Stress partitioning and Main wellbore failure in the West Tuna Area, Heading Gippsland Basin Emma Nelson 1 Richard Authors Hillis 2 Scott Mildren 3

Exploration Geophysics (2006) 37, 215 221 Stress partitioning and Main wellbore failure in the West Tuna Area, Heading Gippsland Basin Emma Nelson 1 Richard Authors Hillis 2 Scott Mildren 3 Key Words: Gippsland Basin, wellbore Key Words: stability, key present-day words stress, finite element modelling ABSTRACT Image logs from the deep intra-latrobe and Golden Beach Subgroups of the West Tuna area in the Gippsland Basin reveal that wellbore failure is restricted to fast, cemented sandstone units and does not occur in interbedded shales. Triaxial testing and analysis of empirically derived, wireline-log based strength equations reveals uniaxial compressive strengths of 60 MPa in sandstones and 30 MPa in shales in the West Tuna area. Conventional analysis of wellbore failure assumes constant stresses in the shales and adjacent sandstones and that breakout is focused in the weaker units. We propose that the flat lying, strong, cemented sandstone units in the West Tuna area act as a stress-bearing framework within the present-day stress regime that is characterised by very high horizontal stresses (S Hmax > S hmin = S v ). Stress focusing in strong sandstone units can result in high stress concentrations at the wellbore wall and account for the restriction of wellbore failure to the strong sandstone units. Finite element methods were used to investigate the stress distribution in horizontal, interbedded strong sands and weak shales subject to a high present-day stress state such as exists in the West Tuna area (S Hmax > S v ~ S hmin ). Modelling using the present-day stress tensor and estimated elastic properties for the sandstones and shales indicates that the present-day stress is partitioned between strong inter-bedded sandstones and weaker shales, with high stress being focussed into the strong sandstones. The stress focusing causes borehole breakout in the sands despite their higher strength. Conversely, stresses are too low to generate wellbore failure in the weaker shales. INTRODUCTION Stress magnitudes and orientations have been shown to vary substantially between lithologies in petroleum fields (Warpinski et al., 1985; Evans et al., 1989; Raaen et al., 1996). It has been proposed that the variation in stress distribution through different lithological units in a sedimentary basin (herein referred to as stress partitioning) depends on relative rock strength and the present-day stress state (Plumb, 1994; Reches, 1998). Plumb (1994) found that the ratio of minimum to vertical present-day stress was 1 BP Exploration (formerly at Australian School of Petroleum), Chertsey Road, Sunbury-on-Thames Middlesex TW16 17LN United Kingdom Phone: +44 (0)1932 739521 Facsimile: +44 (0)1932 738411 Email: Emma.Nelson@bp.com 2 Australian School of Petroleum, The University of Adelaide, SA, Australia, 5005 3 JRS Petroleum Research Pty. Ltd. 1 45 Woodforde Rd Magill, SA, 5072 2 Manuscript received 6 January, 2006. Revised manuscript received Fig. 1. Location of the West Tuna study area in the Gippsland Basin (offshore Victoria). 40% greater in hard carbonate rocks, and 20% higher in hard sandstones, than in the weak shales in sedimentary basins at high present-day stress (reverse stress regimes). Similarly, the ratio of minimum to vertical stress was found to be 4 to 15% higher in weak shales than in strong sandstones in basins in a relaxed present-day stress state (normal fault regimes). Although reliable methods have been developed for estimating present-day stress in individual reservoirs using leak-off tests during drilling or specialised injection tests (such as mini-frac tests), these are often expensive and do not test all lithological units intersected by the wellbore. As such, the present-day stress tensor can often only be examined at a field or basin scale. The low spatial resolution of stress data is a problem in the assessment of wellbore stability and planning of fracture stimulation operations, where the present-day stress tensor is required in both the reservoir and surrounding non-reservoir units (Desroches and Woods, 1998; Smith et al., 2001; Barree, 2004). The interpretation of six image logs covering the deep intra- Latrobe and Golden Beach Subgroups in the West Tuna