RECENT ADVANCES IN THE APPLICATION OF VERTICAL DRAINS AND VACUUM PRELOADING IN SOFT SOIL STABILISATION Buddhima Indraratna Professor of Civil, Mining & Environmental Engineering Director, Centre for Geomechanics and Railway Engineering Faculty of Engineering, University of Wollongong Wollongong City, NSW 2522, Australia

Contents Introduction to PVDs and VP application Role of Smear zone (disturbed soil zone around the mandrel), its assessment and implications Effect of Vacuum Pressure propagation and variation with time (including vacuum removal &reapplication) Experimental Investigations Numerical Modelling and Case History Analysis Advances in Design and Practice Guides

Inward movement due to VP Potential Benefits of Prefabricated Vertical Drains in Soft Formation Clays Surcharge Fill Embankment Settlement Surcharge fill only – no vertical drains Due to PVDs Depth Inward movement due to VP Vertical drains with surcharge Vertical drains with surcharge and vacuum preloading Time Lateral displacement at toe

Drain anchors and Mandrel shapes Installation of PVDs Drain anchors and Mandrel shapes 4

Principles of Vacuum Consolidation Via PVDs Membrane system (e.g. Menard) Membrane less system (e.g. Beaudrain) 5

Soft Foundation Stabilisation by Vacuum Consolidation Surcharge Fill Only (Anisotropic Load) Vacuum Preloading with PVD (Isotropic Loading) Vertical Stress Vertical Stress Slip Surface No Failure Risk of Slope failure is minimized by the use of Vacuum Preloading

Site preparation for Membrane-type Vacuum Consolidation (Courtesy of Austress-Menard) Drain Installation Horizontal drain installation Peripheral bentonite trench Connection between horizontal drainage and vacuum pump Membrane installation

Site preparation for Vacuum Consolidation-Membraneless (Courtesy, CeTeau) Drain Installation Tube connection

Principle of Vacuum Consolidation Governing Equation Consolidation: (a) conventional surcharge loading; (b) idealised vacuum preloading (Indraratna et al. 2005c).

Effect of Vacuum Removal and Reloading on Consolidation After some initial consolidation, putting off the vacuum pump is not going to make the soil swell up again, but the rate of settlement is swiftly retarded. Pumps may have to be switched off from time to time to prevent over-heating. Suction in the drain (240mm from bottom); b) surface settlement surface settlement associated with simulated vacuum loading and removal (Indraratna et al. 2004).

Experimental Evaluation of PVD + VP system Large-Scale, Radial Drainage Consolidometer at Uni. of Wollongong Constant Strain Mandrel Driving Installation of PVDs by the steel mandrel causes smear around the PVD Hydraulic Loading 550mm Diameter 1.2m Height 11

Assessment of the Extent of Smear Zone (Indraratna & Redana, 1998, Sathananthan & Indraratna 2006) Drain Locations of cored specimens PVD and smear zone 12

Evaluation of Smear Effects Permeability Approach Indraratna & Redana, 1998, JGGE, ASCE Vol. 124(2) Water Content Approach Sathananthan & Indraratna 2006, JGGE, ASCE, Vol. 132(7)

Vacuum Propagation Model based on Laboratory Data Soil-drain interface Soil element Lateral Propagation of VP ks < kh If k1 =1, there is no vaccum loss with depth Vacuum pressure distribution patterns in the vertical and lateral directions (after Indraratna et al. 2005).

Cyclic Loading and Soil Consolidation via PVDs Cyclic Loading Actuator (a) Large-scale triaxial rig; (b) soil specimen (Indraratna et al. 2009a). 15

Cyclic Excess Pore Pressure Response of Soft Clay With and Without PVD (Indraratna et al., 2009) Failure of samples T3 T6 Specimens without PVD fail very quickly as the excess pore pressure rises rapidly! 16

Excess Pore Pressure is rapidly created during mandrel intrusion FEM Simulation of Mandrel-driven PVD – Pore Pressure CREATION due to very high plastic strains Excess Pore Pressure is rapidly created during mandrel intrusion Excess PWP dissipates very gradually after mandrel withdrawal in spite of the drain. Mandrel Driving INCREASES effective vertical stress, hence, the lateral permeability decreases within the smear zone (Indraratna et al. 2009, ASCE J. of Geomechanics).

