|Theme||Visible||Selectable||Appearance||Zoom Range (now: 0)|
The use of one-dimensional (1D) electromagnetic (EM) inversion for well placement, reservoir mapping, and planning of multi-lateral wells is now quite common, but it is limited because it assumes continuity of resistivity in all directions except above and below the wellbore. Where formation and fluid distributions are not simple layer cake structures, 1D inversion does not reveal the lateral distribution of target zones. Structures with lateral variability require mapping of the geology in three dimensions (3D). Historically this has been done based on seismic data, but this has limited resolution. 3D EM inversion allows more refined well placement and reservoir mapping. Multi-frequency 3D inversion results in adjacent multi-lateral wellbores can be verified by superimposing the results to identify the same formation and fluid boundaries and can be used in re-planning trajectories of subsequent laterals to consider lateral changes in the position of resistivity boundaries.
This paper presents results from a complex turbidite reservoir, developed using a multi-lateral well with three production branches. 1D EM inversion used in real-time displayed the vertical distribution of the sands, but did not indicate lateral variations, which are expected in this geological environment. Real-time ultra-deep azimuthal resistivity images provided a qualitative assessment of the lateral distribution of the target sands, indicating that the lateral position of the target would need to be considered when planning the second lateral. 3D inversion of memory data was performed within 48 hours of drilling the first lateral to understand the complex sand distribution. The distribution of the geobodies identified from the inversion results was used to re-plan the subsequent laterals for optimal placement and enhanced reservoir contact. All three laterals were inverted independently, providing overlapping datasets that showed similar structural features, giving confidence in the results.
In a complex turbidite reservoir with discontinuous boundaries and significant lateral variations, true reservoir understanding requires 3D inversion of EM data. Multi-lateral wells provide an ideal opportunity to gain high confidence in the inversion workflow with repeatability of the results across multiple overlapping datasets. The reservoir understanding brought by this data enabled more sophisticated well planning and increased reservoir exposure in the subsequent laterals. When available in real-time, 3D EM inversion will facilitate azimuthal, as well as inclination, changes in a well path for optimal placement.
The improved reservoir cognizance provided by 3D inversion of ultra-deep EM data allows lateral variation in the position of geobodies to be considered when planning multi-lateral wells. Previously, 1D inversion only allowed TVD changes to be considered. As many complex geological scenarios are 3D in nature, 3D EM inversion allows changes to the planned azimuth of a well path to also be considered for optimal well placement.
Maraj, Priya (BP) | Huber, Ken (BP) | Itter, David (BP) | Nelson, Jeff (BP) | Rabinovich, Michael (BP) | Youngmun, Alex (BP) | Antonov, Yuriy (Baker Hughes) | Mejia, Luis (Baker Hughes) | Martakov, Sergey (Baker Hughes) | Pazos, Jhonatan (Baker Hughes) | Small, Austin (Baker Hughes) | Tropin, Nikita (Baker Hughes)
A significant oil resource exists within the Schrader Bluff and stratigraphic equivalent West Sak reservoirs located on the central North Slope of Alaska. The Schrader Bluff resource is under development in the Kuparuk River Unit, Milne Point, and within the Prudhoe Bay Unit. These areas have developed reservoirs with oil viscosities up to 200 cp under waterflood and viscosity-reducing miscible gas injection. A grass roots penta-lateral gas-lifted producer was drilled and completed to unlock untapped oil in the northern portion of the Polaris S-Pad Schrader Bluff viscous oil reservoir in Prudhoe Bay, Alaska. This producer was the first multilateral well drilled in the Orion and Polaris Schrader Bluff reservoirs in ≈8 years. The team worked to unlock surface and subsurface opportunities, delivering a substantially lower cost of access than planned. Production rate from Schrader Bluff reservoir wells has a positive correlation with contacted net sandstone. Optimal well placement in each of the five laterals was key to the delivery of a successful project. A novel approach to geosteering was used to maximize net sandstone exposure, utilizing deep azimuthal resistivity and real-time user-guided multilayered inversion modeling.
The total porosity obtained from nuclear magnetic resonance (NMR) logging-while-drilling (LWD) data is typically not affected by lateral motion of the NMR-LWD tool; however, some other deliverables (e.g., bound water, movable fluid, permeability, viscosity) might be affected. This paper introduces a data-based lateral-motion correction (LMC) that uses a four-parameter function to quantify and correct potential lateral-motion effects.
