Ikon's geological approach to well planning

Once, a trap has been defined, reservoir presence and quality, seal presence and effectiveness and source rock and charge have been verified, a prospect may be moved to the drilling phase. For this, a well plan is required. In a well plan (or “pre-drill” model), the pore pressure is determined by analysis of data in previously drilled wells, both measured pressures in porous formations such as sands and high-energy carbonates, and predicted pressures from logs in non-porous layers e.g. shales and low energy carbonates.

The drilling of this well requires that a sufficiently high mud-weight is needed such that it can withstand the pore pressure. In reality, it requires that the lower limit of the mud window must be greater than both the pore pressure and the collapse pressure. An overburden is also key as part of a well plan for a prospect. In the following sections, we highlight how access to an regional interpretation can help reduce the uncertainty in producing this well plan for a prospect i.e. the key elements are (a) pore pressure, (b) overburden, (c) fracture gradient and (d) geomechanical assessment.

Figure 1 Schematic depth/ppg plot illustrating the key components of a well plan i.e. pore pressure, fracture pressure and overburden (after Edwards and O’Connor, 2015).

Reservoir pore pressure

In porous units, Wireline Formation Tests (WFTs) are accurate measurements of pressure. These data can be used to produce overpressure maps for all stratigraphic horizons where data exists. Assessing the geospatial distribution and magnitude of the resulting overpressure allows for the identification of vertical and lateral pressure cells bounded by seals or faults as well as identifying hydrodynamic, connected reservoirs. Seismic interpretation can be used to generate structure maps and edge attributes can be used to guide the positioning of overpressure cell boundaries and assign to them appropriate confidence levels. These maps can then be used to determine the likely pressure in a prospect.

Figure 2 Integration of seismic attribute data (provided by PGS taken from a joint Ikon Science/IHS Global/PGS, 2010) to constrain fault/overpressure cell boundaries and reduce uncertainty in the overpressure cell magnitudes (after Connor et al., 2013). Inset shows an example of confidence limits applied to overpressure cells as well as magnitudes of overpressure in psi.

Non-reservoir pore pressure

For the purposes of well planning, pore pressure prediction is required for the entire well section. Pore pressure cannot be measured directly in low permeability lithologies, such as shales and carbonates, therefore indirect techniques need to be utilised.

Prediction in shales relies on using petrophysical log (typically sonic, density and resistivity logs), drilling data, and seismic velocities. The physical property underlying methods which use these data is porosity, which is predominantly controlled in shales by the interplay between deposition and compaction processes. In general terms, pore pressure is determined by comparing the measured value of shale properties (from log data) with the expected value for that measurement in a normally pressure environment using a Normal Compaction Trend (NCT).

There are several popular industry techniques for predicting pore pressure from logs (Foster & Whalen, 1965; Eaton, 1975; Bowers, 1995). Some techniques are better suited to one log type over another hence it is important to run multiple techniques on multiple log types to build an understanding of uncertainty. Furthermore, these techniques may break down under certain geological conditions, e.g. HP/HT environments or cemented rocks.

Standard pore pressure prediction methodology is not possible in carbonates because dissolution, cementation and other geochemical processes can modify carbonate porosity independent of compaction processes, even at very shallow depths (Mallon and Swarbrick, 2008; Jenkins et al., 2012). Current research is beginning to unlock potential data types that can be linked to pore pressure in carbonates but no clear path has yet been determined.

Figure 3 Pressure vs. equivalent mud-weight plot for a well, offshore Niger Delta (after Ikon Science, 2011). Wireline log data are shown for Gamma Ray (green), Sonic (blue), Density (yellow), Resistivity (red) and a volume log converted from Gamma Ray where, green denotes the amount of shale content. The corresponding pore pressure predictions are shown determined by Eaton (1975) method (blue) and Equivalent Depth method (Foster & Whalen, 1965) (green). Direct pressure data are shown for measured RFTs (red triangles).

Overburden model

Accurate calculations of the overburden gradient are not only valuable for well planning but are important inputs in shale-based pore pressure prediction methods described above. Overburden gradients are dependent on a range of rock density. If the overburden model is derived using average density data from log analysis (from many wells) for different stratigraphic/lithological units, and average values assigned, this model can be easily applied at a prospect location if seismic markers and seismic facies can provide a local model for stratigraphy and lithology.

Figure 4 Stratigraphically defined overburden modelling (after O’Connor et al., 2013). High quality density data is utilised to derive a series of average gradients, one per Stratigraphic interval. Therefore, at a new location, only seismically-derived marker thicknesses are needed to generate a model.

Fracture gradient

The choices to determine fracture strength include using off-the-shelf Gulf of Mexico-derived algorithms such as Matthews & Kelly (1967). These are predominantly empirical and have specific constants determined by the datasets from which they were derived. A more satisfactory approach is to use the actual LOT data to construct a regional model in a similar manner to the overburden, that is, stratigraphically-controlled and related to changes in pore pressure. The advantage is that not all tests are accurately interpreted or reported leading to an uncertainty in the local failure strength of the rock. The chance therefore of having good test data locally is limited. The importance of evaluating the fracture strength correctly can impact directly on calculating sealing capacities, column heights and Well-Bore Stability (or “WBS”) estimates.

Figure 5 Pressure-Depth Plot of showing pore fluid and LOT pressures from data in Mid-Norway (after O’Connor et al., 2013). At Jurassic level, the reservoir (and associated shales) either has close to hydrostatic or very high pore pressure. The Leak-Off Test (LOT) show a corresponding separation. Above the Jurassic, the pore pressures are the same, and the LOT data, a single linear trend.

Borehole stability

Once the above have been defined, they become key inputs into the final well plan where a detailed geomechanical analysis can be undertaken, this analysis determines the stable mud-weight window and an optimum wellbore trajectory (if deviation is required). The aim is to minimise unscheduled events such as lost circulation, pack-offs, stuck pipe etc. Part of this analysis is to also define the collapse pressure. Read more.

Real-time decisions during drilling wells

Once drilling commences, data of many types are recorded which can be monitored to help re-calibrate the pre-drill well plan. These data include mud-weight, D-exponent (a calculated relationship based on data such as weight on bit, rate of penetration etc.), connection gas, cuttings, torque, mud temperature in and out, bit size, and caliper. In addition Logging-While-Drilling (“LWD”) velocity and resistivity data could be part of a Bottom-Hole Assembly (BHA) which, in conjunction with a normal compaction trend defined by analysis of previously drilled wells, can be used to monitor the pressures encountered whilst drilling ahead. Read more.