Song 2009; Potyondy & Cundall 2004; Cho et al. 2007) compared to the contact-bond model.