Low Resistivity Pay in Carbonate ROS
Abu Dhabi 30th Jan. – 2nd Feb. 2005
The following notes summarize the main points raised during the talks and breakout sessions of the Conference. Where a consensus was reached this is also included. These notes are intended to ensure that the volume of good work presented at the Conference is not lost – however since the talks were Confidential – only general – and not specific issues can be mentioned.
Low Resistivity Pay (LRP) and Low Contrast Pay (LCP) are closely related concepts. Low Contrast Pay is the more general term and applies when it is difficult to determine whether a formation will flow oil or water. This may occur because there is little difference between the resistivity of oil and water bearing intervals, or because there is little contrast in the resistivity between pay beds and surrounding non-reservoir rock such as shales. The absolute value of the resistivity in LCP may be moderately high – particularly when the formation water is very fresh.
Low Resistivity Pay is the special case of LCP when the absolute value of the resistivity is also low – below 1-2 Ohm.m as a general rule of thumb. Consequently, it is most frequently encountered in areas with saline formation waters. In extreme cases the resistivity may be so low that pay zones are overlooked since the calculated water saturations are very high – or 100%.
Two end-member types of LRP may be identified:
Type 1. Pay where the water saturation as calculated from the resistivity (using normal Archie methods) is incorrect and overestimates the true water saturation of the formation. The issue is then to find improved methods for calculating the true water saturation – either by modification of the calculation algorithm used from the resistivity, or possibly by obtaining saturation data by an alternative and independent means.
Type 2. Pay where the water saturation as calculated from the resistivity (using normal methods) is correct, but very high. The issue is then to understand why the water is effectively not mobile – and to develop some means for predicting which rocks will flow dry oil and which will flow wet – from rocks with the same water saturations.
Frequently, the truth may represent some combination of the two above explanations, which are best regarded as end-members of a distribution. However, the distinction remains useful since the diagnosis required depends on which of the two causes is the more important.
It was generally agreed however that true water saturations of LRP were only rarely less than 50% — though they may range up to 80% and even higher than this — in some extreme cases.
Five distinct causes of LRP were identified. There was a consensus that they could be ranked in terms of importance and frequency of occurrence. The list below is ordered from least important to most important:
- Rt incorrectly measured. – Rare- It is assumed that the logging tool gives a misleading value for Rt – for example as a result of excessive invasion of saline mud filtrate. Very few examples of this were presented. One example is wipe logging with LWD some considerable time after drilling; when invasion does occasionally exceed the depth of the relatively shallow LWD resistivity tools.
- Conductive Minerals – Rare – This is often the case in LRP clastic reservoirs where pyrite and other conductive minerals are common. Conductive minerals are only rarely found in carbonate reservoirs – and in carbonate systems are largely confined to non-reservoir (carbonate mud) facies.
- Fractured formations – Sometimes – A formation with a fine grain and largely water bearing matrix but oil filled fractures can exhibit low resistivity but flow essentially dry oil.
- Layered formations – Often – Formations composed of thin layers of small pore size rock (micrite etc.) and coarser high permeability rock (grainstone etc.) give a very misleading resistivity response. It is assumed that the capillary pressure is such that oil cannot enter the pores of the fine grain rock. The ratio between the layers of coarse and fine grain rock controls the average water saturation – which may vary essentially arbitrarily to very high values if the coarse grain layers are thin. A standard arithmetic solution of the expected resistivity of such a formation then shows that a simple Archie approach will lead to an overestimation of the water saturation – since the conductive fine grain rock essentially “short circuits” the resistivity measuring current.
- Dual Porosity Systems – Most Common – It is assumed that the formation contains pores of greatly varying pore size – and has a bi or even tri-modal distribution on mercury injection data. The entry pressure of the smaller pores is thus such that oil cannot enter them and they remain water filled. On test oil in the large pores may flow, while capillary bound water in micropores remains immobile. It has often been assumed that micritized grains containing water may be able to “short circuit” the resistivity measuring current. While this is an attractive concept, no irrefutable evidence was presented confirming such effects from SCAL data on the core plug-scale. This may be due to 1) Limited amount of cores analyzed. 2) Plug orientation did not favor a “short circuit” along a possible preferred orientation of micritized grains. 3) Masked by other effects – wettability.
It is worth noting that the three most likely methods – c, d and e all rely on a dual porosity structure. In the case of fractured and layered formations, the large pore sizes (fractures being regarded as of infinite pore size) are physically separate from the smaller pores while in the case of dual porosity systems the two pore sizes are in intimate juxtaposition.
Traditionally the “Transition Zone” is the volume of rock just above the free water level where water saturations are high as a result of limited oil capillary entry into the smaller pores. No clear consensus of how to define the upper limit of the “Transition Zone” could be determined and the term is thus only loosely defined.