area of the Gippsland Basin (Figure 1) revealed that borehole breakout occurs in strong, cemented sandstones and is absent in the weaker shales. This is inconsistent with conventional wellbore stability analysis, which tends to assume that wellbore failure occurs in the weaker units (Hickman et al., 1985; Zoback et al., 1993; Reinecker and Lenhardt, 1999). This paper reviews the presentday stress in the Gippsland Basin and compares the relative strengths of sandstones and shales from the deep intra-latrobe and Golden Beach Subgroups. Finite element modelling is then used to show that the high present-day horizontal stresses that exist in the Gippsland Basin may be partitioned to strong lithological layers and can account for the occurrence of breakout in the cemented sandstones. 215

PRESENT-DAY STRESS IN THE WEST TUNA AREA The orientation of 96 borehole breakouts (194 m in cumulative length) and four drilling-induced tensile fractures (55 m in cumulative length) interpreted from six image logs consistently indicate that S Hmax is oriented ~138 N in the West Tuna area. The vertical stress magnitude was derived from density, sonic and checkshot velocity data from the Tuna-4 exploration well and the West Tuna-39 development well. The two vertical stress profiles indicate that S v is consistent across the West Tuna area, ranging from 20 MPa at 1 km to 66 MPa at 3 km depth (Figure 2). The lower bound to the leak-off pressures from deep intra-latrobe and Golden Beach Subgroups suggests that the minimum horizontal stress gradient (S hmin ) is ~20 MPa/km (close to S v ; Figure 2). Modular Dynamic Test pressures indicate that pore pressure is hydrostatic in the sandstones above 2800 m (Figure 2). The S Hmax magnitude can be constrained using observations of drillinginduced tensile fractures (DITFs) on image logs and the criterion for formation of DITFs in elastic, impermeable rocks in vertical wellbores (Peska and Zoback, 1995; Brudy and Zoback, 1999; Nelson et al., 2005). Assuming the stress magnitudes defined above, then the occurrence of DITFs in the West Tuna area constrains the magnitude of S Hmax to ~ 40 MPa/km (in sands above 2800 m). The present-day stress tensor suggests that the West Tuna area is on the boundary between strike-slip and reverse faulting stress regimes (i.e., S Hmax > S v S hmin ; Figure 2). Fig. 2. Stress depth plot for the West Tuna area. The diamonds represent the depths at which the stress gradients were assessed. S hmin has been determined from leak-off tests (red data), S v from density and check shot data (light green gradient from West Tuna-39, dark green gradient from Tuna-4), S Hmax from the occurrence of drilling-induced tensile fractures (orange data) and P p from MDT tests (blue data). The orientation of S Hmax in the West Tuna area is consistent with previously determined orientations in the greater Gippsland Basin and southeast Australia as a whole (Nelson and Hillis, 2005; Nelson et al., 2006). The high present-day stress magnitudes are also consistent with earthquake focal mechanisms and the neotectonic record, including significant reverse faulting in Southeast Australia (Nelson et al., 2006). The orientation of the horizontal stress, and the high, present-day horizontal stress magnitudes in the Gippsland Basin have been suggested to be a consequence of contemporary oblique compression along the New Zealand plate boundary (Coblentz et al., 1995; Sandiford, 2003; Nelson et al., 2006). ROCK STRENGTH IN THE WEST TUNA AREA Knowledge of the compressive strength of rocks is important to understanding wellbore failure in petroleum fields (Zoback et al., 1985). The uniaxial compressive strength (UCS) of rocks can be determined from laboratory testing of core samples or estimated using empirical relationships based on wireline log data. Both of these methods have been used in this study and are outlined in the section below. Laboratory testing of core Fig. 3. Borehole breakout in a cemented sandstone interpreted on the West Tuna-8 image log. Resistive features around the wellbore appear white, conductive features appear brown or black. The track on the right shows a dynamic presentation of the image log, while the track on the left shows a static image. Note the low gamma ray count, fast sonic velocity, and reasonably high density. The static image shows true resistivity and indicates