P-wave azimuthal AVO variations are not necessarily correlated with the magnitude of fracture density. Shen (1998) showed that the elastic properties of background rocks have an important effect on P-wave azimuthal AVO responses. Results from 3D finite-difference modeling show that azimuthal AVO variations at the top of gas-saturated, fractured reservoirs, which contain the same fracture density, are significant in the reservoir model with small Poisson's ratio contrast. Analytical solutions indicate that azimuthal AVO variations are detectable when fracture-induced reflection coefficients can generate a noticeable perturbation in the overall reflection coefficients. Varying fracture density and saturated fluid content can lead to variations in AVO gradients in off fracture strike directions. Shen's numerical results also show that AVO gradients may be significantly distorted in the presence of overburden anisotropy caused by vertical transverse isotropy media, which suggests that the inversion of fracture parameters based on an individual AVO curve would be biased without correcting this influence. He recommended that azimuthal AVO variations could be effective for detecting fractures; model analysis studies and combination of P-wave NMO velocities are more beneficial than using reflection amplitude data alone.

Shen (1998) studied various elastic parameters of five sandstones, whose characteristics are summarized in Tables 5.2 and 5.3. He studied for two aspect ratios, i.e., 0.01 and 0.05. Figure 5.3 shows δ^(v) as a function of fracture density for a gas-saturated sandstone with fracture aspect ratio of 0.01. Results are shown for an aspect ratio of 0.05 (Figure 5.4). These results show that for gas-saturated sandstones, δ^(v) is insensitive to aspect ratio. For the water-saturated case, however, absolute values of δ^(v) increase with fracture density and crack aspect ratio. This latter case (Figures 5.5 and 5.6) also shows a range of values for different samples, as compared to gas-filled case that show little dependence on sample types. It is also noted that δ^(v) is dependent on V_p/V_s of isotropic, unfractured sandstones. The smaller the V_p/V_s, the larger the absolute value of δ^(v) obtained.

ε^(v) shows similar characteristics to δ^(v). The difference is that ε^(v) is the function of the V_p of the background medium. In gas-saturated, fractured sandstones, ε^(v) is sensitive to fracture density alone. In water-saturated, fractured sandstone the absolute value of ε^(v) increases with both fracture density and aspect ratio. The smaller the Vp, the smaller the absolute value ε^(v) obtained. ε^(v) as a function of fracture density and crack aspect ratio in gas- and water-saturated, fractured sandstones is shown in Figures 5.7 through 5.9.

Parameter γ^(v), is different from both δ^(v) and ε^(v) in that it measures the degree of shear wave splitting at vertical incidence. Figures 5.10 through 5.13 show that γ^(v) has little dependence on fluid bulk modulus and crack aspect ratio and is the parameter most directly related to fracture density. Therefore, for parallel, penny-shaped cracks, the shear wave splitting parameter, γ^(v), can provide direct information about fracture density with least ambiguity.

These findings show that the variations of parameters δ^(v) and ε^(v) are sensitive to fluid content. For gas-saturated fractures, δ^(v) and ε^(v) vary with the fracture density alone, making them an effective indicator of fractures. On the other hand, for water-saturated fractures, the magnitudes of δ^(v) and ε^(v) depend on fracture density, crack aspect ratio, and elastic properties of background rocks. Whereas the shear wave splitting parameter, γ^(v), is insensitive to both fluid content and crack aspect ratio. It is the parameter most related to crack density. P-wave NMO velocity is controlled by the vertical P-wave velocity, angle between crack normal and survey line, and the parameter δ^(v).

During overbalanced drilling, when a sweet spot is encountered, the mud pressure is greater than the fluid pressure contained in the fracture resulting in a flood of drilling mud into the fracture. When the drill bit intersects the fracture, drilling mud will flow into the fracture. Because the matrix permeability is low, leak off into the matrix is minimal and the mud loss is entirely due to fluid flow through fractures.

This mud flow is reflected in the rheology of the returning mud, thereby, creating a correlation between surface-observed rheology and fracture density as well as fracture geometry (Figure 5.15).

Lietard et al. (1999) provide type curves describing mud-loss volume versus time that can be used to determine the hydraulic width of fractures through a curve matching approach. The type curves are based on an analysis of the local pressure drop in the fracture (Lietard et al., 1999):

The most useful aspect of UBD is in the insight gained during UBD. The deliberate underbalanced pressure difference between the drilling fluid and the formation pressure causes an inflow of formation fluid into the wellbore along the entire drilled section. This can act as a tracer for dynamic reservoir characterization.

