Unconventional resources merit strong attention in the oilfield today. Great value is placed on understanding the amount and viscosity of the oil which in turn guides the production strategy. We demonstrate a workflow employed in Middle East carbonate reservoirs containing intermediate to heavy oil. The workflow combines dielectric and nuclear magnetic resonance (NMR) log data, and it is complemented by laboratory measurements. The applicable range of viscosities is tens to thousands of centipoises. Integrating dielectric and NMR log data provides a more accurate determination of saturation and viscosity. The workflow does not require any change to the processing of the raw data and is simple to implement. We demonstrate the workflow using log data along with correlations from lab measurements on core and bulk fluids. The workflow solves three key challenges for heavy oil analysis. The first challenge is the overlap of heavy oil and bound water NMR signals, compounded by a radial variation of saturations near the wellbore. Conveniently, the physics of dielectric and NMR measurement results in depths of investigation of 1 to 4 inches. Dielectric-based analysis provides the total water-filled porosity, which improves the NMR-based saturation and fluid characterization. The second challenge comes from the effect of restricted diffusion of water molecules in pores. We show that a restricted diffusion model is required to obtain accurate saturations and viscosity. The third challenge is the reduced NMR sensitivity to viscosity for heavy oil. Lab measurements on core and bulk oil samples provide the parameters required to link NMR to viscosity. We describe a workflow based on dielectric and NMR log data for a more accurate determination of fluid saturations and oil viscosity. The workflow takes into account a) the combined dielectric and NMR data; b) the carbonate restricted diffusion effect inherent in the NMR response for carbonates; and c) laboratory validation and interpretation support using core plugs in wells where the dielectric data was not available. We use log data covering a range of heavy oil viscosities to demonstrate the workflow in four wells.

Summary Textural and wettability variations are two main vexing problems in carbonate reservoirs. Adding heavy oil presents an incremental challenge to the reservoir description. Ineffectual analysis often leads to poor completion and high water production. Integrating multifrequency-dielectric and nuclear-magnetic-resonance (NMR) logs leads to improved carbonate-reservoir evaluation. Dielectric and NMR logs have similar depths of investigation, both probing the flushed zone. It is difficult to distinguish bound water from heavy oil in rocks from NMR tools alone. The addition of accurate water-filled-porosity (ϕW) measurements from the dielectric log allows us to differentiate the fluids. Using ϕW as a reference and accounting for restricted molecular diffusion in small pores, NMR analysis on the basis of diffusion relaxation provides more-accurate oil viscosity and fluid volumes. The computation of viscosity from NMR assumes that the oil relaxation is dominated by bulk relaxation; oil-wetting of pore surfaces causes viscosity to be overestimated. Conversely, the wetting state can be inferred if the viscosity is known. Dielectric dispersion with frequency provides a measure of the water-phase tortuosity MN_De, which in turn is affected by texture and wettability. We compared the in-situ NMR signal of oil to the surface NMR measurement of a bulk-oil sample. We also identified a correlation between MN_De and the Archie cementation exponent, mainly driven from NMR. This leads to interesting conclusions about texture and wettability, which are found to be coherent with the existing reservoir data. Together with the viscosity profile, the comparison between the pore-fluid volumes in the flushed zone and the bulk volume of water from deep resistivity is used to identify the zones of movable oil and to estimate residual oil saturation. The method was applied to data sets from three wells in a shallow carbonate reservoir. The estimated viscosity was validated by pressure/volume/temperature (PVT) analysis. In two wells, the prediction of movable oil and water zones was confirmed by downhole test results. In the third well, the free-water zone was confirmed by production logging.