that the resistivity is quite high in the sandstone unit. Two laboratory strength measurements were available from triaxial tests undertaken on the deep intra-latrobe and Golden Beach Subgroups in the West Tuna area (Nelson, 2002). These sandstone units are characterised by low porosity, low permeability (<5 md), high resistivity, low gamma ray count ( ~ 60 70 gpi) and a high sonic velocity (Δt = 60 85 µs/ft), which suggests that they are cemented and likely to be hard and strong (Figure 3). Failure envelopes constructed from the multi-stage triaxial test data indicate that the sandstones have a cohesion (C) of ~ 13.5 MPa and a co-efficient of friction (µ) of ~ 0.90 (Figure 4). The uniaxial compressive strength (C 0 ) is not directly measured in a triaxial test but can be determined by substituting C and µ into equation (1) below. Equation (1) indicates that the average uniaxial compressive strength of the two sandstone samples is ~ 60 MPa: ( ). (1) C 0 = 2C ( 2 + 1) 0.5 216

Unfortunately no laboratory testing of UCS in shales from the deep intra-latrobe or Golden Beach Subgroups was available in the Gippsland Basin. Triaxial test data was available for one shale sample from the Lakes Entrance formation ( ~ 2400 m depth; Figure 4). The failure envelope derived from the triaxial test results indicates a cohesion of ~ 8 MPa and a coefficient of friction of 0.6. When these values are substituted into equation (1) the uniaxial compressive strength is determined to be ~ 30 MPa. Shales in the deep intra- Latrobe and Golden Beach subgroups are characterised by their high gamma count (> 100 gpi) and relatively slow sonic velocity (Δt = 90 100 µs/ft). Analysis of wireline log data over the sampled interval reveals that the shale tested is (petrophysically) very similar to those at depth (GR = 100 gpi, Δt = 90 µs/ft). Hence the shale tested is believed to be representative of those observed in the image logs in this study. Empirical relationships Since only three UCS tests were available in the Gippsland Basin, generic empirical relationships were used to determine UCS from sonic logs in sands and shales from three West Tuna wells. Log-based lithological discrimination was based on gamma ray (GR), density (RHOB), sonic velocity (DT) and caliper (CALI) logs. The logs were first examined to determine if the lithology was coal (distinguished by low DT, RHOB and GR), and if not the GR was used to discriminate sandstone and shale. Sandstones were interpreted where GR < 80 API, shales were interpreted where the GR > 100 API. The sonic velocities from each lithology were then substituted into the sand or shale empirical relationships: Fig. 4. Mohr circles and Mohr-Coulomb failure envelope from triaxial testing of a cemented sandstone in the Golden Beach sub-group and a shale from the Lakes Entrance formation. The sandstone sample has a µ approximately equal to 0.9, a cohesive strength approximately equal to 13.5 MPa and a compressive strength of 60 MPa. The shale sample has a µ approximately equal to 0.6, a cohesive strength approximately equal to 8 MPa and a compressive strength of 30 MPa. (2) (3) high horizontal present-day stress environment such as exists in the West Tuna area. Stresses Around a Vertical Wellbore As a well is drilled, the wellbore wall must support stresses previously carried by the removed rock. This causes a stress concentration about the borehole that depends on the orientations of the wellbore and of the far field present-day stress (Kirsch, 1898; Jaeger and Cook, 1979; Amadei and Stephansson, 1997). Three principal stresses act on the wall of a vertical well (Kirsch, 1898), these are: the radial stress (σ rr ) which acts normal to the wellbore, where UCS is in MPa and Δt is in microseconds per foot (µs/ft). An example UCS log from the West Tuna-39 development well is shown in Figure 5. The McNally (1987) equation was developed using strength data from sandstones in a shallow coal field environment in the Bowen basin however the range of UCS calculated using the McNally equation for the intra-latrobe and Golden Beach sandstones was found to be ~ 59 147 MPa (average Δt = 60 85 µs/ft) which compare very well to the laboratory derived values. The Chang (2004) equation was based on a global shale data set and determines the range of UCS for the intra-latrobe and Golden Beach shales to be ~ 24 32 MPa (average Δt = 90 100 µs/ft) which is consistent with lab-derived UCS and verifies that shales are much weaker than the sands in the West Tuna area. STRESS DISTRIBUTION AT THE WELLBORE WALL Borehole breakout forms when the circumferential stress at the wellbore wall exceeds the compressive strength of the rock (Peska and Zoback, 1995). From this perspective it is somewhat counter-intuitive that breakout forms in the strong sandstones rather than the weak shales in the West Tuna area (Figure 5). The following sections outline the Kirsch equations that describe stress concentration at the wellbore wall and wellbore failure. We will show that shales would be expected to fail prior to sands using a conventional Kirsch approach to wellbore stability in the West Tuna area. Finite element methods are then used to investigate stress variation by lithology (and rock strength) in a the axial stress (σ zz ) which acts parallel to the vertical wellbore axis, and, the circumferential stress (σ θθ ), which acts orthogonal to σ rr and σ zz. The stress at the wall of a wellbore drilled through elastic, homogeneous and isotropic rock can be described by the Kirsch (1898) equations. The magnitude of the stresses depends on the magnitude of the far field stresses, the radius of the wellbore (R), distance from the wellbore (r) and the pore pressure (P p ). For the full Kirsch equations and a comprehensive description of wellbore stresses, the reader is referred to Jaeger and Cook (1979). The Kirsch equations may be simplified if only the stresses at the wellbore wall are considered (i.e., where R = r). The simplified Kirsch equations are described in Moos and Zoback (1990). The Kirsch equation for circumferential stress at the wall of a vertical well can be written: = ( S ' H max + S ' h min ) 2( S ' H max S ' h min )cos2 P (4) where P is the difference between mud pressure and pore pressure (P w -P p ), θ is the angle between the S Hmax azimuth and north and S Hmax and S hmin are effective stresses (Moos and Zoback, 1990). If the circumferential stress is plotted with respect to position around the wellbore wall (Figure 6) it can be shown that σ θθ is maximum when θ = 90 (i.e., at the azimuth of far-field S hmin ), 217

circumferential stresses are calculated to be 0 and 80 MPa respectively assuming mud-weight in the well is in balance with the pore pressure (P w = P p ). The uniaxial compressive strengths determined for sandstones and shales in the West Tuna area have also been plotted on Figure 7. Figure 7 illustrates that borehole breakout is more likely in shales than in sandstones in the West Tuna area if a conventional approach (which considers a homogeneous stress distribution, independent of lithology) is assumed. This is contrary to what is observed in the West Tuna area. One possibility that explains the systematic occurrence of borehole breakout in the strong sandstone units is that stress is unequally distributed between different lithologies in the West Tuna area. We next investigate the likelihood of this possibility through finite element modelling. Finite Element Modelling A possible mechanism that explains the observed wellbore failure in the West Tuna area is that present-day stress is higher in the sandstone units than in the interbedded shales. Drilling-induced tensile fractures and leak-off tests have been used to constrain the present-day stress tensor in sandstones in the West Tuna area. However, since no leak-off tests or wellbore failure exist in the shales, there is no data on stress magnitude in the shales. A finite element model was constructed to investigate the concept of stress variation by lithology (stress partitioning), and to predict the state-of-stress in the shales in the West Tuna area. Linear elastic constitutive equations were used so that the results of the finite element model could be directly compared with the conventional approach to wellbore stability (using the Kirsch equations). Pore pressure and mud-weight have been accounted for by assuming an in-balance well and applying effective stresses at the model boundaries. The finite element modelling presented herein is only intended to investigate the concept of stress partitioning in the West Tuna area and is not intended to be fully predictive. Fig. 5. UCS log derived from sonic log data (blue), gamma ray log (red) and interpreted borehole breakouts (black) from the West Tuna- 39 development well. The logs show that borehole breakout occurs in sandstone (low GR) intervals with high compressive strength (high UCS). and minimum when θ = 0 (i.e., at the azimuth of far-field S Hmax ). Equation (4) can therefore be simplified for the conditions where θ = 0 and θ = 90 (Equations 5 and 6 respectively). ( ). max = 3S H max S h min P = 90 ( ). min = 3S h min S H max P = 0 Borehole breakouts form due to shear failure at the wellbore wall when the maximum circumferential stress around the wellbore exceeds the compressive strength of the rock (Peska and Zoback, 1995; Barton et al., 1998). The circumferential stress is maximised at θ = 90, hence breakouts form in the wellbore at the azimuth of S hmin (Figure 6). Equation (5) can be modified to represent the criterion for formation of breakouts in elastic, impermeable rocks in vertical wellbores such that: (5) (6) max = 3S H max S h min P C 0, (7) where C 0 is the uniaxial compressive rock strength. Using Conventional Methods to Assess Wellbore Failure The present-day stress tensor determined in the West Tuna area (S Hmax = 40 MPa/km, S hmin = 20 MPa/km, Pp = 10 MPa/km) can be substituted into equation (4) for all orientations around the wellbore wall (Figure 7). The minimum and maximum The stress concentration around the wellbore in a single sandstone layer predicted by the finite element code DIANA, was verified against the Kirsch equations for stress concentration around a vertical well. A 3D, linear elastic model of a wellbore drilled in a homogeneous block was constructed (TNO Construction Research, 2000). Structural boundary conditions were defined by imposing displacement constraints at three of the model boundaries (in the xy plane, yz plane and the zx plane). The three effective principal stresses previously derived (S' Hmax = 30, S' hmin = 10 and S' v = 10 MPa/km) were applied perpendicular to the remaining three model boundaries. The model boundaries were applied in this way to represent the high horizontal present-day stresses in the West Tuna area due to the compressional plate boundary at New Zealand. The model boundaries were constructed approximately six well diameters from the wellbore wall so that boundary effects were minimal. The perturbed near-wellbore stress environment is generally considered to extend three wellbore diameters from the well (Fjær et al., 1992). A Poisson s ratio of 0.25 and an elastic modulus of 40 GPa are considered average values for sandstones and were applied in the model construction (Lama and Vutukuri, 1978). The model geometry was discretised into elements using meshing algorithms inbuilt in DIANA. The modelled circumferential stress at the wellbore wall in a homogeneous sandstone reservoir is presented in Figure 8. The modelled circumferential stress is consistent down the wellbore and is a maximum of 80 MPa/km at the azimuth of S hmin (blue) and a minimum of 0 MPa/km at the azimuth of S Hmax (red). The circumferential stresses at the wellbore wall calculated using the finite element approach are consistent with those derived from the Kirsch equations (Table 1 and Figure 7). 218

Nelson, Hillis,Heading and Mildren Left Running Stress partitioning Right in therunning Gippsland Heading Basin Fig. 6. Circumferential stress diagram showing that breakouts form when the circumferential stress exceeds the compressive strength of the rock and drilling induced tensile fractures form when the circumferential stress is less than the tensile strength of the rock. Fig. 8. Finite element model of circumferential stress about a vertical wellbore drilled through homogeneous rock. Blue represents high stress, red represents low stress. Sandstone Shale Wellbore wall (MPa/km) 80 σθθmax 0 σθθmin 25 σθθmax σθθmin 5 Fig. 9. a) 3D finite element model of interbedded sandstone and shales. The sandstone layers (red) were assigned an elastic moduli of 40 GPa and a Poisson s ratio of 0.25. The shale layer (orange) was assigned an elastic moduli of 0.86 GPa and a Poisson s ratio of 0.3. b) Circumferential stress at the wellbore wall derived from the finite element model. Far-field (MPa/km) SHmax Shmin SHmax Shmin 40 20 20 15 Table 1. Far-field (total) stresses and circumferential stresses at the wellbore wall in modelled sandstone and shale layers calculated using FE methods. Once the finite element code