Download files
Abstract
In this thesis, an experimental and numerical study on the hydraulic fracturing
propagation is conducted using the 3D printing techniques and the lattice modelling.
The first half of the thesis presents the experimental studies. Firstly, the potential of 3D
printing on rock mechanics is explored by Polylactic acid (PLA) material, but its plastic
property inhibits the direct simulation of rocks. Then, more brittle powder-based
gypsum-like 3D printed specimens are successfully fractured by the split Hopkinson
pressure bar (SHPB) system for investigating the dynamic crack propagation and
coalescence. In addition to the high accuracy and flexibility in specimen preparation, the
desired fracture paths captured by the high-speed camera affirm the suitability of the
powder-based material on fracturing test. Finally, this material post-processed with
cement curing agent was applied in hydraulic fracturing tests. The desirable brittle
fracture patterns driven by the fluid pressure were obtained.
In the second half of the thesis, numerical studies are conducted by lattice modelling. A
fundamental study on the crack propagation by the distinct lattice spring model (DLSM)
was carried out. The relationship on the fracture criterion between the DLSM and linear
elastic fracture mechanics is investigated for the first time. The work involves the
correlation between the stress intensity factor (SIF) and spring deformation, the
influence of the particle size on fracture toughness, and the relationship between the
micro-spring failure and critical SIF. Following this, a damage-plasticity model was
developed and implemented into DLSM to enhance the predictability on dynamic
fracturing. The new constitutive model is capable of predicting the initial and peak fracture toughness, crack speed, and crack patterns. Lastly, a hydro-mechanical model
by coupling the DLSM and the lattice Boltzmann method (LBM) was proposed to
simulate hydraulic fracturing propagation. A direct coupling algorithm is applied with
parallelization of both DLSM and LBM. The coupled model is validated through a
series of benchmark problems. The hydraulic fracturing results predicted by the model
are well matched to the experimental observations, and the model is promising on
simulations in complex geological conditions.