Numerical modelling of rockfall evolution in hard rock slopes

Abstract

The aim of this thesis has been to model small rockfalls in order to better understand where, when and why they occur. High-resolution monitoring of rock slopes has revealed clustering of rockfalls through space and time, suggesting interactions, whereby one detachment from a rock slope influences the nature of those that follow. This observation contrasts with the more conventional idea of rockfalls as time-independent, discrete events that occur in response to an identifiable trigger. As the processes that give rise to observations of rockfall clustering are not well established, this thesis takes the opportunity to bring together current understanding of the controls on rockfalls with ideas around the progressive development of failure in brittle rock in an attempt to explain these patterns. The representation of these processes at scales comparable to high resolution field monitoring has not previously been attempted. Therefore this thesis has developed an approach using numerical modelling to simulate rockfalls as spatially and temporally-dependent sequences of events, to better explain the underlying mechanisms.
This study begins with the analysis of a high-resolution inventory of rockfalls, concentrating on identifying patterns in rockfall occurrence. Analyses of this data reveals patterns of rockfalls that cannot be explained by environmental conditions or local geology alone. Evidence has been collected that demonstrates that rockfalls cluster in space and time, and that through time rockfall scars grow upward and outward in a consistent manner. The results of this analysis are used to inform numerical modelling that explores the mechanics driving small rockfalls, focussing upon the impact of a detachment on the location and timing of future rockfalls. Numerical modelling of idealised rock slope sections was conducted using Slope Model and demonstrated that the timing and location of failure in a rock slope could be considered as a function of accumulated damage, represented by fracture. The results suggest that time-dependent failure and associated mechanisms of stress redistribution and damage generation are one possible explanation for the propagating sequences of contiguous failures observed.
Finally, this thesis has taken an exploratory approach to modelling rockfalls through the development of a new deterministic, numerical model that simulates rockfall evolution using cellular automata. This rockfall model allows the patterns and associated underlying mechanics of small rockfalls to be explored in detail using a reduced complexity approach. Critically rockfalls are modelled in a 2.5D slope face perspective to allow both rockfalls and their effects to interact across the rock slope through time. The model operates at a relatively high spatial and temporal resolution to consider the full range of rockfall characteristics that have been observed. The outputs of the model are compared with the two-year monitoring data to address key questions regarding the competing roles of endo- and exogenic forcing on rockfall occurrence. The results of the rockfall modelling shows that a consideration of stress redistribution from small scale rockfalls and time-dependent weakening provides a possible explanation for the size distribution of rockfalls, their location and timing, and the resulting changes to slope profile form as observed in the field. This has implications for how rock slopes are monitored and modelled to determine the potential for future rockfalls to occur.