and methodology were verified using the model above, a simplified 3D, linear elastic model of a wellbore drilled through alternating strong sandstone and weak shale layers was constructed (Figure 9a). The stress boundary conditions were applied such that the present-day stress tensor in the sandstone (in the far-field and at the wellbore wall) was consistent with that determined previously (SHmax ~ 40 MPa/ km, Shmin ~ 20 MPa/km and Sv ~ 20 MPa/km; σθθmax = 80 and σθθmin = 0). Based on Lama and Vutukiri (1978) the upper and lower sandstone layers were assigned a Poisson s ratio of 0.25 and an elastic modulus of 40 GPa (red in Figure 9a). The middle 219 Fig. 7. Plot of circumferential stress and average compressive rock strengths for sandstones and shales in the West Tuna area. Assuming a conventional Kirsch model for wellbore failure (equivalent stresses in the sandstone and shale); shales are more likely to fail than sandstones in the West Tuna area. shale layer was assigned a Poisson s ratio of 0.35 and an elastic modulus of ~ 10 GPa (orange in Figure 9a). The results of the multilayer 3D finite element model are shown in Figures 9b and 10 and support the hypothesis that the stress may be partitioned between stronger sandstone units and adjacent shales. The present-day stress is focused in the sandstones, where horizontal, interbedded, lithologies of varying strengths are subjected to high present-day horizontal stresses. The modelled sandstone units are at high stress equivalent to the total far-field tectonic stress (SHmax = 40 MPa and Shmin = 20 MPa; Table 1) whilst the total stress in the shale prone units is lower (SHmax = 20 MPa and Shmin = 15 MPa; Table 1). This mechanism for stress partitioning is analogous to stiff boards separated by sponge in a vice. The strong units (sandstones) take up the stress load whilst the sponge (shale) remains relatively unstressed and undeformed. IMPLICATIONS FOR WELLBORE STABILITY The finite element modelling of interbedded sandstones and shales shows that where far-field stress is low in the shales, the stress concentration at the wellbore wall is low (and more isotropic). Low stress concentration and low stress anisotropy

a northeast southwest (σ h ) azimuth as shown in Figure 11. The stresses acting on a horizontal well deviated in the southeast northwest (σ H ) azimuth are isotropic and hence this is a relatively stable drilling trajectory. The risk of breakout formation in the shale is low in all directions because of the lower stress anisotropy in the shales. CONCLUSIONS Fig. 10. Circumferential stress in the a) sandstone units and b) shale units derived from petroleum data and finite element modelling. The circumferential stress at the wellbore wall in the sandstone exceeds the compressive strength of the sandstone that should result in the formation of borehole breakout. Conversely, the circumferential stress in the shale does not exceed the compressive strength in the shale and hence no borehole breakout would be expected. Conventional analysis of wellbore failure (assuming constant stress in all lithologies) predicts that borehole breakout occurs in weak shales rather than strong sandstones in the West Tuna area of the Gippsland Basin. However, analysis of image logs has revealed that wellbore failure only occurs in strong, cemented sandstones in the West Tuna area. Rock strength testing and sonic log-derived uniaxial compressive strengths suggest that sandstones are much stronger than shales in the West Tuna area. A finite element model of horizontal, interbedded, strong sandstones and weak shales was constructed and subjected to a high horizontal far-field present-day stress load equivalent to that believed to exist in the Gippsland Basin. The model revealed that strong sandstones act as a stress-bearing framework when subjected to high horizontal stresses. The sandstone units effectively shield the shale-prone ( weak ) sections from the high horizontal stress load resulting in the shales being at a much lower and less anisotropic stress state than the sandstones. Stress partitioning (higher stress in the sandstones relative to the shales) results in high stress concentrations at the wellbore wall in the sandstones whilst the stress concentrations in the shales are much lower. We suggest this mechanism accounts for the high degree of borehole breakout in the