The most important piece of data is the rate of fluid flow from the formation into the wellbore. Very rarely are flow rates actually measured at bottomhole. In almost all cases, the formation fluid-flow rate is estimated from surface measurements of inflow and outflow of the drilling fluid. The difference between the mud injected into the wellbore and the outflow of mud from the annular section is often estimated as the formation fluid-flow rate. Methods to account for the expansion of gas due to changes in temperature and pressure must be taken into account to obtain accurate data (Aremu and Osisanya, 2008). Other data that models tend to utilize include bottomhole pressure, ROP, formation porosity, wellbore diameter, and wellbore length. Models generally provide profiles of formation permeability and pore pressure versus depth. Norbeck (2012) presented field data, showing correlation between UBD data and fractures. He extracted field data originally reported by Myal and Frohne (1992). The report investigates the effectiveness of directional drilling in a tight gas formation located in the Piceance Basin of western Colorado. The formation is known to be highly naturally fractured, and consequently the decision was made to drill a large section of the well underbalanced in order to reduce formation damage and lost circulation. As can be seen from Figure 5.17, at least 10 major gas shows were detected during drilling. These gas shows were attributed to the presence of natural fractures intersected by the wellbore. An increase in mud density not only led to suppression of the gas shows, but also prevented extraction of gas show data and their correlation with fracture distribution. This type of correlation can lead to depiction of the formation fracture network.

Norbeck (2012) proposed two criteria that can be combined to develop a correlation between mud log data and reservoir properties. They are:

Both criteria are related to open fractures that contribute to flow directly. Furthermore, fractures are uniquely correlated if the formation is tight with negligible permeability. In order to estimate natural fracture permeability, several assumptions have to be made:

The following estimates can be obtained from the drilling and mud log data:

If an assumption about the radial extent of the fracture can be made, then fracture aperture can be determined as follows:

Then, fracture permeability can be estimated by the following equation that is derived from Poiseuille law, as applied in parallel plate flow.

Norbeck (2012) demonstrated through study of six wells that this technique is both practical and accurate. Data from six wells that were drilled underbalanced were collected. Conductive natural fracture zones are determined for each well, thereby estimating fracture permeabilities. The wells were from two tight gas shale formations, one located in the United States and the other in Canada. The lateral sections of these horizontal wells range from 3000 to 6000 ft. The lateral sections of all wells were drilled using oil-based mud.

In order to verify accuracy of the analysis, a special technique had to be applied because no borehole image logs were available. The main validation technique used in that study took advantage of a commonly used horizontal drilling technique in which wells are drilled in parallel. Six wells selected for the study constituted three sets of parallel wells drilled to similar elevations. The spacing of these wells was between 500 and 800 ft. The parallel well sets should penetrate similar natural fracture systems. The results of the natural fracture identification analysis for each parallel well set are compared to determine if any patterns exist that may be indicative of the orientation of natural fracture planes.

In the first field case study, first two wells (Well A-1 and Well A-2) were chosen from the same field. The horizontal spacing between these two wells is roughly 800 ft. Well A-1 runs in S–N orientation and Well A-2 runs in the opposite direction. Well A-1 was drilled toe down and Well A-2 was drilled toe up. The geometric properties and the average reservoir pressure for each well, obtained from diagnostic fracture injection testing, are listed in Table 5.4. The targeted pay zone is roughly 175-ft thick.

The analysis of Well A-1 indicates that 10 conductive natural fractures were intersected during the drilling process (see Table 5.5). Two of these natural fractures are within very close proximity to each other and are considered a single conductive natural fracture zone. In total, nine natural fracture zones are present along the lateral of this well. As can be seen in Figure 5.19, the stretch of lateral between 14,000 and 15,500 ft MD contains no conductive natural fracture zones. This is a primary example of the insight that can be gained from this type of analysis. This zone can be selected for creating sweet spots through hydraulic fracturing. The fracture apertures range from 13 to 53 μm. The cross-plot indicates a general positive correlation between mud pit volume peak and gas peak for these conductive natural fracture zones (see Figure 5.20).