Pore pressure variation during mandrel installation Locations of pore pressure transducers 18

Maintain geometric equivalence Analytical and Numerical Simulation Multi-drain Analysis and Plane Strain Conversion Field condition: Axisymmetric 2D plane strain FEM Maintain geometric equivalence Reduce the convergence time and require less computer memory Must give the same consolidation response

Conversion of an Axisymmetric Unit Cell into Plane Strain Indraratna et al., 2000 & 2005 Conversion must give the SAME time-settlement curve

No conversion (some published work) After conversion

Normalized average excess pore pressure in axisymmetric condition with vacuum (Indraratna et al., 2005), CGJ kh ks -p0 -k1p0 z l ds/2 de/2 Smear zone Undisturbed zone Vacuum pressure distribution C L = pore pressure at time t (average values) = time factor = undisturbed horizontal permeability = smear zone permeability = initial pore pressure = average applied vacuum pressure

Vacuum + Surcharge alone Degree of Consolidation: Pore pressure Based Models (Indraratna et al. 2008) Surcharge alone Vacuum alone Vacuum + Surcharge alone

Case Study 1: Port of Brisbane Ground Improvement Dredging for Reclamation fill PVD installation Marine boundary and sandy platform Vacuum stabilised area

Essential Design Aspects (selected section) Service load @ 25 kPa, Max. residual settlement @ 250 mm over 20 yrs. Sea wall and future development area Soil Properties Plan view 25

Time-Settlement and Pore Pressure Response More than doubled settlement obtained with VP at the same time scale Higher k promotes greater PWP dissipation (a) Settlement and (b) excess pore pressure for non-vacuum site (a) Settlement and (b) excess pore pressure for a typical vacuum site

Effect of OCR and clay thickness on residual settlement Reduction in lateral movement due to VP Effect of OCR and clay thickness on residual settlement Lateral Displacement reduction due to vacuum application 27

Case Study (2): Trial Embankment Stabilized with PVD and Vacuum Preloading, Ballina Bypass, Australia No vacuum Vacuum Instrumentation layout Typical soil properties

Test Embankment Cross Section -70 kPa vacuum and max. 8 m surcharge applied at this site

Performance : soil properties before and after vacuum application The void ratio, compressibility Index and water content decrease significantly in the initial 17m. Beyond that, only a marginal decrease is observed. PVD+VP system is mainly effective at the upper regions of the clay.

Performance: Lateral displacement No Vacuum Lateral movement decreases due to vacuum, even at higher fill heights. Ratio of lateral movement to fill height is a better indicator of the stability provided by VP

(Indraratna et al 2005, Int. J. of Geomechanics, ASCE, 114-124) Case Study (3): Test Embankment Stabilized with PVD and Vacuum Preloading in Soft Bangkok Clay, Thailand (Indraratna et al 2005, Int. J. of Geomechanics, ASCE, 114-124) -60 kPa design vacuum and max. 2.5 m surcharge 32

Vacuum Simulation (selected section) (Indraratna and Redana, 2000) Model A: Conventional analysis (no vacuum; only surcharge fill) Model B: Vacuum pressure is adjusted according to field measurement and reduces linearly to zero at the bottom of the drain (k1= 0) Model C: Perfect seal (vacuum pressure was kept constant at -60kPa after 40 days); vacuum pressure varies linearly to zero along the drain length (k1= 0) Model D: No vacuum loss along the drain length (k1=1) Field measurements Model C: Assumed VP 33

Surcharge Only Settlement Excess pore pressure 34

Weathered Crust is much stiffer in reality than the assumed properties Lateral Movements at Embankment Toe Weathered Crust is much stiffer in reality than the assumed properties Key Advantages: Surcharge fill height reduced from 4.0m to 2.5 m Time for 95% consolidation reduced from 12 months to 4 months. 35

3D vacuum pressure propagation across the boundaries of treated zone Effect of vacuum application (negative movements) may extend more than 10 m from the edge of the embankment Lateral movement A A A

Class A Prediction (Indraratna et al, ASCE, JGGE, 2010) Case Study 4: Railway Applications: FEM Analysis of Short PVDs at Sandgate Class A Prediction (Indraratna et al, ASCE, JGGE, 2010) Very Soft Alluvial Clay Soft Silty Clay Rapid dissipation of excess pore pressure Curtailing lateral displacement 37