The objective of the LMC is to improve the accuracy of the final T2 porosity distribution. The LMC extends the operational range of the NMR-LWD method and enables more advanced petrophysical NMR applications. In addition, the LMC approach can be used to quantify the lateral-motion effect and to mark intervals where the motion effect is too severe to be fully corrected.
The LMC was developed by analyzing the potential lateral-motion effect on numerous NMR-LWD data in combination with numerical simulations. By using drilling dynamic simulations of a complete bottomhole assembly (BHA) under realistic drilling conditions, motion paths of an NMR sensor were calculated, which were then used to simulate NMR signals. Actual data and the NMR simulation indicate that lateral-motion effects can be adequately described by a four-parametric function. Two parameters describe an exponential decay, while the other two parameters describe a periodic variation of the amplitude. The motion effect function was integrated into the forward matrix of the NMR joint inversion, and a non-linear optimization algorithm was used to determine the four motion parameters and, if present, to compensate for lateral-motion effects.
Although motion paths are typically complex, the motion periodic characteristic relates to the revolutions per minute (RPM) of the BHA. The amplitude of the motion mainly depends on the drilling regime (e.g., from “smooth” drilling to whirl), the gap between the BHA stabilizers and the borehole, and the borehole inclination. Numerical simulations show that the NMR motion effect is small-to-negligible if an NMR-LWD tool has a low magnetic field gradient and has adequate stabilization.
The correction method was tested on synthetic and real NMR-LWD data from more than 30 runs with different realizations of lateral motion. The approach is robust and works for all data sets. The magnitude of the lateral motion effect is reliably shown by the implemented multiple quality control indicator. Some examples of synthetic and actual NMR data with and without LMC are included.
The inversion that includes the LMC improves the quality of the T2 spectrum, which is important for standard and advanced NMR applications, such as accurate calculation of volumetrics (e.g., movable fluid, bound/irreducible fluids) and pore size distribution, as well as improved estimation of fluid viscosity and formation permeability.
Abstract Objective /Scope: The objective of the study was to examine the effectiveness and results of geosteering horizontal wells in the Green River Formation of the Uinta Basin in northeast Utah using a deep directional resistivity tool. Using real-time Logging While Drilling (LWD) measurements, geosteerer maximized exposure in open and marginal lacustrine sands while simultaneously provided imaging information that described the horizontal target zone thickness and identification of faults or folds. This tool provided geosteerers data to map the subsurface without drilling additional vertical delineation wells. Methods /Procedures /Process: Steering within the lacustrine sands of the Uinta Basin is often difficult because the primary target sands can be laterally discontinuous and below seismic resolution where seismic data is available. Changes in channel direction or thickness combined with local formation dips or faulting create complex log signatures which make geosteering within target sands difficult. Using a deep directional resistivity tool that imaged the resistivity contrasts at bed boundaries of the target zones, allowed geosteerers to remain in the higher resistivity reservoir rock with minimal directional changes. Inversion information enabled the measurement of the top and base of the target zone without drilling up or down which also mitigated a risk of exiting the target zone. Results /Observations /Conclusions: By using azimuthal resistivity information while drilling, the horizontal wells drilled were geosteered more effectively in the highest quality reservoir rock with lateral lengths of one mile or greater. In addition, an increase in the lateral length can potentially support drilling multiple wells from the same location which has a smaller surface footprint in a development program. The real-time LWD data allowed the laterals to be steered in zone for longer and the data was used post well to create new estimations of sand thickness in areas with sparse offset well data. Using the inversion technique of the deep directional resistivity tool, formation thicknesses were interpreted to be thinner or thicker than seen on the offset well data and faults were more easily identified. Applications /Significance/Novelty: Using an advanced inversion technique in real-time enabled the mapping of the reservoir thickness while drilling, and because the inversion isn't dependent on a prior model, a more accurate representation of the stratigraphy can be mapped with confidence. Based on the information provided by the inversion, it was possible to map target thickness as the borehole drilled away from well control. In addition, after encountering a fault, the deep reading inversion allowed for a smooth re-landing back in the target zone. The data was easily visualized and shared over the internet in real-time allowing decisions to be made quickly without slowing the drilling operation.