The three main causes of low resistivity pay all rely on limited entry of oil into the smaller pores. Thus it may be expected that higher in the column (above the “Transition Zone”) oil would gain access to the smaller pores – and both oil saturations and resistivity would rise. Thus LRP is a “Transition Zone” phenomenon and would not be expected in identical rock higher in the column. Several presentations also addressed the concept of co-existing pore systems. Each pore system is percolating and posses its own entry pressure and transition zone. The system with the larger pores (macro-pores or vugs) would have very low entry pressure and a very short transition. This system would also dominate permeability and fluid flow. Hence, even when the meso and micro-pores are at the base of their transition zone, it is possible the macro-pores are at irreducible saturation and contribute to oil flow at very low water-cut.
Similarly, because of the higher buoyancy pressure as a result of the greater density difference between gas and water LRP is only rarely encountered in carbonate gas reservoirs.
This is defined as LRP where a traditional water saturation calculation using the resistivity log gives misleadingly high water saturation. A number of methods of overcoming this were proposed:
- Different Archie parameters – The solution to the LRP problem may simply be that a or m values appropriate to normal reservoir rock are inappropriate in the case of LRP and a well conceived SCAL program may be a sufficient solution. (See also section “The phenomenological nature of the Archie equation”.)
- A non-Archie resistivity approach may be taken – by for example dividing the pore space into different pore type categories and defining separate resistivity/saturation relationships for each category.
- NMR may be used to define the pore structure and so provide input into a non-Archie resistivity method.
- NMR may be used to make a direct measure of the formation fluids. For this to be quantitative, a means of demonstrating that invasion has not occurred or for accurately correcting for it must be found. Given that in LRP formations the oil generally resides in the largest pores and is thus highly fugitive invasion may be expected to render this technique invalid in many cases. Nonetheless, flushed zone measurements of flushed zone saturation such as NMR and dielectric (below) can be useful for evidence of oil, a signal to look further.
- Dielectric measurements have the potential to provide non-Archie estimates of saturation – but are subject to similar concerns to those of the NMR about the possible extent of invasion.
- If the LRP flows essentially dry oil then the water must be largely immobile. Consequently, a core carefully cut using OBM may be expected to maintain the formation water saturation. The Dean Stark technique then may be employed to directly measure the water saturation.
- Pulsed neutron logs may provide a non-Archie method of water saturation evaluation – either in open hole if invasion problems can be overcome – or in cased hole after completing the well. (But see section on filtrate dissipation in LRP formations.)
- If (and only if) layering is the cause of LRP then a vertical resistivity measurement – such as is provided by some recently introduced vector resistivity tools – has the potential to more accurately quantify the hydrocarbon saturation. Layering will cause a strong anisotropy between the horizontal and vertical resistivities that can be readily detected.
- Conductive minerals may be detected and quantified by numerous (frequently core based) methods. These are not considered further since conductive minerals are not a major factor in the case of Carbonate reservoirs.
This is defined as LRP where the water saturation as calculated from resistivity is broadly correct – and the issue is to demonstrate that oil is the mobile phase. The following solutions were proposed:
- Model the formation taking proper account of relative permeability effects. Formations with multimodal pore size distributions frequently show remarkably low water relative permeabilities up to surprisingly high water saturations due to relative permeability hysteresis. i.e in the transition zone the imbibition kr curve should start from the actual Swi for that height from the primary drainage curve. The imbibition curve relevant higher up the oil column should not be used.
- If oil is the mobile phase then a careful look at the shallow resistivity reading will show invasion – exactly as for normal pay. The separation between the different resistivity depth readings will be less than normal because of the low volume of mobile oil and the effect may be further masked by uncertainty in the effective salinity of the invaded zone. This salinity will be some unknown average of the mud and formation water salinities. If care is taken to use mud of the same salinity as the formation water this latter effect may be removed and it is easier to get a clear mobile oil signature from the shallow resistivity curve.
- To summarize: If mud filtrate is more saline than formation water:
- If formation contains movable oil: Rxo < Rt. If formation is wet: Rxo < Rt.
- If mud filtrate has the same salinity as formation water:
- If formation contains movable oil: Rxo < Rt. If formation is wet: Rxo = Rt.
- If mud filtrate is less saline than formation water:
- If formation contains movable oil: Rxo < Rt. If formation is wet: Rxo > Rt.
- Hence, it may be that the best solution is to use fresher mud.
It is normal to use different Archie parameters for different layers and/or rock types. Even in the event of a strong variation of “m” with porosity a sufficiently detailed subdivision can always be found that allows “a” and “m” values to be used that are in agreement with the core measurements.
In LRP the water saturation is always high – and this means that of necessity Archie is operating in a region where “n” is of limited impact. Archie can reconstruct the core-measured resistivity response whenever the “n” curve as seen on a standard log-log RI vs Sw (n) plot is sufficiently straight. A straight line can approximate any curve over a sufficiently short interval. As noted above given the high water saturations expected in most LRPs it is generally the case that the “n” curve is sufficiently straight – over the limited saturation interval of interest – that an “n” value can be found that reconstructs the SCAL measured resistivity response.
Thus almost by definition Archie “a”, “m” and “n” values can be found that reconstruct the SCAL core measurements and consequently the Archie equation has sufficient generality that it can still be used to provide accurate saturation calculations as long as suitable electrical parameters are used.