Similarly, a total of nine conductive natural fracture zones were identified for Well A-2. The relevant results are listed in Table 5.6. The zones are relatively evenly spaced along the lateral (see Figure 5.21). The fracture apertures are generally smaller than for Well A-1, ranging from 15 to 38 μm. These estimates could be largely due to the higher level of underbalance maintained while drilling Well A-2. Additionally, the observed rise in mud pit volume due to the presence of these fractures is relatively low. The cross-plot does not show a strong trend for the relationship between mud pit volume peak and gas peak for these fractures, however, it is a positive correlation (see Figure 5.22).

Similar results were obtained for other fields as well. On average, the computational tool identifies between nine and ten conductive natural fracture zones for each well. For each conductive natural fracture zone, the fracture aperture and fracture permeability was estimated. The estimated fracture apertures all lie within the expected range of values (i.e., 10–1000 μm). Overall, although the estimates of fracture aperture may not be entirely accurate, they should be considered as a lower bound for the true fracture apertures.

In absence of other means of validation, patterns in the locations of conductive natural fracture zones between wells were used. Existence of such patterns would confirm tectonic continuities that are essential for history matching of the diagenesis involved to validate the results, because each pair of wells should penetrate similar geologic conditions. Once validated the existence of fractures (or sweet spots), orientation of fractures could be used to refine reservoir characterization. For the case in question, two out of the three parallel well pairs exhibit strong patterns. From visual inspection of the results obtained for Field A, two dominant patterns can be observed, as can be seen in Figure 5.23. A total of seven pairs of natural fractures are aligned at an orientation of roughly N65°E (see Figure 5.24). Only one identified natural fracture from Well A-2 does not have a corresponding feature in Well A-1. For each of the seven natural fracture planes identified, the estimated apertures of the corresponding pair of natural fractures compare well, with the exception of pairs four and five (see Table 5.7).

Usually, during the field recording phase it is possible to make a preliminary individuation of levels that can be potentially fractured. These are very often associated to:

Borehole imaging tools provide an image of the borehole wall that is typically based on physical property contrasts. There are currently a wide variety of imaging tools available, though these predominately fall into two categories: resistivity and acoustic imaging tools.

Resistivity imaging tools provide an image of the wellbore wall based on resistivity contrasts (Ekstrom et al., 1987). Resistivity imaging tools consist of four- or six-caliper arms with each arm ending with one or two pads containing a number of resistivity buttons. Resistivity imaging tools provide one with the same information on borehole diameter and geometry as the older dipmeter tools, however, the resistivity buttons also allow high-resolution resistivity images of the borehole wall to be developed. There are a wide variety of wireline resistivity imaging tools available, some of the more common tools are the Formation MicroScanner (FMS; from Schlumberger), Formation MicroImager (FMI; from Schlumberger), Oil-Based MicroImager (OBMI; from Schlumberger), Simultaneous Acoustic and Resistivity tool (STAR; from Baker Atlas), Electrical MicroScanner (EMS; from Halliburton) and Electrical Micro Imager (EMI; from Halliburton). Furthermore, recent years have seen the development of a range of logging while drilling (LWD) or measurement while drilling (MWD) resistivity image logging tools, such as the Resistivity At Bit (RAB; from Schlumberger) and STARtrak (from Baker Inteq). For more details on resistivity image logging tools see Ekstrom et al (1987) or Asquith and Krygowski (2004).

Acoustic tools, on the other hand, emit high frequency sonar waves. The acoustic imaging tool then records the amplitude of the return echo as well as the total travel time of the sonic pulse. The acoustic wave travel time and reflected amplitude are measured at numerous azimuths inside the wellbore for any given depth. These data are then processed into images of the borehole wall reflectance (based on return echo amplitude) and borehole radius (based on pulse travel time). There are a wide variety of acoustic imaging tools available, some of the more common tools are the Borehole Televiewer (BHTV; from Schlumberger), Ultrasonic Borehole Imager (UBI; from Schlumberger), Circumferential Borehole Imaging Log (CBIL; from Baker Atlas), Simultaneous Acoustic and Resistivity tool (STAR; from Baker Atlas), Circumferential Acoustic Scanning Tool-Visualization (CAST-V; from Halliburton) and the LWD/MWD Acoustic Caliper tool (ACAL; from Halliburton).

The use of AFIT has been implemented to characterize the permeability of feed zones in oil and gas and geothermal wells. A description of the AFIT tool is given by McLean and McNamara (2011):

In practice, the interpretation of AFIT data is quite sophisticated. Planar natural fractures appear as sinusoids in the imaged data set, as shown in Figure 5.26.