sandstones, and absence of borehole breakout in the shales of the Latrobe Group and Golden Beach Subgroups in the West Tuna area. Fig. 11. Breakout risk diagrams constructed using the analytically determined present-day stress tensor in the sandstone (S Hmax orientation = 138 N, S Hmax ~ 40 MPa/km, S hmin ~ 20 MPa/km and S v ~ 20 MPa/km), and the modelled stress tensor in the shale (S Hmax orientation = 138 N, S Hmax ~ 16 MPa/km, S hmin ~ 14 MPa/km and S v ~ 15 MPa/km). Red indicates high risk of breakout, blue indicates low risk of breakout. decreases the propensity for wellbore failure during drilling (Figures 9b and 10). Assuming breakouts develop when the maximum circumferential stress exceeds the uniaxial compressive strength of the rock, then the finite element model suggests that the propensity for breakout is high in the sandstone (where σ θθmax ~ 80 MPa/km) and lower in the shale (where σ θθmax ~ 25 MPa/km). The contrast in circumferential stress between the modelled strong sandstones and weak shales can explain the occurrence of wellbore failure in the strong lithologies in the West Tuna area and the absence of wellbore failure in the weaker shale-prone lithologies (Figure 10). The modelling suggests that optimum mud weights to prevent well collapse for drilling should be planned considering the present-day stress in the sandstone units. The risk of breakout development in wells of all orientations can be assessed in terms of the rock strength (compressive strength, C 0 ) required to prevent breakout formation normalised to S v and the present-day stress tensor (Brudy and Zoback, 1999; Figure 11). A breakout risk diagram has been constructed using the present-day stress tensor determined analytically in the sandstones and the stress tensor derived from the finite element modelling in the shales in the West Tuna area. Blue colours indicate that low rock strength (or low mud weight) is required to prevent breakout. Red colours indicate that high rock strength (or high mud weight) is required to prevent breakout. The risk of breakout formation in the sandstones in West Tuna is greatest for a vertical well and for wells deviated in The finite element modelling undertaken herein illustrates that different lithological units may behave differently under an applied tectonic stress load. Hence, lithology should be considered when assessing wellbore failure on image logs, and indeed when using petroleum wellbore data such as injection tests for present-day stress analysis. Field and basin scale present-day stress tensors should be treated with caution during well planning or other operations which require knowledge of the present-day stress tensor (e.g., hydraulic fracturing or waterflooding). ACKNOWLEDGMENTS The authors thank Glen Nash, Michael Power, Adem Djakic, and Wayne Mudge (ExxonMobil) for supplying image log data and providing useful insight. Mark Tingay and Mike Dentith are thanked for their comments, which greatly improved the manuscript. Wilde FEA Ltd is thanked for the use of DIANA and the ASEG RF for supporting Emma Nelson s PhD project. REFERENCES Amadei, B. and Stephansson, O., 1997, Rock Stress and its Measurement: Chapman and Hall. Barree, R.D., 2004, Fracture modeling with GOHFER using extensive real-data input: Oilasia Journal. Barton, C.A., Castillo, D.A., Moos, D., Peska, P., and Zoback, M.D., 1998, Characterising the full stress tensor based on observations of drilling-induced wellbore failures in vertical and inclined boreholes leading to improved wellbore stability and permeability prediction: Australian Petroleum Production and Exploration Association Journal, 38, 466 487. Brudy, M. and Zoback, M.D., 1999, Drilling-induced tensile wall-fractures: Implications for determination of in situ stress orientations and magnitudes: International Journal of Rock Mechanics and Mining Sciences, 36, 191 215. 220

