Well Testing of Fracture Corridors in Naturally Fractured Reservoirs for an Improved Recovery Strategy ^†

Abstract

Naturally fractured reservoirs (NFRs) account for a significant portion of the world’s oil and gas reserves. Among them, corridor-type NFRs, characterized by discrete fracture corridors, exhibit complex flow behavior that challenges conventional development strategies and reduces recovery efficiency. A review of previous studies indicates that failing to identify these corridors often leads to suboptimal recovery, whereas correctly detecting and utilizing them can significantly enhance production. This study introduces a well-testing technique designed to identify fracture corridors and to evaluate well placement for improved recovery prediction. A simplified modeling framework is developed, combining a local model for matrix/fracture wells with a global continuous-media model representing the corridor network. Diagnostic pressure and derivative plots are used to estimate corridor properties—such as spacing and conductivity—and to determine a well’s location relative to fracture corridors. The theoretical analysis is supported by numerical simulations in CMG, which confirm the key diagnostic features and flow regime sequences predicted by the model. The results show that diagnostic patterns can be used to infer fracture corridor characteristics and to approximate well positions. The proposed method enables early-stage structural interpretation and supports practical decision-making for well placement and reservoir management in corridor-type NFRs.

1. Introduction

Naturally fractured reservoirs (NFRs) hold a major share of the world’s oil and gas resources and continue to play a vital role in global hydrocarbon production [1]. These reservoirs are marked by natural fractures that can either enhance fluid flow or create barriers, depending on their distribution, connectivity, and interaction with the surrounding rock matrix. Their complex and heterogeneous nature often makes reservoir characterization, well placement, and the design of efficient recovery strategies difficult [2,3]. Within the classification of NFRs, two main types are commonly distinguished: conventional NFRs and corridor-type NFRs, as shown in Figure 1. Conventional NFRs generally have a widespread network of small fractures distributed more uniformly, while corridor-type NFRs contain distinct, high-conductivity fracture corridors separated by low-permeability matrix regions, which are often referred to as exclusion zones [4]. This difference in structure has a significant impact on how these reservoirs should be developed and managed.

In conventional NFRs, optimizing the spacing between wells is key to maximizing drainage and minimizing interference between wells. However, in corridor-type reservoirs, the priority shifts to strategic well placement to target high-conductivity corridors. Wells drilled into these corridors typically show higher productivity because of improved fluid flow, whereas wells drilled into exclusion zones often underperform [6]. Correctly identifying and characterizing fracture corridors is, therefore, essential to designing effective recovery strategies. Traditional techniques such as drilling, coring, and logging have been used to find these corridors, but they are often invasive, costly, and time-consuming [7]. In recent years, well testing, especially pressure transient analysis, has gained attention as a non-invasive and efficient method for detecting fracture corridors and evaluating their properties [8,9,10].

This study‘s objective is to use a simple well-testing technique to identify fracture corridors in naturally fractured reservoirs and to determine the properties needed for the design of effective recovery strategies. In this technique, based on the pressure drawdown pattern, we shall distinguish between fracture and matrix wells, estimate some reservoir properties, and identify important parameters for well placement. Conceptually, using this technique in the well can affect the manner in which the well placement is managed.

2. Methodology

2.1. Well-Placement Strategy in Corridor-Type NFRs

In conventional NFRs, the matrix provides significant storage capacity, while diffuse fractures act primarily as high-permeability pathways that facilitate fluid movement. The fluid transfer between the matrix and fractures follows a dual-porosity or dual-permeability flow model, and it is governed by shape factors that describe the matrix–fracture interaction [11]. Their development strategy involves optimized well spacing to ensure sufficient drainage with minimum well interference. Enhanced recovery methods, such as waterflooding or gas injection, are often applied but require good matrix–fracture connectivity for success.

Corridor-type NFRs, on the other hand, exhibit discrete, highly conductive fracture corridors that dominate fluid flow, which are separated by low-permeability matrix regions that are often called exclusion zones. In these reservoirs, the fracture corridors are laterally extensive and vertically continuous, resulting in strongly anisotropic flow behavior that is highly directional and that is governed by the orientation and spacing of the corridors. Unlike conventional systems, the key to successful development in corridor-type NFRs is strategic well placement rather than well spacing.