Of course, such a phenomenological approach to the Archie equation cannot provide physical insights into the true cause of the problem. For this other non-Archie approaches can be applied — for example methods involving the explicit subdivision of the pore space into differing categories.
While drilling, water from the mud filtrate usually enters the formation, and changes the salinity of the water trapped in even the smallest pores – since there are no capillary effects to hinder its entry into them – even if this entry is only diffusive. Once the casing has been run any filtrate that has displaced oil may slump away and the oil return until the formation has returned to its equilibrium saturation as defined by the saturation height function. However once slumping is complete there is no longer any fast mechanism for the salinity of the water trapped in the smallest pores to return to that of the formation. Remember that once oil has returned to the major pore network it will greatly reduce the water relative permeability – and diffusion rate is a strong function of relative permeability. The consequence is that mud filtrate may be effectively trapped in the smallest pores with only a very slow diffusive means of dissipation.
In water bearing formations the water relative permeability will not be reduced and diffusive dissipation will not be hindered as in LRP. In conventional reservoirs, the volume of filtrate held in the smaller pores will be less, and less diffusion will be required. Therefore, LRP reservoirs may have the potential for having unusually long filtrate dissipation times – and indeed examples were presented of filtrate still being present twenty years after drilling.
In order to use PNL logs in cased hole as a non-Archie means of quantifying the water saturation in LRP it is thus necessary to find a method to demonstrate that the filtrate has effectively dissipated. No reliable methods of doing this were presented. Pulsed neutron CO logging is a means around this if saturations have returned to equilibrium, while salinity remains affected.
Certainly without electrical SCAL data and sometimes even with it there is a risk of overlooking LRP if the only method for determining saturation is from the resistivity log. It is thus necessary to use further methods for indicating the presence of hydrocarbons. Some suggestions were:
- Mud logging, incorporating new geochemical sensors and mass-spectroscopy technology.
- Capillary pressure measurements on cuttings – done on the rig. Multimodal porosity distributions would then suggest the possibility of LRP intervals.
- Wireline formation pressure tests would be expected to show an oil gradient in LRP intervals and so provide a useful backup for resistivity data. (See also “Transition Zone” below.)
- Samples taken from a wireline formation tester may give an indication of the mobile phase – though examples of misleading results were given if a conventional probe was used. Particularly in layered LRP intervals, the fluid produced may depend whether the probe set on a fine or coarse-grained layer. A dual packer arrangement may give more reliable results.
New LWD PNC logs show promise and should be evaluated as direct measurement of saturation.
A key requirement for success is that the sensor reaches the formation prior to significant invasion.
Evidence was presented showing the risks of misleading pressure gradients and contact definition due to mixed wettability and deep mud filtrate invasion. These effects have been successfully modeled in a dynamic simulator and must be taken into account in addition to the “normal” issues of super-charging in low permeability carbonates.
- The pressure gradient seen in close proximity to the free water level can be a function of the mud type used for drilling. If OBM mud is used it is presumed that the mud filtrate creates preferential contact with the oil phase – and an oil gradient may be seen – whereas if WBM is used the mud filtrate can create preferential contact with the water phase – and a water gradient may be seen.
- In extreme cases, the mud filtrate may make effective contact with the non-miscible phase (via a capillary meniscus) and the pressure gradient may be offset from its true value.
Pore size distribution may be derived from NMR T2 in the case of single wetting phase and very shallow invasion. Mixed wettability and/or the presence of two phases i.e. remaining oil and mud filtrate (in an invaded zone) or residual oil and water in a flushed zone disturbs the determination of pore size distribution. The NMR responds to pore size, not pore throat size, which controls permeability. In carbonates, there is a limited relationship between pore size and pore throat size, permeability prediction from NMR is therefore difficult. There is work going on to try to derive pore size from restricted diffusion.
Efforts to derive fluid saturations from NMR require calibration with lab NMR. It is important to take into account the effects of temperature and salinity on the NMR measurements. Laboratory NMR experiments should therefore be conducted at reservoir conditions.
With a new generation of NMR logging tools now available, there is a need to investigate relations between different generation NMR tools and their interpretation models. Future direction for research includes laboratory work on dynamic NMR and 3D flow imaging.
Enables visualization of grain sizes in 3D across the range from 2.5 micron -5 mm.
Using the 3D images the following parameters can be modeled:
- Rock fabrics
- Pore size distribution
- Pore throat distribution
- NMR “T2” model
- K3D (lattice Boltzman)
- Drainage Kr + Pc
- Pore network model (local)
- Macro-pores and Meso-pores (but not Micro-pores)
- Resistivity modeling (without contribution from Micro-pores)
This technology provides a “digital” alternative to conventional core analyses. It may be used to investigate rock from reservoirs with unusual log responses. Results will be available much quicker than SCAL, which may take more than 1 year for a complete cycle of Petrophysical measurements.
The way forward focuses on obtaining larger images: higher resolution and larger samples (to capture heterogeneity) and tie in with Effective Medium Models (EMM).
It was agreed that it would be time-consuming to establish such a Consortium and that Confidentiality concerns over data and competitive edge issues might well limit its effectiveness.
Better to hold another Topical Conference or Workshop; Same place – Next year?