Chang, C., 2004, Empirical rock strength logging in boreholes penetrating sedimentary formations: Geophysical Exploration, 7, 174 183. Coblentz, D.D., Sandiford, M., Richardson, R.M., Zhou, S., and Hillis, R.R., 1995, The origins of the intraplate stress field in continental Australia: Earth and Planetary Science Letters, 133, 299 309. Desroches, J. and Woods, T.E., 1998, Stress measurements for sand control: SPE/ISRM Eurock, Trondheim, Norway. Society of Petroleum Engineers, SPE/ISRM 47247. TNO Building and Construction Research, 2000, DIANA 7.2, program and users manual (CD-ROM). Evans, K., Engelder, T., and Plumb, R.A., 1989, Appalachian Stress Study 1: A detailed description of in situ stress variation in Devonian Shales of the Appalachian Plateau: Journal of Geophysical Research, 94, 1729 1754. Fjær, E., Holt, R.M., Horsrud, P., Raaen, A.R., and Risnes, R., 1992, Petroleum related rock mechanics: Developments in Petroleum Science, vol. 33: Elsevier. Hickman, S.H., Healy, J.H., and Zoback, M.D., 1985, In situ stress, natural fracture distribution, and borehole elongation in the Aurburn geothermal well, Auburn, New York: Journal of Geophysical Research, 90, 5497 5512. Jaeger, J.C. and Cook, N.G.W., 1979, Fundamentals of rock mechanics, Chapman and Hall. Kirsch, G., 1898, Die Theorie der Elastizitat und die Beaurforisse der Festigkeitslehre: Zeitschrift des Vereines Deutscher Ingenieure, 42, 797 807. Lama, R.D. and Vutukuri, V.S., 1978, Handbook on mechanical properties of rocks testing techniques and results: Series on Rock and Soil Mechanics, vol. 3: Trans Tech Publications McNally, G.H., 1987, Estimation of coal measures rock strength using sonic and neutron logs: Geoexploration, 24, 381 395. Moos, D. and Zoback, M.D., 1990, Utilization of observations of wellbore failure to constrain the orientation and magnitude of crustal stresses: application to continental, deep sea drilling project and ocean drilling program boreholes: Journal of Geophysical Research, 95, 9305 9325. Nelson, E.J., 2002, The in situ stress field of the West Tuna area, Gippsland Basin: Implications for natural fracture permeability and wellbore stability: B.Sc.Hons thesis (unpublished), The University of Adelaide. Nelson, E.J. and Hillis, R.R., 2005, In Situ stresses of the West Tuna Area, Gippsland Basin: Australian Journal of Earth Sciences, 52, 299 313. Nelson, E.J., Hillis, R.R., Reynolds, S.D., Sandiford, M., and Mildren, S.D., 2006, The present-day state-of-stress in Southeast Australia: Australian Petroleum Production and Exploration Association Journal, 46, 283 305. Nelson, E.J., Meyer, J.J., Hillis, R.R., and Mildren, S.D., 2005, Transverse drillinginduced tensile fractures in the West Tuna area, Gippsland Basin, Australia: implications for the in situ stress regime: International Journal for Rock Mechanics and Mining Sciences, 42, 361 371. Peska, P. and Zoback, M.D., 1995, Compressive and tensile failure of inclined wellbores and determination of in situ stress and rock strength: Journal of Geophysical Research, 100, 12791 12811. Plumb, R.A., 1994, Variations of the least horizontal stress magnitude in sedimentary rocks: 1st North American Rock Mechanics Symposium, North American Rock Mechanics Association. Raaen, A., Hovern, K., Joranson, H., and Fjaer, E., 1996, FORMEL: A step forward in strength logging: SPE Annual Technical Conference, Denver, Colorado. Society of Petroleum Engineers, SPE 36533. Reches, Z., 1998, Tensile fracturing of stiff rock layers under triaxial compressive stress states: Journal of Rock Mechanics and Mining Science, 35, 4 5. Reinecker, J. and Lenhardt, W.A., 1999, Present-day stress field and deformation in eastern Austria: International Journal of Earth Sciences, 88, 532 550. Sandiford, M., 2003, Neotectonics of southeastern Australia: linking the Quaternary faulting record with seismicity and in situ stress: in Hillis, R.R. and Muller, R.D. (eds.), Evolution and dynamics of the Australian plate, 107 119. Smith, M.B., Bale, A.B., Britt, L.K., Klein, H.H., Siebrits, E., and Dang, X., 2001, Layered Modulus Effects on Fracture Propagation, Proppant Placement, and Fracture Modeling: SPE Annual Technical Conference and Exhibition, New Orleans, Louisiana. Society of Petroleum Engineers, SPE 71654. Warpinski, N.R., Brannagan, P., and Wilmer, R., 1985, In-situ stress measurements at U.S. DOE s multiwell experiment site, Mesaverde Group, Rifle, Colorado: Journal of Petroleum Technology, 37, 527 536. Zoback, M.D., Moos, D., and Anderson, R.N., 1985, Well bore breakouts and in situ stress: Journal of Geophysical Research, 90, 5523 5530. Zoback, M.D., Apel, R., Baumgartner, J., Brudy, M., Emmermann, R., Engeser, B., Fuchs, K., Kessels, W., Rischmuller, H., Rummel, F., and Vernik, L., 1993, Uppercrustal strength inferred from stress measurements to 6 km depth in the KTB borehole: Nature, 365, 633 636. 221