Data processing software allows for characterization of geologic features including strike and dip, fracture aperture, and fracture density. The signal amplitude can be used to distinguish between open and closed fractures. Low amplitude signals are seen as dark features on the acoustic image and often interpreted as open. High amplitude signals are seen as light features on the acoustic image and are thought to be attributed to mineral fill. These high amplitude features are usually considered as closed fractures that do not contribute to flow. While McLean and McNamara (2010) report a good level of correlation between measured feedzone fluid velocity and fracture aperture determined from AFIT, the fracture apertures reported range from several centimeters to greater than 50 cm. This implies that the resolution of AFIT can at best distinguish fractures of roughly 1–2 cm. This is nowhere near the level of resolution quality necessary for fracture characterization in highly fractured tight gas reservoirs.

Resistivity image logs, also called FMI logs, have been documented as an improved technique to characterize geologic features along the wellbore. These techniques make use of a tool that places an electrode at constant electrical potential against the borehole wall and measuring the current. Picture 5.2 shows the device. The tool is a small-diameter imaging tool that can be deployed with or without a wireline. The high sampling density of these tools (e.g., 120 samples per foot) provides extremely high resolution in the image quality. Microresistivity imaging tools have the ability to visualize features down to 2 mm in width. Careful interpretation of resistivity image logs can provide helpful information about the geologic conditions near wellbore, including dip analysis, structural boundary interpretation, fracture characterization, fracture description, and fracture distribution.

Data obtained from all three of these image log techniques can be analyzed to gain useful insight about the natural fracture system and state of stress system near wellbore. Barton and Zoback (2002) present an approach for discriminating natural fractures from drilling-induced fractures from different types of image logs, and using this knowledge to determine the state of stress in situ. It has been well documented that drilling-induced tensile fractures will form in the azimuth of the maximum horizontal principal stress. If the three in situ principal stresses can be determined and the formation fluid pressure is known, then a Coulomb failure analysis can be applied to the natural fractures identified in the image logs. Shear and effective normal stresses acting on each fracture plane can be determined from knowledge of the orientation of the fracture plane with respect to the orientations of the principal stresses. The Mohr–Coulomb failure envelope for fractures is determined from laboratory measurements on prefractured rock. The failure line is constructed assuming no cohesion and using the friction angle of the prefractured rock. Poles to fracture planes are then displayed on the Mohr diagram. Critically stressed fracture planes lie above the Mohr–Coulomb failure envelope (Figure 5.27). Barton and Zoback (2002) report a strong correlation between critically stressed fracture planes and hydraulic conductivity of the fractures. These findings indicate that only a small percentage of the total number of fractures are likely to contribute to flow.

Of relevance is also the fact that total porosity can only be determined through core analysis. For unconventional gas perspective, this is of great consequence because as much as 50% recoverable gas can go unaccounted without definitive knowledge of porosity. Conventional assessment of porosity through GR analysis can be vastly misleading in unconventional reservoirs.

Practically all aspects of core analysis are different in unconventional reservoirs from conventional reservoirs. They are highlighted as follows:

The most prominent of the above is the one related to relative permeability graph. In a conventional reservoir, it is clear that there is relative permeability in excess of 2% to one or both fluid phases across a wide range of water saturation. In traditional reservoirs, critical and irreducible water saturation occur at similar water saturation values. Under these conditions, the absence of common water production usually implies that a reservoir system is at, or near, irreducible water saturation, low-permeability reservoirs, however, one can find that over a wide range of water saturation, there is less than 2% relative permeability to either fluid phase. In these reservoirs, the lack of water production cannot be used to infer irreducible water saturation. In some very low-permeability reservoir, there is virtually no mobile water phase even at very high water saturations. The term “permeability jail” coined by Shanley et al., 2004 describes the saturation region across which there is negligible effective permeability to either water or gas.

Low-permeability reservoirs are usually characterized by high to very high capillary pressures at relatively moderate wetting-phase saturations (Figure 5.46). In many cases, wetting-phase saturations of 50% (close to critical gas saturation, S_gc ) are associated with capillary pressures in excess of 1000 psia, suggesting that a large number of pore throats are less than 0.1 μm in diameter and are of the micro- to nanoscale. In many low-permeability sandstone reservoirs, wetting-phase saturation continues to decrease with increasing capillary pressure.

