Summary Downhole mud/shale interaction can only be properly understood if rock mechanical, shale hydration, and fluid transport phenomena are taken into account. This paper presents a review of Koninklijke Shell E&P Laboratorium's research on borehole stability in shales. Mechanisms relevant to shale stability, including pore pressure penetration (the gradual increase in pore pressure resulting from high mud weight), capillary threshold pressures, compressive and tensile failure, post failure stabilization, hydration stress, inhibition, and osmotic phenomena are discussed. We attempt to integrate these mechanisms into a comprehensive model for shale behavior. Introduction Borehole instability in shales is a major source of trouble, time, and cost during drilling. Problems generally build up over time, beginning with fragmentation at the borehole wall, followed by transfer of fragments to the annulus, and culminating in such problems (if hole cleaning is insufficient) as a sticky or tight hole, packing off, hole fill, and stuck pipe. Consequences may include losing the hole, having to sidetrack, inability to log, and poor cementations because of excessive washouts. New technologies (e.g., horizontal, slim-hole, and coiled-tubing drilling) will not resolve borehole instability problems; they will lead to at least as many problems as conventional drilling. This paper addresses the mechanisms behind shale failure and the transfer of failed material to the annulus. A proper understanding and prediction of hole cleaning is equally important but requires separate treatment.1,2 Causes of Borehole Instability in Shale Borehole instability in shale is a complex phenomenon. Five basic problem areas can be distinguished:drilling through a naturally fractured shale,drilling through a brittle shale and inducing fragmentation through drillstring vibrations,causing shale failure with too high a mud weight (tensile fracturing),causing shale failure in a compressive mode through insufficient mud-weight support, andcausing shale failure in a tensile mode through hydration stress. In Case 1, the harm is already done, and only postfailure stabilization and optimum hole cleaning can help relieve problems. In Case 2, reduction of drillstring vibrations may offer additional relief.3,4 The third case, involving classic tensile fracturing, is not considered. This paper addresses the mechanisms underlying compressive and tensile failure and postfailure stabilization of shales. First, the effect of low shale permeability on borehole stability is discussed, followed by presentation of a poro elastoplastic model for rock mechanical behavior. How postfailure stabilization may alleviate shale problems is discussed, and finally, mud/shale interaction in terms of the shale's intrinsic hydration stress is addressed. Shales, like other materials, only fail if the effective stress state exceeds the failure envelope. This is valid for compressive and tensile failure at the borehole wall and also for failure of cuttings and cavings (disintegration, dispersion). Therefore, the models presented here are formulated in terms of downhole pressure, stress, and strength effects. Shales are relatively ill-defined rocks and may include both highly cemented shaly siltstone and weak gumbo-type shales consisting primarily of hydratable clays. Major differences in shale behavior can be attributed to these differences in composition. We define shale as a low-permeability rock where the matrix consists, at least partially, of clays. How Low Permeability Affects Shale Behavior Hydraulic Flow Through Shale. Shales have permeabilities ranging from ˜1×10–6 to 1×10–12 darcy. Because of these low permeabilities, no "normal" fluid loss occurs and no filter cake builds up on the borehole wall. Instead, gradual equilibration between mud and pore pressures takes place unless a barrier is present at the borehole wall. In the case of microfractures, a shale behaves as a dual-permeability medium, with high permeabilities in the microfractures and low permeabilities in the bulk of the material. With mud in overbalance, equilibration takes place from the wellbore to a semi-infinite medium and results in transient pore pressures, penetrating from the wellbore outward. Note that only a minor amount of filtrate invasion is required to raise the pore pressure over a considerable trajectory away from the wellbore. When drilling in overbalance, pore pressure penetration invariably leads to a less-stable situation at the borehole wall. To measure the rate of mud-filtrate invasion as a function of filtrate and shale composition, we have developed the microfiltration cell (Fig. 1). The principle of the test is simple: a confined shale core sample is put into contact with a simulated pore fluid on one side and with mud on the other side. Overbalance is applied to the mud, and the rate of pressure increase at the pore-fluid side is measured. The method has been used to screen muds and mud additives for their capacity to reduce or to prevent pore pressure penetration. The pressures applied were 3.5 and 0.35 MPa on the mud and pore-fluid sides, respectively. Mud is preceded by distilled water as a baseline measurement. After a set period (from 1 to 7 days, depending on the test), the pressure is bled off, the water is replaced by mud, the fluids are repressurized, and the test continued. Tests carried out so far on Pierre shale cores have already yielded interesting results. Fig. 2 shows the results of a test with a 3% KCl/bentonite/partially hydrolyzed polyacrylamide (PHPA) mud. The rate of pore pressure penetration is very similar to that of water. Micron-sized (and larger) particles and high-molecular-weight polymers apparently cannot plug off the shale surface, let alone invade the pore system. This can easily be understood because of the ˜1- to 10-nm pore size and the general rule of plugging by particles - one-third to one-seventh of the pore diameter. Fig. 3 shows the results of a test with sodium silicate solution (water glass). In this case the "mud" can reduce the rate of pressure penetration to almost zero. The silicate reacts with the divalent ions present in the shale pore fluid to yield a gelatinous precipitate that plugs off the pore system. Fig. 4 shows the results of a test with a 25% sucrose solution. Again, the rate of pressure penetration is significantly reduced. Apparently, the sugar molecules are small enough to enter the pore system and impart high viscosity to the invading filtrate. A similar advantageous effect of sugar was seen earlier in cutting-disintegration tests.5 Fig. 5 shows the results of a test with oil-based mud (OBM). Clearly, the OBM does not penetrate the shale. An explanation is given in the next paragraph.

