Search

This study outlines a probabilistic model based on artificial neural networks applied to the deeply-buried karsted carbonates of the Ordovician Yingshan Formation, which represent significant oil reservoirs in western China. The complexity of both rock type prediction and 3D facies modeling of paleokarst fillings, which are hosted within the cavities, drives the need to apply innovative techniques for identifying new oil plays. Due to the high heterogeneity of clastic fillings and patchy continuity of the karst patterns, physical evaluation of these reservoirs is extremely complex. We propose the Democratic Neural Networks Association (DNNA) as the probabilistic technique to solve these challenges. This technique simultaneously runs several artificial neural networks in parallel and combines seismic data and well logs. The resulting probable facies volume is expected to provide an appropriate distribution and delineation of clastic fillings (i.e., conglomerates, fine-grained sandstones, siltstones, mudstone, dolomite fragments, and sparry calcarenite) and unfilled or empty spaces. This calculated volume is then used as a reliable input data to condition trend analysis on a very fine geological grid, in order to model the complex patterns in question. The static model obtained shows that, the probabilistic distribution of each filling has the same orientation as karst system. Likewise, spatial dimensions similar to the proposed analogue model of these patterns (vertical and horizontal scales) are delineated. Finally, we validated prediction results by comparing them with the interpreted karst facies of a well not initially considered in the 3D model. The results indicating that the DNNA technique proves to be a useful innovative tool for generating realistic depictions of fillings deposited within deeply-buried paleokarst.

The characterization and modeling of fractures in karsted carbonate reservoirs are challenging due to the mechanical interactions between fractures and irregular karst elements. This study proposes an integrated approach for characterizing and modeling large-scale fractures within deeply buried fractured karst carbonates of the Yingshan Formation (Ordovician in age) located at the Tahe oilfield, Tarim Basin (western China). This investigation implements a promising workflow to stochastically model the fracture sets interpreted from borehole image logs and compute both fracture porosity and permeability tensor in karsted carbonates. For this case study, we have constructed a Discrete Fracture Network (DFN) by combining two structural seismic attributes (i.e. local structural dip and ant-tracking). We have also applied the automatic fault extraction (AFE) algorithm. The results indicate that this method detects a greater number of discontinuities associated with fractures compared to the conventional methods employed in this kind of reservoir (e.g. seismic coherence and curvature). The fracture set with higher values of fracture porosity and permeability coincided with the orientation of maximum horizontal stress, WSW-ENE. In the model, the highest values of fracture intensity correspond to the most productive oil wells. Finally, this study highlights the importance of fracture modeling for evaluating the distribution of the most favorable production areas, so-called sweet-spots, in karsted carbonates reservoirs.

Fault roughness constitutes a key element in the understanding of earthquake nucleation, and surficial asperities on the fault plane play a critical role in slip dynamics and frictional behavior during the seismic cycle. Since it is not generally feasible to recover fault roughness profiles or maps directly at the seismogenic sources, faults at the Earth’s surface are typically used as analogues. However, these analogue fault surfaces are often subjected to weathering and erosion, which in turn, reduces their representativeness as seismogenic faults. Rupture along active faults episodically exposes “fresh” fault planes at the Earth’s surface, which represent the best available targets for the evaluation of fault roughness generated at seismogenic depths.Here we present a study conducted on a splay of the Mt. Vettore fault system in the Central Apennines, Italy, along a vertical transect that includes both a weathered and freshly exposed portion of the fault. The latter was exposed after the dramatic Mw 6.5 shock that hit the area on 30 October 2016. We have produced a highly detailed model (i.e., point cloud) of a section of the fault using structure from motion-multiview stereo photogrammetry to assess its roughness parameters (i.e., the Hurst fractal parameter) and to determine the extent to which these parameters are affected by weathering assuming that they had similar fractal characteristics when reaching the surface.Our results show that weathering can modify the value of the fractal parameters. In particular, by independently analyzing different patches of the fault, we have observed that the smoother and recently exposed portions have an average Hurst exponent of 0.52 while the average Hurst exponent of zones with more prolonged exposure times is 0.64. Accordingly, we conclude that by using high-resolution point clouds, it is possible to recognize patches of faults having a similar intensity of deterioration attributable to weathering.