The relationship between relative permeability, capillary pressure, and position within a trap in an unconventional reservoir are shown in Figure 5.47 (Shanley et al., 2004). It shows a reservoir body that thins and pinches out in a structurally updip direction. In this case, significant water production is restricted to very low structural positions near the free water. In many cases, the effective permeability to water is so low that there is little to no fluid flow at or below the free water level. Above the free water level, a wide region of little to no fluid flow exists. Farther updip, water-free gas production is found.

In terms of porosity, overburden pressure plays a significant role in unconventional reservoirs. This is unlike carbonate reservoirs or conventional sandstone reservoirs. Figure 5.48 shows little change in porosity with net increase in stress.

For unconventional reservoirs, such as shale formation, tight sand, etc., porosity is affected by overburden stress. Mechanical compaction is thus an inevitable consequence of burial and basin evolution. Often, this effect is modeled as a function of depth that is most affected by the overburden. However, depth is only a position coordinate that specifies present-day location. Consequently, depth is a poor measure of the processes that have acted upon a sedimentary section through time, which is the primary function that governs all features of a formation. In that sense, everything, including diagenetic process, remains dynamic. Conventional models of shale compaction relate porosity to effective stress using the empirical relationship between void ratio and effective stress established in soil mechanics (Burland, 1990; Yang and Aplin, 2004). Okiongbo (2011) explored the effect of petroleum generation by evaluating the variation in porosity and effective stress in the Kimmeridge clay formation (KCF) above and within the oil window.

Figure 5.49 shows porosity and effective stress relationship for above and below oil window. A common interpretation of the effective stress–porosity relationship above and within the oil window is that the loss of porosity is faster in the pregeneration zone than within the oil window. In addition, cementation and deeper burial create a relatively stiffer matrix for the KCF sediments within the oil window making it more difficult to compact at high effective stresses. In presence of organic matter, porosity may be impacted by biogenic products. The presence of fractures and fissures alters the dynamics of porosity and pathways for escape of organic gaseous products. The exact nature of the origin of porosity is not known, but there are factors that are not well understood, particularly in the context of unconventional reservoirs. Figure 5.49 also shows a correlation developed by Yang and Aplin (2004) as well as Skempton (1970) and Burland (1990). These models express porosity as a unique function of clay and its natural, fine-grained clastic sediments. Yang and Aplin's (2004) model is based on well data from the North Sea as well as data derived from the studies of Skempton (1970) and Burland (1990). Although the clay content is assumed to be the same (∼65%) in all the data sets, a remarkable difference exists in terms of their organic carbon content. Skempton (1970) and Burland (1990) data sets are low TOC samples (TOC < 1 wt%), the North Sea data set used by Yang and Aplin, (2004) have TOC ranging between 1% and 5wt%, while Okiongbo (2011) data set that are portrayed with red and orange lines have TOC ranging between 5% and 10wt%. In addition, Yang and Aplin's (2004) data set exclude chemically compacted sediments. Above the oil window, the Figure 5.49 indicates similarity between the trend of Yang and Aplin (2004) and that from this study on extrapolation to low effective stresses (<5 MPa). Significant variation only occurs at effective stresses less than five.

Within the oil window, the Figure 5.49 shows a close agreement between Yang and Aplin's data and Okiongbo data. Overall, the following factors play a role:

While the first three components have been investigated with some details, thermal effects have not received much attention. Ehrenberg et al. (2009) related geological age of rocks with their diagenesis in order to assess the impact of temperature on porosity. Figure 5.50 shows the effect of geological age on porosity. The impacts of age and lithology are captured in this figure. For petroleum reservoirs, however, the presence of hydrocarbon adds complications. The release of gas increases pore stress, and any escape and migration by hydrocarbon fluids add complications to the nature of porosity. This dependence is different in carbonate formations from silicate formations (including clay and shales). While depth adds to the overburden stress adding to the compaction of pores, it also adds temperature that affects hydrocarbon-bearing pores in a manner different from water-bearing pores. Adds to this is the chemical compaction and cementation that are directly affected by thermal conditions as well as overburden stress.

