We develop a neural network workflow, which provides a systematic approach for tackling various problems in petroleum engineering. The workflow covers several design issues for constructing neural network models, especially in terms of developing the network structure. We apply the model to predict water saturation in an oilfield in Oman. Water saturation can be accurately obtained from data measured from cores removed from the oil field, but this information is limited to a few wells. Wireline log data are more abundantly available in most wells, and they provide valuable, but indirect, information about rock properties. A three-layered neural network model with five hidden neurons and a resilient back-propagation algorithm is found to be the best design for the saturation prediction. The input variables to the model are density, neutron, resistivity, and photo-electric wireline logs, and the model is trained using core water saturation. The model is able to predict the saturation directly from wireline logs with a correlation coefficient (r) of 0.91 and an error of 2.5 saturation units on the testing data.
Due to a surge in oil demand over the past years and its impact on prices, major heavy oil accumulations are gaining special attention in the Middle East. Knowledge of the fluid characteristics is critical to determine a suitable recovery process for heavy oil reservoirs. Oil viscosity is the property that most affects producibility and ultimate recovery in these settings, but it is the most difficult property to determine. This is due to the challenges associated with obtaining representative oil samples in increasingly heavy oil reservoirs. Additionally, the oil viscosities in these reservoirs exhibit large variations within the same formation. Hence, downhole in-situ measurements are key in establishing viscosities and their variation within the reservoir. Current practice in heavy oil accumulation is to extract oil samples through crushing of cores at surface, or sampling through changing the temperature of the near wellbore in the hope that the fluid will flow. These techniques provide viscosity ranges that help in uncertainty testing of various recovery processes but are only point-data measurements. Nuclear magnetic resonance (NMR) logging is also used to determine in-situ oil viscosity accurately in conventional light oil reservoirs. However, this technique has limitations in heavy oil settings because the extremely fast relaxivity of the heavy oil components results in very short transverse or relaxation time (T2) values that are not detectable by the NMR tool. The short T2 time also means that the heavy oil and bound-water signals, overlap resulting in an ambiguous oil signal. A novel technique combines data from 3D NMR and dielectric dispersion tools to determine in-situ oil viscosity. This was based on recently published NMR viscosity correlations. The primary output from the dielectric tool is the total water volume. Using this output combined with the 3D NMR data makes it possible to accurately derive bound-water, free-water, and oil volumes. These are necessary inputs into NMR-based viscosity correlations. This technique yielded continuous logs of heavy oil viscosities within the same viscosity range as measurements obtained from the core data.
In the past decade, Fiber-Optic (FO) based sensing has opened up opportunities for in-well reservoir surveillance in the oil and gas industry. Distributed Temperature Sensing (DTS) has been used in applications such as steam front monitoring in thermal EOR and injection conformance monitoring in waterflood projects using (improved) warmback analysis and FO based pressure gauges are deployed commonly. In recent years 1 significant progress has also been made to mature other, new FO based surveillance methods such as the application of Distributed Strain Sensing (DSS) for monitoring reservoir compaction and well deformation, multidrop Distributed Pressure Sensing (DPS) for fluid level determination, and Distributed Acoustic Sensing (DAS) for geophysical and production/injection profiling. For the latter application, numerous field surveys were conducted to develop the evaluation algorithms or workflows which convert the DAS noise recordings into flow rates from individual zones. The applicability of a new graphical user-interface has been expanded to include smart producers and injectors that allows the user to visualize (in real time), QC and evaluate the DAS data. Also, the evaluation methods for the use of DTS for warmback analysis have been significantly improved.There are still improvements to be made in enabling Distributed Sensing infrastructure, such as handling and evaluation of very large data volumes, seamless FO data transfer, the robustness & cost of the FO system installation in subsea installations, and the overall integration of FO surveillance into traditional workflows. It will take some time before all these issues are addressed but we believe that FO based applications will play a key role in future well and reservoir surveillance.In this paper we present a recent example of single-phase flow profiling using DAS. The example is from a long horizontal, smart polymer injector operated by Petroleum Development Oman (PDO).
There have been a number of major heavy oil discoveries in Oman in recent years. In order to devise efficient and cost effective recovery mechanism careful and detailed subsurface understanding of these fields is critical. To this end, petrophysical understanding plays a critical role, as it represents a basic building block of the static and dynamic models. The field under study is a fractured carbonate reservoir with high viscous oil. It is believed that this reservoir has gone through various cycles of drainage and imbibition. Thus, in addition to the complex geology, understanding of fluid distribution and fluid mobility are among major challenges that detailed petrophysical evaluation needs to address. Understanding these parameters will help determine the feasibility of the recovery methodology to be adopted. This paper details a novel petrophysical workflow that integrates 3D NMR, multi array/multi frequency dielectric measurements, borehole images, and core analysis. The core analysis focused on capillary pressures, Dean-Stark, and rock typing. Fracture studies included detailed image analysis and extensive fall off test for understanding the nature and distribution of the fracture network in the reservoir. The wealth of well data coupled with geological and dynamic data reduced the overall reservoir properties and fluid distribution uncertainties.Dielectric data provided resistivity independent saturations validated by Dean-Stark data. Combining dielectric and 3D NMR data allowed better formation characterization and fluid type evaluation and their present day distribution. Additionally, this combination indicated that water is not at an irreducible state in the reservoir. This was supported by the core saturation height function which indicated that present day saturation should be much higher if the reservoir was in drainage mode. These results were crucial to evaluate development options, underlying uncertainty/risks of this reservoir, and design optimum future data acquisition requirements. TX 75083-3836, U.S.A., fax +1-972-952-9435
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.