Fluid flow through a single fracture is traditionally described by the cubic law, which is derived from the Navier-Stokes equation for the flow of an incompressible fluid between two smooth-parallel plates. Thus, the permeability of a single fracture depends only on the so-called hydraulic aperture which differs from the mechanical aperture (separation between the two fracture wall surfaces). This difference is mainly related to the roughness of the fracture walls, which has been evaluated in previous works by including a friction factor in the permeability equation or directly deriving the hydraulic aperture. However, these methodologies may lack adequate precision to provide valid results. This work presents a complete protocol for fracture surface mapping, roughness evaluation, fracture modeling, fluid flow simulation, and permeability estimation of individual fracture (open or sheared joint/pressure solution seam). The methodology includes laboratory-based high-resolution structure from motion (SfM) photogrammetry of fracture surfaces, power spectral density (PSD) surface evaluation, synthetic fracture modeling, and fluid flow simulation using the Lattice-Boltzmann method. This work evaluates the respective controls on permeability exerted by the fracture displacement (perpendicular and parallel to the fracture walls), surface roughness, and surface pair mismatch. The results may contribute to defining a more accurate equation of hydraulic aperture and permeability of single fractures, which represents a pillar for the modeling and upscaling of the hydraulic properties of a geofluid reservoir.

We present the results of 3-D monitoring of strain along active faults outcropping in the central Apennines, where recent and historical strong earthquakes occurred, using TM-71 extensometers during the period 2002–2017. We discuss the fault slip development as well as kinematics particularly in the light of local recent seismicity. The recorded fault slip corresponds to the minimum and maximum principal stress with a SW-NE orientation and slightly dipping to SW. Moreover, the detected fault slip is non-linear and affected by remarkable transient periods of acceleration, which lasted up to several months. The recorded accelerations were induced by alternation of extensional, reverse and horizontal displacement due to switching extensional and compressional stress/strain state and coincide with periods of major seismic events that recently affected the central Apennines. This phenomenon is being investigated. In the case of the moderate L'Aquila 2009 and Norcia 2016 earthquakes, the recorded fault slips can be characterized as pre-seismic.

This work assesses the impact of the stratigraphic and structural heterogeneities on the fluid distribution and migration within a deformed carbonate reservoir analogue. The study was carried out in the abandoned Roman Valley quarry (Majella Mountain, central Italy), well-known for its historical bitumen extraction. The studied rocks consist of ramp carbonates of the lower member of the Bolognano Formation (Oligocene-Miocene in age) mainly composed of grainstones, and secondary packstones and wackestones. In the quarry, the exposed rocks are crosscut by two high-angle oblique-slip faults striking WNW-ESE with up to 40 m of throw. With the aim of deciphering the contribution of both matrix and fracture pore space on the fluid storage and migration, both laboratory measurements and field-based Discrete Fracture Network (DFN) models (cubes with 4-m per side) have been carried out. The obtained dual-porosity/permeability model, which includes the lithological and the structural heterogeneities observed at the location, was used to run flow simulations under different scenarios of well locations. The fluid distribution in the matrix, resulting from these flow simulations, is consistent with field observations where bitumen localizes within the most permeable lithofacies (grainstones). In the fault zones, the fracture network causes an important fluid flow anisotropy, enhancing the flow in a direction subparallel to the faults. The flow orthogonal to the fault zones is controlled by type and lateral continuity of fault rocks, where fault breccias represent conduits and cataclasites localized barriers.

In deformed porous carbonates, the architecture of the pore network may be modified by deformation or diagenetic processes altering the permeability with respect to the pristine rock. The effects of the pore texture and morphology on permeability in porous rocks have been widely investigated due to the importance during the evaluation of geofluid reservoirs. In this study, these effects are assessed by combining synchrotron X-ray computed microtomography (SR micro-CT) and computational fluid dynamics. The studied samples pertain to deformed porous carbonate grainstones highly affected by deformation bands (DBs) exposed in Northwestern Sicily and Abruzzo regions, Italy. The high-resolution SR micro-CT images of the samples, acquired at the SYRMEP beamline of the Elettra - Sincrotrone Trieste laboratory (Italy), were used for simulating a pressure-driven flow by using the lattice-Boltzmann method (LBM). For the experiments, a multiple relaxation time (MRT) model with the D3Q19 scheme was used to avoid viscosity-dependent results of permeability. The permeability was calculated using Darcy's law once steady conditions were reached. After the simulations, the pore-network properties (effective porosity, specific surface area, and geometrical tortuosity) were calculated using 3D images of the velocity fields. These images were segmented considering a velocity threshold value higher than zero. The study showed that DBs may generate significant heterogeneity and anisotropy of the permeability of the evaluated rock samples. Cataclasis and cementation process taking place within the DBs reduce the effective porosity and therefore the permeability. Contrary to this, pressure dissolution and faulting may generate connected channels which contribute to the permeability only parallel to the DB.