The increasing amounts of water being produced from oilfields, and the increasing need or necessity to return it to the reservoir it ori-ginated from, are posing a challenge to the industry. A fundamental question in Produced Water Reinjection (PWRI) is: "How clean is clean ?", or perhaps even more succinct : "How clean is fit for purpose ?" There is no universal correct answer to this question, as it depends on specific variables, largely intrinsic properties of a reservoir, its produced fluids and the contaminants that eventually end up in the produced water. PWRI for reservoir management purposes must find the balance between injector plugging and the extent of induced fractures [1], for which duty the available simulation models have been found to be wanting as they do not adequately describe leak-off dynamics. Recognizing this constraint, Shell decided that there is a need for unambiguous empirical data that do not suffer from the limitations associated with commercially available coreflood services. A coreflood rig was designed and built, featuring an accurate, automated high injection pressure capability and a much extended test duration. This test rig has since 2007 been field tested at two oilfields, main findings being:-Steady-state conditions can only be achieved in long term exposure windows-Leak-off dynamics at high dP cannot be extrapolated from experiments at low dP-Filtercake permeability depends on the permeability of the flooded core-Membrane tests (Barkman-Davidson) fail to produce representative filtercake properties Although the test rig performed satisfactorily, prospects for further development were identified, leading to a hardware upgrade and quality monitoring instrumentation that is expected to produce even better results in imminent field tests. It is expected that the now gleaned information will provide much improved input for simulators modelling fractured injection. 1. Introduction Injectivity decline is a major issue in most water floods, on a global basis some 80% of injection targets are being met. The 20 % shortfall is largely due to a mismatch between water quality and recipient reservoir. Where targets are being achieved, however, it is often doubtful whether reservoir sweep satisfies projected needs, for the very same (mismatch) reasons. Sub-specification water quali-ty may cause the injected water to predominantly enter highly permeable zones, or create large fractures in unwanted directions. When contemplating PWRI for reservoir management purposes, the cost of achieving a minimum required injection water quality (i.e. complexity of water treatment system) must be weighed against achievable incremental reservoir yield. Establishing these minimum quality needs requires knowledge of leak-off dynamics in an actual injection system. Current thinking is that chemistry and wettability related effects invalidate the traditional view on mechanistic particulate plugging of permeable media. The composite chemical landscape depends on the properties of reservoir rock, injected fluid and its contaminants, but also on additives that are always being used to facilitate key processes. As these interactions cannot easily be modelled, water qualities are being placed in a actualized perspective by resorting to empirical techniques, such as core flood tests. Many core flood tests have been undertaken over the past few decades [2], but unfortunately only few sufficiently approximated, let alone validated actual injection practices. Recognizing this, Shell developed a core flood test rig, designed to provide accurate, fully automated test facilities, capable of con-ducting tests at high injection pressure, at virtually unlimited test duration.

Through the application of horizontal well technology in a giant shallow, un-consolidated oil-rim reservoir, Shell Gabon significantly increased field output and recovery at reduced cost, while minimising the environmental impact of field development. Systematic well design changes were implemented particularly in completion hardware and mud- and completion fluids and latest coiled tubing-technology was introduced in efforts to improve completion efficiency. The paper summarises the observations and practices resulting in the improved completion technology and gives examples of the production logging tool selection and interpretation techniques developed. An outline is given of techniques which require further development to control over the inflow distribution in un-consolidated oil-rim reservoirs such that the recovery of the field can be achieved with the minimum number of wells. Introduction The Rabi field, discovered in 1985 with a current estimated STOIIP of some 1.5 109 stb and key characteristics as given in Table 1, was initially planned to be developed with vertical wells. These wells, however, experienced a significant decline, which is typical for an oil-rim situation where bean-back is the primary method used to prevent excessive gas production and limited oil recovery. Where possible, gassed-out zones were shut-off with production continuing from lower completion intervals but oil recovery per well remained low. Measures to maximise the distance to the gas-oil and oil-water contacts in the vertical wells included the successful use of sand consolidation allowing for limited completion heights as opposed to the use of gravel-packs with a normal completion height of some 10–20m in the (initially) 46m oil column, Ref. 1. Nevertheless, with the limited recovery of vertical wells (which have an average ultimate recovery of some 4.5 106 bbl per well), drainage of the remaining reserves would require some 180 additional (vertical and deviated) wells. Production history with the first horizontal wells drilled in Rabi in the period 1990-91, showed that these wells could be drilled successfully in the shallow un-consolidated reservoir sands. These wells also, although imperfect as far as completion efficiency is concerned, proved to have a significantly higher oil recovery. P. 279^

The use of integrated system modelling of wells and facilities and the application of performance improving chemicals and novel techniques allowed Shell Gabon to increase Rabi field output by nearly a factor of two over the initial projections. Optimisations are carried out in a consistent and systematic manner without jeopardising system integrity and safety standards. These efforts increased project profitability while obviating the need for major system extensions. Some (identified) upgrading measures are still required to reduce long-term operating costs and to reduce deferments.