Figure 5.51 shows overburden stress alone would decrease the porosity. One must caution, however, unconventional reservoirs cannot be modeled uniquely with overburden stress. The degrees of “stress sensitivity” will be a function of the lithology and the pore throat size distributions. Rocks with “slot pores” will be more stress sensitive than rocks with more round pore throats. Slot pores are created by quartz overgrowths during diagenesis. Whenever fractures (natural or induced) are present, the presence of fractures can alter the flow behavior very differently from conventional reservoirs. This has tremendous impact on permeability. Figure 5.52 shows gas permeability at net overburden pressure as a function of gas permeability at ambient pressure (both Klinkenberg corrected). For high-permeability (10–100 mD) core plugs, the permeability under the original overburden pressure is slightly less than the value of unstressed permeability for that same core plug. However, as the permeability of the core plugs decreases, the effect of net overburden pressure on the core plug increases substantially. For the core plugs that had values of unstressed permeability of around 0.01 mD, the values of permeability under net overburden stress were about an order of magnitude lower, or 0.001 mD. The lower permeability rocks are the most stress sensitive because the lower permeability core samples have smaller pore-throat diameters than the higher permeability rocks. As overburden stress increases, the diameter of the pore throat decreases. Because the permeability of a rock is roughly proportional to the square of the diameter of the pore throat, the permeability reduction in low-permeability rocks is much more dramatic than in high-permeability rocks. This behavior would explain the trapping mechanism presented earlier in this section. It also explains why core permeabilities have little relevance when it comes to fractured formations. Figure 5.52 also shows that a fractured formation would show much steeper dependence on overburden stress. This behavior indicates that flow rates would be increased under low stress conditions.

In order to avoid this problem, filtering should be performed. Any such filter has to be custom designed for a specific field, if not specific well. For that, following data should be collected.

At first, the full 3D schematic of the reservoir should be constructed. As depicted in Picture 5.9, the idea is to transit from microscopic scale to reservoir scale, which involves developing the scaling laws.

This schematic is not quantitatively accurate but represents overall trend. In order to decide on the schematic as well as the type of flow that is expected in the reservoir, geophysicists, geologists, and engineers must work together.

Based on that, one of the following averages should be used to construct the permeability (k) × net pay (h) correlation.

For parallel flow, arithmetic average should be used:

For series flow, harmonic average should be used:

For flow without a particular pattern (random flow), geometric average should be used:

At this point, well test results should be gathered and history of permeability variation with time should be collected. Quite often, well test permeability becomes a function of time. Such behavior gives out useful clues as to the nature of fluid flow in the reservoir. The knowledge of the operators, combined with the construction of geological history, and fluid flow behavior all contribute to determining the true nature of reservoir. For instance, consistent decline in reservoir permeability, as evidenced from well test data can occur because of fracture closure, asphaltene deposition, or formation of gas pockets in the neighborhood of the well. It is important to conduct laboratory tests in order to determine natural reaction to cores to stress. Figure 5.74 shows an example of such a test. In this laboratory simulation with artificially consolidated cores, the original stress corresponds to initial conditions of a reservoir, which is subject to little stress because of overall balance of forces. This is expected because over geological time, any reservoir reaches a state that can be termed steady state in conventional terminology. This stress is increased over time as more and more fluid is extracted from the reservoir. During this time period, original permeability can be reduced to less than 50% of the initial permeability. Many field history support this observation. For instance, in several of the giant oil field of Hassi Messaoud, Algeria, similar decline in well test permeability has been observed. The decline is steeper for more heterogeneous zones, the ones that have profuse fracturing.

As can be seen in Figure 5.74, the core is crushed if the axial stress is very high (exceeding 100 MPa), resulting in sudden increase in permeability. Even then, however, original permeability is not restored. In theory, such occurrence of partial restoration of permeability can occur in the field. Several fields report such behavior. Almost all of these have fractures with intense network of secondary cementation.

In order to characterize a fractured formation, one has to know the following:

One of the most useful tools of designing local filters is the compilation and plotting of hk from well tests and cores. In case of relatively homogeneous formation, the correlation between the two will follow the trend of a 45^° straight line. Figure 5.74 demonstrates this point. On the other hand, if there are open fractures with significant aperture and length in the reservoir, the data points will fall over the straight line (meaning HKR is greater than unity, where HKR = ratio of hk_welltest/hk_core), the highest points representing maximum departure from the median line and highest frequency of open fractures. On the other hand, when points are located under the median straight line (meaning HKR less than unity), it signals the existence of secondary cementation that caused fractures to be plugged (Figure 5.75). For this case, the fraction of closed fractures over open fractures for various locations must be determined before a useful filter can be constructed.

This information is crucial to developing the filter. The filter uses the following data:

Natural fracture distribution map should be an indicator of how induced fracture will behave. Therefore, the data collected on natural fractures should be processed for designing an induced fracture.

Stress data collected during drilling as well as cuttings and well site logs should be all combined to develop enhanced understanding of reservoir fracture properties.