This dissertation presents an integrated design process for paste backfill based on three inter-related engineering approaches;​ centrifuge modeling, analytical modeling and numerical modeling. This strategy, developed to (1) establish free-standing heights for different paste fill recipes and stope conditions, and (2) assess sillmat stability, can be regarded as the main contribution to paste backfill support systems design. This work demonstrates the effectiveness of centrifuge modeling as a primary tool used to dynamically test paste fill performance on a time-dependent and three-dimensional basis. The methodology incorporates various conditions such as stope geometry, stope wall roughness, stope wall closure, fill reinforcement and surcharge loading for paste fill free-standing heights and sillmat behaviour and stability performance. The effect of binder agents and curing time on paste fill stability is also presented.

As in rock engineering, an integrated approach is the only practical method to rigorous and accurate paste backfill design. Analytical, numerical and centrifuge physical modeling are incorporated into the design process, and are used to describe or predict the behaviour of paste backfill, as a structural element that serves not only as ground support during adjacent mining, but is also utilized as sillmat elements to support excavations during overhead mining or sill pillar recovery. Improper design of these structures may cause fill mass failure, resulting in extensive economic losses associated with loss of production and ore dilution, and in safety problems. The design process, illustrated in case studies performed for establishing stable free vertical fill faces and sillmats for the Golden Giant Mine, demonstrates the effectiveness of the approach not only in describing or predicting the support performance of backfills during mining, but also in assessing arching effects, failure modes and fracture mechanics involved in fill mass failure.

Further, the design principles applied for the centrifuge facility used in the physical modeling appear to be appropriate as demonstrated by the number of centrifuge runs conducted involving models of varying payloads and gravity accelerations. As well, a stress error minimization approach developed for the centrifuge facility indicated that a model height to effective radius ratio of 1/3.07 would provide a stress error, between the model and the prototype, of less than 1%.

Finally, the applicability of the integrated design approach demonstrated in the case studies indicated the following:​

• The various approaches used in the integrated design are complementary:​ centrifuge and numerical modeling can be effectively used for assessing failure modes and failure mechanisms of paste fill support systems;​ the analytical modeling approach is useful in providing some approximate parameters for predicting the behaviour of paste fill support systems.
• Principal failure modes in paste fill exposed faces are shear failure followed by sliding block, somewhat resembling a circular slip;​ failure modes in sillmats are slip failure, sill rotation and caving. The rupture plane mechanism in sillmats occurring in centrifuge models is a function of strain softening (decrease in cohesion, friction angle and tensile strength) conditions.
• Stability improves for fill materials placed in narrow stopes, with rough wall conditions, relative to those placed in wider stopes.
• Wall closure improves stability of paste fill support systems:​ higher closure stresses are required in wider stopes to enable stability of the fill mass to develop.
• Effective shear strength effects (effective cohesion and effective friction angle parameters) increase with increases in cement content, wall roughness and stope wall closure.

It is postulated that application of the developed methodology will result in economic and effective paste backfill designs for the mining industry.