To estimate Rt under a variety of different logging conditions and in different formations, a simple three-parameter, step-profile invasion model is often used. This model consists of a flushed zone of resistivity Rxo and a sharp boundary at diameter di, with the uninvaded zone of resistivity Rt. Three independent, borehole-corrected resistivity measurements with appropriately chosen depths of investigation contain enough information from the formation to reliably solve for Rt using this model. Measurements with the following features should be chosen: small, correctable borehole effects; similar vertical resolutions; and well-distributed radial depths of investigation--one reading as deep as practical, one very shallow reading, and one intermediate reading. In conductive muds, the Dual Laterolog (DLL) Resistivity– Rxo combination tool provides simultaneous measurements suitable for evaluating Rt, Rxo, and di. It should be said that the value of Rt in a given bed is an interpreted parameter, and is almost never measured.
Traditional methods of monitoring reservoir behavior, including reservoir simulation and history-matching with production rates and pressure, can produce nonunique solutions for reservoir behavior in the interwell regions. In some instances, the uncertainty can be significant, and additional information is needed to optimize production and improve estimates of ultimate recovery. In many cases, the effect of the changing reservoir pressure and/or saturation on seismic data can be used to map the changing pattern of these reservoir properties by obtaining seismic data repeatedly during production of the reservoir. With care, seismic data obtained for other purposes (such as regional exploration) can sometimes be used for time-lapse seismic monitoring, but new data are often obtained from seismic experiments designed particularly to monitor the reservoir. The desire to minimize differences in acquisition parameters between surveys has led, in some cases, to permanent installation of sensors in the oilfield.
The ability of seismic reflection technology to image subsurface targets is possible largely through the geometry of sources and receivers. A method similar to triangulation is used to place reflections in their correct locations with (more-or-less) correct amplitudes, which can then be interpreted. The amplitudes are indicative of relative changes in impedance, and the seismic volume can be processed to yield impedances between the reflecting boundaries. These traces were recorded at different offset distances, and the travel times for seismic waves traveling to and from a given reflecting horizon varies with that distance (Figure 1). If the overburden through which the seismic waves pass is of constant velocity, then the time-variation with distance is a simple application of Pythagorean geometry, and the shape of the reflector on a seismic "gather" of traces is hyperbolic.
Integral transforms are useful in solving differential equations. A special form of the linear integral transforms, known as the Laplace transformation, is particularly useful in the solution of the diffusion equation in transient flow. The following fundamental properties of the Laplace transformation are useful in the solution of common transient flow problems. For the Laplace transform to be useful, the inverse Laplace transformation must be uniquely defined. In this operation, p(t) represents the inverse (transform) of the Laplace domain function, .
Resistivity is the one of the most difficult formation parameters to measure accurately because of the complex changes that occur during and after drilling a well and that may still be occurring during logging. The various components of the downhole environment may have strongly contrasting resistivities, some of which cannot be measured directly, and their physical dimensions may not be readily available. Figure 1 shows an idealized relationship of the main environmental components. There is no direct measurement of Rt. It must be inferred from the multiple-depth resistivity measurements.
The new-generation oil-base mud (OBM) microresistivity imagers provide photorealistic high-resolution quantified formation imaging. One of the existing interpretation methods is based on composite processing providing an apparent resistivity image largely free of the standoff effect. Another one is the inversion-based workflow, which is an alternative quantitative interpretation, providing a higher quality resistivity image, button standoff, and formation permittivities at two frequencies. In this work, a workflow based on artificial neural networks (NNs) is developed for quantitative interpretation of OBM imager data as an alternative to inversion-based workflow.
The machine learning approach aims to achieve at least the inversion-level quality in formation resistivity, permittivity, and standoff images an order of magnitude faster, making it suitable for implementation on automated interpretation services as well as integration with other machine learning based algorithms. The major challenge is the underdetermined problem since OBM imager provides only four measurements per button, and eight model parameters related to formation, mud properties, and standoff need to be predicted. The corresponding nonlinear regression problem was extensively studied to determine tool sensitivities and the combination of inputs required to predict each unknown parameter most accurately and robustly. This study led to the design of cascaded feed-forward neural networks, where one or more model parameters are predicted at each stage and then passed on to following steps in the workflow as inputs until all unknowns are accurately obtained.
Both inverted field data sets and synthetic data from finite-element electromagnetic modeling were used in multiple training scenarios. In the first strategy, field data from few buttons and existing inversion results were used to train a single NN to reproduce standoff and resistivity images for all other buttons. Although the generated images are comparable to images coming from inversion, the method is dependent on the availability of field data for variable mud properties, which at the moment limits the generalization of the NNs to diverse mud and formation properties.
