15th ICSMFE: Workshop #7 organised by TC2 Centrifuge Testing Some Achievements and Challenges Malcolm Bolton Cambridge University
What are the “unique selling points” of centrifuge testing? Cheap replication of the response of geo-structures Weight-induced stresses; gradients and rotations Coupling of intergranular and pore fluid pressures Access to severe distortions, drainage of clay, 3D Realistic boundary conditions; –free surface, cracks –soil-structure compatibility –seismic shaking
Some achievements: actuators CITY University tunnelling kit TIT excavator NUS compaction sand piler UWA-COFS drag anchor deployment TUD wave generator CCORE static liquefaction actuator UCD earthquake simulator CU-SC atmospheric chamber
CITY U. tunelling swinging platform motor slip rings Acutronic 661 Centrifuge model container perspex window camera lights swinging platform swing hangers tunnel clay sample with targets on front surface
(a) left camera (b) centre camera (c) right camera 3D Surface settlements from digital image analysis
TIT excavator
(a) Casing and sand hopper. (a) SCP model on centrifuge. Figure 16a – b. In-flight SCP installer. NUS sand piler
UWA - COFS drag anchor deployment
TUD wave generator
Clay Embankment Liquefaction Event Geometry and Instrumentation Settlement Gauge Piezometer DAS BBW Sand Deposit BAW Sand Deposit Tilt Sections Compacted Sand Contained Sand CANLEX CCORE static flowslide actuator
After 2 Shakes Before UCD earthquake shaker
Some challenges Replicating the mechanics of processes Broadband data stream to characterise soil response Realistic soil profiles - grading, layering, OCR etc Assessment of size and rate effects The atmospheric boundary condition
Atmospheric loading of embankments: boundary conditions
Atmospheric loading of embankments: pore water pressures
Atmospheric loading of embankments: cyclical deformations
Modes I and II fracture mechanics
PIV reveals slope deformation mechanisms And remember - this was not “collapse”
Rubblisation
Discussion point We must not make complex models just for fun Our task is to reproduce essential complexity And then observe events intensively Ground layering is sometimes essential Anisotropy may be essential, and is automatic Construction process is essential In clays, the atmospheric boundary may be essential We need to make simple mechanistic models Numerical simulation would be useful but lags behind, and can not therefore be considered essential…..