University of Twente Student Theses


Applicability of the semi-structured mass movement mechanism for rockfall runout modelling

Dhakal, Om Prasad (2021) Applicability of the semi-structured mass movement mechanism for rockfall runout modelling.

[img] PDF
Link to full-text:
(only accessible for UT students and staff)
Abstract:Rockfall and avalanche processes of various scales are a major natural hazard in mountainous areas throughout the world. There is a lack of clear methodology that could approach both simulating a small scale rockfall and large scale rock avalanche. The requirement of this single method is crucial when incorporating rockfall phenomena into a multi-hazard modelling framework. During such events, flow and fall processes of various scales occur and require a generalized model to capture complex interactions and transitions during runout. This research looks at the functionality of the newly developed semi-structured mass movement model by applying it to three case study sites. The selected case study sites provide enough variability when it comes to the runout extent, initiation volumes, and slope morphology. The models in use are Rockyfor3D to represent trajectory rockfall modelling. This industry-standard tool for rockfall dynamics was calibrated for two of the three study sites. OpenLISEM hazard 1.0 and OpenLISEM hazard 2.0 represent granular flow mechanism and structured mass flow mechanism, respectively. The OpenLISEM models are compared to the reference data from Rockyfor3D in order to assess the accuracy and applicability of generalized semi-structured mass movement models to rockfall and rock avalanche dynamics. Using a gradient descent algorithm, extensive calibration and sensitivity analysis was carried out. Simulation accuracies were calculated in terms of kinetic energies and impact pressure for each model compared to the Cohens Kappa inter reliability index. The accuracy calculated in reference to the validated model gave Kappa value of 0.66, 0.47 and 0.57 for the study case of Acheron, Andorra and Barcelonnette, respectively. Since the spatial accuracy was not enough to delineate the model's applicability, analysis based on the simulated velocities, impact to elements at risk and fragmentation behaviour of each model were carried out. The results showed at least a two-fold underestimation of velocities by the OpenLISEM hazard 2.0 model compared to the RF3D when the runout slope is greater than 40 degrees. Further, the estimation based on elements at risk showed at least a three-fold underestimation of kinetic energy values by the OpenLISEM hazard 2.0 model. On the contrary, the OpenLISEM hazard 2.0 model accurately simulates the fragmentation behaviour. Furthermore, the non-diffusive transport of the solid and breakage based on the stress-strain relationship led to realistic results compared to a field inventory. In conclusion, the research suggests further research and improvement of the OpenLISEM hazard 2.0 model. The design and planning for rockfall phenomena based on the model output are not suggested within the current framework. The part for improvement would be 1) modelling of the free fall, tumbling and rolling phenomena. 2) non averaged frictional forces. 3) Third impact against trees.
Item Type:Essay (Master)
Faculty:ITC: Faculty of Geo-information Science and Earth Observation
Programme:Geoinformation Science and Earth Observation MSc (75014)
Link to this item:
Export this item as:BibTeX
HTML Citation
Reference Manager


Repository Staff Only: item control page