In the second strategy, we utilized the synthetic responses from a finite element model (FEM) simulator for a wide range of standoffs, formation, and mud properties to develop a cascaded workflow, where each stage predicts one or more model parameters. Early stages of the workflow predict the mud properties from low formation resistivity data sections. NNs then feed the estimated mud angle and permittivities at two frequencies into next stages of the workflow to finally predict standoff, formation resistivity, and formation permittivities. Knowledge of measurement sensitivities was critical to design the efficient parameterization and robust cascaded neural networks not only due mathematically underdetermined nature of the problem but also the wide dynamic range of mud and formation properties variation and the measurements. Results for processed resistivity, standoff, and permittivity images are presented, demonstrating very good agreement and consistency with inversion-generated images. The combination of two strategies, training on both synthetic and field data, can lead to further improvement of robustness allowing customization of interpretation applications for specific formations, muds, or applications.
A high risk of suboptimal well placement exists in new field development where seismic uncertainty can be great. Recent ultradeep resistivity measurement developments provide great benefits for identifying and optimizing the well path position within a given stratigraphic sequence. This paper presents a case study in which an operator planned to place wells 10 m TVD below the reservoir top because of seismic uncertainty of the top reservoir pick. To help mitigate this subsurface risk, the field development plan required real-time well placement optimization, using both standard formation evaluation data and an ultradeep azimuthal resistivity service. In this case-history, the ultradeep inversion canvases could be used to identify the well path position within the reservoir, as well as provide sufficient confidence to steer the well closer to the reservoir top than originally planned.
Multiple geological models, created from nearby offset wells and seismic grids, represented the expected seismic uncertainty of 5 to 15 m TVD. To identify the optimal measurement setup for real-time operations, resistivity modelling illustrated the effect of frequency and spacing on the data, producing multiple inversions for each geological scenario. After drilling began, real-time inversions for the ultradeep resistivity data were initially qualified using standard formation evaluation data, including both deep azimuthal resistivity and azimuthal density images. Multiple inversion canvases from various spacings and frequencies identified several formation features, including distances to the top and base of the reservoir. The quantified uncertainty of these results assisted in the evaluation of the inversion quality.
When close to the reservoir top, the wellbore position indicated in the ultradeep inversion canvases matched the interpretation from the conventional logs, which provided increased confidence in the inversion canvas results at distances farther away. This enhanced reservoir knowledge enabled the operator to progressively raise the well path to 5 and to 2 m TVD from the reservoir top. Except for strategic geosteering decisions based on expected faults positions from the seismic data, the operator made most well-placement decisions, across multiple wells, using ultradeep resistivity data. The high data quality and close collaboration within the subsurface team quickly led to high confidence in the inversion results. Integrating the full suite of available data, from shallow to ultradeep measurements in a comprehensive interpretation, provided better reservoir understanding, resulting in optimal well placement.
This paper presents formation evaluation results used within an integrated well-placement optimization service from a new field development. The integrated data qualified the results for an ultradeep resistivity tool. Confidence in the tool results enabled the operator to place wells much closer to the reservoir top than initially planned, in an area of seismic uncertainty.
Cossa, Alessandro (Eni S.p.A.) | Manassero, Elena (Eni S.p.A.) | Mele, Maurizio (Eni S.p.A.) | Musca, Claudio (Eni S.p.A.) | Tarchiani, Cristiano (Eni S.p.A.) | Arrigoni, Veronica (Vår Energi) | Halset, Gjertrud (Vår Energi)
In a deep-water green field’s development, horizontal wells are drilled to exploit the reservoir with the aim of reducing drilling cost and time to first oil.
During geosteering operations, Ultra Deep Azimuthal Electro-Magnetic (E.M.) measurements permit investigating the reservoir around the borehole up to a maximum depth of 30 m, in a good resistivity contrast environment. Acquired data are inverted on a vertical section, providing multi-boundary reservoir mapping along well path.
With such a depth of investigation, the reservoir mapping is an excellent bridge between conventional logging-while-drilling (LWD) and seismic images. The integration of acquired wellbore data with high-resolution attributes, from seismic inversion, maximizes well placement results when operating in complex subsurface geology and expands the perspectives of geosteering application.
A workflow to calibrate the reservoir structural and stratigraphyc setting has been assessed, via integration of seismic and Borehole Data. Enhancement of reservoir geometry interpretation during geosteering provides revised structural surfaces suitable for a quick update of the velocity model and a depth-calibration of all the seismic attributes used to steer wells.