A methodology to assess the effects of structural heterogeneities below seismic resolution in porous carbonate grainstones on reservoir performance during production is developed by integrating structural analysis, power law distributions, up-scaling, and numerical techniques. The novelty of the methodology consists of accounting for the buffering effects on permeability caused by compactive and cataclastic deformation bands. By using this methodology, a 3D deterministic field analogue and a 3D Discrete Fracture Network (DFN) representations of the reservoir/aquifer were built first, and then single-phase, steady-state fluid flow numerical experiments in an equivalent porous medium framework were performed. The field analogue is located along the Northwestern coast of Sicily (Italy) where Lower Pleistocene porous carbonate grainstones are crosscut by a strike-slip fault system. This fault system is made up of two conjugate sets of strike-slip shear structures recognized as single compactive shear bands (CSB), zones of compactive shear bands (ZB) and well-developed faults (DF), with discrete slip surfaces and cataclastic material. The permeability of these structures is up to three orders of magnitude less than the surrounding porous carbonate rocks. The fluid flow numerical experiments have been performed on the two aforementioned reservoir/aquifer descriptions to assess the effect of Sub-Seismic Resolutions Faults (SSRF), such as those observed in the outcrops, on fluid flow during production from a well, injection production in Enhanced Oil Recovery (EOR), and up-scaling to large cell size for regional flow simulation. Comparison of draw-down modeling in the DFN and the deterministic models show that results are similar with the exception of wells located in areas of intense strain localization with ZB and DF. The use of the DFN model is therefore an acceptable representation of the heterogeneities induced by SSRF in a reservoir/aquifer. Results of numerical computations show that, in structurally complex areas, SSRF-related ZB and DF might represent a drilling risk because they can enhance draw-down during production and EOR activities.

The modelling of natural fracture in reservoirs requires, as input data, the results of previous detailed and accurate analysis of the 3D fracture network. These data could be derived from well logs and production tests (which however limit our understanding of the fracture geometry, intensity and distribution) and outcrop analogues. Data obtained applying scanline and scanarea methodologies on rocks exposed at the surface, in fact, allow the construction of more representative numerical models of natural fractured reservoirs. This paper focuses on with the (DFN) modelling of natural fractures associated to strike-slip faults crosscutting tight carbonates, which are exposed along vertical walls and pavements of an inactive quarry of the Murge area, southern Italy. The studied outcrops expose the inner structure of two conjugate fault zones striking WNW-ESE and NNW-SSE, respectively. DFN models were built according to the spatial and dimensional properties computed for the natural fracture network. The results of these models show that the overall fault permeability is 3-to-4 orders of magnitude higher than the host rock permeability. The fault damage zones form the main fluid conduits, with the highest permeability values computed for fault-parallel fluid flow. Such a pronounced permeability anisotropy obtained for the fault damage zone is mainly related to the fracture dimension, both lengths and heights, and their aperture values.

The work tackles the control exerted by a sub-seismic fracture network on both secondary porosity and correspondent permeability of outcropping tight carbonates. Taking advantage of excellent 3D exposures, located in the Murge Plateau of southern Italy, the fracture network is investigated at different scales of observation. The rock multi-layer is made up of 10's of cm-thick, sub-horizontal, laterally continuous limestone beds crosscut by stratabound fractures, non-stratabound fractures, and small faults named as persistent fracture zones with low amounts of offset. Stratabound fractures consist of bed-perpendicular joints and sheared joints, non-stratabound fractures of incipient, cm-offset, sub-vertical faults, whereas the 10's of cm-offset persistent fracture zones are made up of 10's of m-high, m-thick fractured damage zones. The aforementioned structural elements localize within discrete carbonate units bounded by primary features such as bed surfaces, prominent surfaces and sedimentary breccia horizons. Such interfaces therefore affected the fracture stratigraphy of the limestone rock, and thus impact the fluid flow properties of the carbonate multilayer by compartmentalizing deformation. In the field, the fracture network is investigated by means of scanline and scan area methodologies to document the orientation, intensity, height distribution, mechanical aperture and roughness of individual fractures exposed along vertical outcrops and pavements of abandoned quarries. Then, Discrete Fracture Network (DFN) models of representative geocellular volumes are built, according to the different scales of analysis, to compute both fracture porosity and correspondent permeability (K xx , K yy , K zz ). Results of such a work show that the most prominent non-stratabound fracture set forms the major control on fluid storage and migration at the scales of single beds and bed-packages. At a larger scale, we document that fluid migration mainly occurs along the persistent fracture zones, which enhance the fault-parallel flow. As a whole, the persistent fracture zones form localized fluid conduits embedded within carbonate matrices that show isotropic fluid flow properties.