CONCLUSIONS Vacuum preloading rapidly decreases excess pore pressure, and directly increases effective stress with time. Conventional surcharge models simulate the increase in total stress, and associated increase in the excess pore pressure. Smear effects adversely affect PWP dissipation, and the application of VP and corresponding increase in the hydraulic gradient partially compensates for this. PVDs in combination with vacuum and surcharge fill curtail lateral movements and provide stability for the superstructure. Excessive VP generates high inward movement causing tensile zones. Sophisticated 3-D numerical modelling is not required if appropriate conversion to 2D plane strain can be made for most sites for multi-drain analysis. Exceptions would be marine boundaries and corners. Field monitoring for VP sites is essential to ensure performance, and to establish any time-dependent variation of VP distribution.

CONCLUSIONS Vacuum preloading rapidly decreases excess pore pressure, and directly increases effective stress with time. Conventional surcharge models simulate the increase in total stress, and associated increase in the excess pore pressure. Smear effects adversely affect PWP dissipation, and the application of VP and corresponding increase in the hydraulic gradient partially compensates for this. PVDs in combination with vacuum and surcharge fill curtail lateral movements and provide stability for the superstructure. Excessive VP generates high inward movement causing tensile zones. Sophisticated 3-D numerical modelling is not required if appropriate conversion to 2D plane strain can be made for most sites for multi-drain analysis. Exceptions would be marine boundaries and corners. Field monitoring for VP sites is essential to ensure performance, and to establish any time-dependent variation of VP distribution.

Acknowledgement Australian Geomechanics Society (AGS) Dr Geng Xueyu and Dr Cholachat Rujikiatkamjorn during compiling and editing of a vast amount of data from the past 15 years of research in vertical drains and vacuum preloading conducted at University of Wollongong (UOW). More than a dozen past and present research students who have contributed to the contents of this lecture directly and indirectly. Australian Research Council – Linkage and Discovery project funding Research Collaborations with many Industry Partners and Institutions over the years: Queensland Department of Main Roads, Port of Brisbane Corporation, Roads & Traffic Authority, Coffey Geotechnics, Asian Institute of Technology, Thailand; Polyfabrics, Geofabrics, ARUP, Douglas Partners, Snowy Mountains Engineering Corporation, RailCorp, ARTC, Chemstab, Queensland Rail and Austress-Menard Technical staff, University of Wollongong

Thank You

Settlement close to the embankment centreline Practicing Engineers’ Dilemma - Disparity between Excess Pore Pressure and Settlement Indraratna, Balasubramaniam & Ratnayake, Journal of Geotechnical Engineering, ASCE, Vol. 120, No. 2, pp. 257-273, 1994.. Rate of excess pore pressure dissipation influenced by: high visco-plastic strains clogging of drains malfunctioning piezometer tips Sudden high rainfall Settlements may continue to occur, when excess pore pressure is still not dissipated. Solution: Increase hydraulic Gradient towards drains by applying vacuum pressure Settlement close to the embankment centreline

PWP/initial total stress Extent of the smear zone based on permeability measurement and finite element prediction Predicted and measured pore pressure during vertical drain installation

Applications (4): 3D FEM Application: Land Reclamation Stabilized with PVD and Vacuum Preloading Tianjin port , China (Rujikiatkamjorn, Indraratna and Chu 2008, Int. J. of Geomechanics, ASCE) Soil Profile 44

Soil Profile, Embankment Cross Section & Instrumentation Embankment Plan View & Instrumentation 45

Settlement Predictions Depth (m) l k n e0 gs kN/m3 kv 10-10 m/s kh,ax k’h,ax kh,ps k’h,ps OCR 0.0-3.5 0.12 0.03 0.3 1.1 18.3 6.67 20 5.91 1.46 1-1.1 3.5-8.5 0.14 0.25 1.0 18.8 13.3 40 11.8 2.92 1.2-1.5 8.5-16.0 0.20 0.04 1.35 17.5 1.2-1.6 16.0-20.0 0.10 0.02 0.27 0.9 18.5 1.67 5 1.48 0.365 1.1-1.4 Soil parameters Settlement Predictions 46

3D FEM mesh 2D FEM mesh (converted)

Finite element analysis: Vertical settlement Soil parameters Surface Settlement Predictions 2009 E H DAVIS LECTURE Buddhima Indraratna