The paper demonstrates how to establish a compositional gradient in Dhirubhai-26 (MA) retrograde gas-condensate reservoir situated in a deep water territory of KG Basin, India, with the help of real-time Downhole Fluid Analysis (DFA).The reservoir consisted of an oil rim sandwiched by a large gas cap and bottom water. Very low contamination samples were captured downhole, using a Wireline Formation Tester (WFT), in Single Phase PVT bottles. This success was achieved through the combined use of DFA results, PVT data and a tuned Equation of State (EoS).
PR Peneloux (T) EoS was tuned to the PVT data and used to populate the PVT properties of the reservoir. The resultant fluid composition versus depth was correlated to the DFA results to verify the EoS simulated compositional gradation. Subsequently a reliable GOC was established using the combined result of WFT pretest derived gradients, PVT property derived gradients and saturation pressure curves. Establishing the existence or non-existence of compositional gradient and reservoir compartmentalization in this reservoir were key prerequisites for the formulation of an optimum Field Development Plan (FDP) and estimation of the associated financial implications.
The methodology adopted in this paper is applicable to any gas-condensate reservoir that exhibits similar PVT properties. It would help to evaluate a more accurate in-place reserve estimate and reduce the reservoir risk, which in turn would lead to an optimum FDP (that would maximize the oil recovery and returns).
The paper highlights the value of wireline conveyed DFA tools and latest wireline sampling methods coupled with a tuned EoS for an accurate PVT description of complex reservoir fluids. It cannot only detect the presence of compositional gradients at an early time but also help refine the PVT model to establish the gradation in composition, thus providing a cost-effective solution for deep water ventures.
Formulation of an optimal field development plan mandates a proper PVT description of reservoir fluids, among many other requirements. Recently it has been observed in quite a few publications10,12 that reservoir fluids from a single hydrocarbon (HC) column often exhibit complex compositional behaviors due to the interplay between various forces like gravitational, thermal and chemical, just to name a few. While composition variation versus depth is more common, areal variation has also been reported4. Prevalence of these forces, either independently or in combination, may result in a varying HC composition with depth. Lars et. al.3, in their treatise have reviewed all the major mechanisms responsible for a compositional gradient in hydrocarbon reservoirs and existing PVT models for the different mechanisms, published till date. In their review they have concentrated only on the Zero-Mass flux models. However several published papers5,6,7,8,9 exist where mathematical models for convection, diffusion and mass flux have been formulated.
Normally, for lighter oils (>35°API), compositional grading is caused by gravity induced component migration that balances the gravitationall and chemical forces2. For moderately heavier oils (20–30°API), asphaltene segregation2 is a primary driving force, and in the case of heavy oils (< 20°API) it is often the aftermath of loss of light ends through bio-degradation2.
The existing practice for investigating these complexities of reservoir fluid distribution includes laboratory experiments performed on multi-depth samples of representative formation fluid. But often a conclusive quantification of this fluid behavior may not be realized due to the lack of a sufficient number of samples versus depth. While specialized mathematical/statistical analyses10 of the static pressure surveys help in revealing possible cases of compositional gradient, induction of latest DFA tools have appended real-time compositional fluid analysis and fluid scanning to the list of gears that help ascertain the presence of compositional gradient with solidarity17,18. The ability of DFA tools to estimate compositional and PVT properties, in real-time, speeds up the diagnostics and also allows contamination monitoring (which is indispensable for a representative HC sample from a well drilled with OBM).
Once the mechanism of the compositional gradient is identified, the fluid system needs to be modeled using a compositional model. In certain cases a 3D Black-oil model2 may also be used as per Hanafy et. al., but the application is limited to a black oil system where there is only a slight variation in its heavy ends with depth. An accurate formation fluid composition, required for the EOS definition, can be obtained either from the compositional analysis of a low contamination bottomhole sample or through a recombined sample (recombined using the Equilibrium Contact Mixing Method)11.
The EoS model presented in this paper is based on the bottom-hole sample collected at a very low draw-down (approx. 8 psia) and thus was considered a representative sample from the test depth. The concept of Equilibrium Contact Mixing can be used in further simulation work to obtain a representative composition at GOC to represent the whole reservoir.
The final step in the exercise is running a composition versus depth simulation using a PVT simulator (based on the Schulte24 saturation pressure gradient model) and the tuned EOS. The fluid compositions and DFA results serve as hard data used to test the validity of the composition versus depth experiment. With the existing commercial PVT simulators, temperature gradient may be used as a means of fine tuning the composition-depth experiment results to match the lab/DFA results from the fluid sampling/scanning stations.
Optical spectroscopy based fluid analysis is an industry accepted technology widely used for discriminating between different types of fluids downhole, viz. oil, gas and water, based on the difference in their absorption peaks in the Near Infrared Region (NIR)13, as seen in Figure 2. Contamination of bottomhole liquid HC samples, due to miscible Oil Based Mud (OBM) filtrate, can also be estimated using the results from these tools14,15. Recent developments in DFA technology have enabled the real-time fluid characterization of reservoir fluids into five compositional groups (C1, C2-C5, C6+, CO2 & water) using NIR spectroscopy19,20 and differentiation between gaseous and liquid phase using the principle of fluorescence Figure 3.
Employing a suitable combination of DFA devices provides a timely analysis of the reservoir fluids, with robust results, and is invaluable for ensuring an appropriate 'Chain-of-Custody' for downhole samples. First-hand knowledge of the fluid characteristics while it is being sampled results in a big logistical reward - a more flexible sampling program which has the potential of significant cost savings in deep water territories.
The Live Fluid Analyzer13 (LFA)* and Compositional Fluid Analyzer19,20 (CFA)*, used in combination, provide both in-situ characterization of reservoir fluids and quantification of filtrate contamination in real time. In addition to that, the CFA, with its fluorescence detector, is an invaluable instrument to aid acquisition of downhole samples of a retrograde gas, under single-phase conditions. Fluid properties like GOR/CGR, derived indirectly from the NIR absorption characteristics of the flowing fluids, are very useful as OBM contamination level indicators as well.
The Dhirubhai-26 (MA) oil-field, situated offshore Andhra Pradesh, India is a deep marine reservoir located in an upper Jurassic to lower Cretaceous zone. It is a fluvial to lacustrine deposit with the principal reservoir sections dominated by areally restricted channel sands. Thin, laterally extensive alternations of siltstone and mudstone also constitute sections of the reservoir.