The Greater Burgan field has been producing since 1946 from a series of highly permeable Cretaceous reservoirs. Recently, a series of more complex reservoirs has been reassessed using advanced logging and wireline formation tester (WFT) technologies. The techniques employed in the reassessment include fluid-quality (viscosity and presence of tar) mapping using nuclear magnetic resonance (NMR) log data and shallow invasion measurements using multifrequency, multispacing dielectric data. In addition, the dielectric logs provide a direct measurement of the Archie m exponent in water zones. Improvements in formation evaluation achieved by integrating these results with conventional logs included better differentiation of moveable from residual hydrocarbon, identification of variations in formation water salinity, and maps of oil-quality variation versus depth and across the field. These techniques were applied to three case studies. In the first case study, formation evaluation was conducted in an Upper Cretaceous carbonate formation of unknown water salinity. The combination of dielectric logs and NMR enabled identification of water-bearing and residual oil zones where formation water salinity could be determined. The analysis revealed increasing water salinity with depth. Dielectric logging also provided a direct evaluation of Archie's m exponent in the water zones, in the absence of special core analysis. The NMR highlights variations in oil quality from one well to another. In the second case study, a viscous oil layer located in the middle of a water zone in a Middle Cretaceous reservoir was evaluated. Moveable oil was identified by radial oil saturation variation close to the borehole detected by the dielectric log measurement. The interpretation was verified by the first oil sample ever recovered in this layer. In the final case study, in a Lower Cretaceous reservoir, dielectric measurements provided accurate estimates of residual oil saturation required for planning enhanced oil recovery projects. The results obtained from the application of the multifrequency dielectric dispersion and diffusion-NMR as confirmed with WFT sampling bring new insight to the evaluation of challenging formations within the Burgan field.

The Jurassic Najmah-Sargelu of west Kuwait can be thought of as a "hybrid" between a conventional and an unconventional reservoir. These systems form an increasingly important resource for operators, but their performance is unpredictable because matrix permeability is in the micro-Darcy range and production depends on natural fractures. Success depends on how well the static models are aligned to the dynamic production, and the effectiveness of a fit-for-purpose multistage completion on project economics. In this work we present our lessons learnt in production modelling these reservoirs and the coupling between reservoir simulation and the discrete fracture network (DFN). Our reservoir models were constructed using a highly integrated approach incorporating data from all scales and disciplines (drilling, geophysical, geological, reservoir and production) and the production simulations were run using dual porosity and black oil models. As expected, the DFN played a key part of this effort. An iterative approach was used to adjust the DFN so that it was consistent with production observations. However, in all cases care was made to ensure the new DFN honoured the seismic, geological, well log and drilling data from which it was generated. Final, smaller adjustments were made to the simulation model at the log scale to match PLT data. We used uncertainty analysis to run hundreds of simulation cases and found that the character of the natural fractures is quite well imprinted in the observed production data, particularly pressure buildup data. This gave us a better understanding of whether the natural fractures are diffuse and laterally extensive away from the wellbore or if they are localized close to the wellbore. Where reservoir simulation history matches inferred laterally extensive natural fractures, an good correlation was obtained with the natural fracturing from the DFN. This correlation was poor where natural fracturing was confined to a smaller depth interval (as observed from PLT), and is a result of the limitation in seismic resolution to resolve these natural fractures. The lessons learnt from our work helps towards improved understanding of production mechanisms of these reservoirs and their natural fracture networks. This, together with higher resolution azimuthal seismic, advanced wellbore characterization data and multistage completions are the desired key ingredients for technically enhancing production in these reservoirs.

Resistivity measurements in high angle and horizontal (HaHz) wells are sometimes inappropriately labeled as 'misleading' when events such as early water breakthrough are observed in spite of measuring high resistivity along the well trajectory. Applying appropriate interpretation techniques that account for geometric effects resolves the interpretation issue but is time consuming. For recent HaHz campaign in North Kuwait, a workflow was required in near real time to optimize completions.In order to efficiently complete a horizontal well it is important to identify the presence of any free water and permeability variations along the well trajectory. Zonal isolation is required in case free water is encountered and completion mechanisms such as Inflow Control Devices (ICDs) are required for effective linear compartmentalization. We present here a workflow which was used to guide completions in the siliciclastic Burgan reservoirs in Raudhatain field.The workflow uses a new-generation logging-while-drilling (LWD) tool that provides sigma and spectroscopy measurements in addition to triple-combo measurements. A comparison of sigma measurements acquired during and after drilling allows identification of any moveable water along the well trajectory. This technique is valid only in wells drilled using oil-base mud and has been used to design zonal isolation to avoid water breakthrough. Spectroscopy measurements and conventional logs are used to derive k-Lambda (matrix) permeability to identify the permeability profile along the well trajectory. This information is used to identify permeability variation to optimize completions.We present several case studies in the sequence in which the workflow evolved. The first case study highlights the limitations of using only triple combo LWD measurements to decide completion intervals as water production was seen across high resistivity zones on production logs. The other case studies have advanced measurements acquired while-and post-drilling to aid in completion optimization decisions.