We describe an application of the workflow to an infill well, targeting channel and crevasse splays deposits drilled through a structurally complex oil field in the Norwegian offshore. The availability of seismic attributes (probabilities of facies and petrophysical properties) allowed improving the overall results of the well placement operation.
Reservoir mapping identifies in real time Geo-bodies crossed by the well and within the range of investigation of Ultra Deep E.M. tool based on tool configuration, frequencies analysed and resistivity contrasts of the rocks. Stratigraphic correlation with offset wells, using conventional LWD data supported by Image Log interpretation, allowed allocating resistivity boundaries in terms of stratigraphic surfaces. These data are then integrated in near real time to depth calibrate maps, update the velocity model, hence the depth image of seismic attributes.
After depth calibration, Geo-bodies recognized on seismic show a good correspondence with those identified on the resistivity inversion and a detailed correlation of the heterogeneous fluvial sand was possible, even in presence of minor faults. In this challenging structural and stratigraphic environment, the correlation supported decision making during well operations to target the well on the pay sand.
The application proves that a detailed stratigraphic interpretation is an achievable goal in real time to steer successfully the well and to be used afterwards in a detailed reservoir model update.
Until recently, reservoir characterization methods in the industry were limited to use of seismic technologies in exploration of oil and gas and had a very constrained role in production and development. In the past, using characterization for development fields was considered a very perilous task. Technological advancements and the risk-averse mindset have significantly expanded the application of reservoir characterization. Today, reservoir characterization is the basis of any development plans made for a commercial field.
Development of 3D reservoir modeling techniques to generate field development plans (FDPs) marked a step-change in reservoir characterization methods. Introduction of geostatistics and numerical simulation made it possible to build precise models to generate realistic field development scenarios. This is the state-of-the-art seismic-to-simulation method of reservoir characterization used in FDPs today. However, the struggle to estimate reservoir properties spatially away from the well continues.
Surface seismic data provide excellent areal coverage but do not provide the vertical resolution required for a fine-scale reservoir model. Geostatistical methods reduce the uncertainty in spatial distribution of petrophysical properties from pseudo-point supports (wells) but are not calibrated spatially between the wells. Correspondingly, the fluid saturation distribution and the parameters used in dynamically calculating the same during numerical simulation are not calibrated in the interwell space.
This paper details necessary data acquisitions and methods of calibration of 3D reservoir model to reduce uncertainty in the interwell space. The data acquisition methods have been available for some time, but have rarely been electronically incorporated in the 3D reservoir model and have been largely used to analytically guide the modeling and its inferences. A logical way of interpreting the results of acquisitions and calibrating the 3D reservoir model cell-by-cell is detailed in this paper.
In a conventional formation evaluation process, the mud-filtrate invasion in the near-wellbore region is considered a bias that requires a well-log correction before any petrophysical evaluation. The developments presented in this paper show that the invasion zone is a valuable source of information to estimate dynamic properties that generally come only from core measurements, such as permeability, relative permeabilities, capillary pressure curves and formation factor.
In this approach, the invasion process is not simulated in itself, as it would lead to a very unstable inverse problem within the time frame of the logging. On the contrary, it considers the fluids in the invaded domain as radially equilibrated and solves the fluid distribution governed at first-order by capillary pressures. Due to the multimodality of the inverse problem and the uncertainties related to the mud-filtrate parameters, the invasion zone is jointly inverted with the vertical capillary equilibrium at field-scale describing the vertical water saturation profile in the reservoir for each facies. The following workflow is then used: First, the invasion is solved in the water intervals while inverting the resistivity logs. The resolved parameters are the local volume of filtrate, pseudopermeabilities and cementation factors at each depth. At the end of this step, we get an insight of the number of petrofacies and the correlation between permeabilities and porosities inside each of these. Second, the inversion in itself is carried out in the hydrocarbon zone by exploiting the grouping from the first step. The vertical capillary equilibrium is added and updates permeabilities (absolute and relative) as well as capillary pressure models for each facies.
In the context of this paper, we present a vertical well and consider a radial oil-based mud invasion. We also assume isotropic petrophysical parameters. The final results are compared to all available sources of data, such as NMR, WFT and cores for permeabilities, formation factor and capillary pressure curves.
The ultimate added value of such an approach is to bridge static and dynamic petrophysical parameters from a single source of data: logs. It provides a reliable first guess of petrophysical and reservoir parameters at an early stage of the well evaluation. It also ensures an overall consistency of the formation model for the whole range of facies and fluid configurations. The technique can even help in the formation heterogeneity and petrophysical upscaling when run in a multiwell configuration.