The strategy of well placement in the corridor-type NFR depends upon the geological setting. In reservoirs with no bottom water, the preferred well location is the corridors rather than the exclusion (matrix) zone. Wells drilled into fracture corridors, referred to as fracture wells, show superior productivity due to their high permeability pathways, whereas wells drilled into exclusion zones (with diffuse fractures), known as matrix wells, suffer from limited production and delayed pressure support. Moreover, high oil recovery from fracture wells results from the connectivity of fracture corridors, which exhibit permeability at up to two orders of magnitude higher than that of a diffuse fracture matrix, resulting in better sweeps and a higher cumulative recovery [12].

In another study [13], a horizontal well placed within fracture corridors achieved nearly the same recovery as three wells placed in the matrix zone. The authors showed that in corridor-type NFRs, the wells intercepting the corridors (fracture wells) are far more productive, so they need wider spacing (i.e., fewer wells), unlike in conventional NFRs, where denser well spacing improves recovery. (Increasing the number of wells from one to three improved recovery by 10% when fracture corridors were present, compared with a 90% gain when corridors were absent.) Their study further showed that fracture wells with wider spacing (600 ft) matched the recovery of matrix wells with denser spacing (300 ft), significantly reducing the well count and capital investment without compromising performance.

Interestingly, in the bottom-water corridor NFR, the best-performing wells should be drilled into the matrix zone rather than corridors [14]. The effect of well placement was studied using the probability of well location (matrix well or fracture well) of randomly placed wells in a corridor-type NFR statistically equivalent to the actual reservoir with a known distribution of corridors. (The statistically equivalent NFR has uniform spacing and an aperture of the corridors equal to their expected values derived from their known distribution in the actual reservoir.)

The underlying reservoir model for this study [14] was a bottom-water drive system with an oil–water contact at 6040 ft depth, matrix permeability of 40 md, and fracture corridor permeability of 2000 md. The reservoir was simulated using a hybrid approach: a discrete single-porosity model for the near-well zone and a dual-porosity dual-permeability (DPDP) model for the outer reservoir. The wells produced for 20 years without going below the bubble-point pressure of 1000 psi. The results in

Table 1. Effect of well placement on recovery in a corridor-type NFR with bottom water. Reproduced with permission from Samir Prasun and Andrew Wojtanowicz, ASME 2020 39th International Conference on Ocean, Offshore and Arctic Engineering; published by American Society of Mechanical Engineers, 2020 [14].

Fracture Corridor Spacing, ft	Fracture Corridor Size, ft	Well Placement Probability, %		Well Recovery, %
		Fracture	Matrix	Matrix Well	Fracture Well
19	7.9	42	58	40	35
56	15.5	28	72	37	28
100	23.8	24	76	35	23.5
150	31.3	21	79	34.5	22

show that in all cases, matrix wells recover more oil than fracture wells, and the difference increases with corridor size and spacing. The results suggest that in the corridor NFR with bottom-water the best development strategy would be to drill most of the production wells in the matrix while minimizing well placement in the corridors. They are, however, specific to the bottom-water model used in the study and not verified in the field.

In summary, the optimal development strategy of the corridor-type NFRs requires targeted well placement either in the corridors, for reservoirs with no bottom water, or in the matrix (exclusion zone), for bottom-water reservoirs. Since the fracture corridor spacing is distributed throughout the reservoir, the exact location of the wells is not known in advance and can only be determined in a step-wise fashion using well testing. Hence, continuous improvement of the next-well placement, based upon the knowledge gained from the previous wells’ testing, would be the best strategy for maximizing the overall recovery from a corridor-type NFR.

2.2. Simplified Well-Testing Model for Corridor-Type Naturally Fractured Reservoirs

To simulate fluid flow in corridor-type NFRs, the standard dual-porosity or dual-permeability model is insufficient due to the complex heterogeneity of the flow system—micro-heterogeneity of the matrix with diffuse fractures, and macro-heterogeneity from the corridors. However, the well’s inflow is local and can be simplified in three steps.

(1) Using a single-porosity flow model of the matrix with diffuse fractures. In the model, diffuse fractures are combined with the matrix due to their slightly higher permeability and cross-flow effects.

(2) Considering explicitly only the fracture corridors next to the well.

(3) Fusing all other corridors into the matrix system.

This approach balances the need to capture local heterogeneities with the computational efficiency of a larger-scale simulation.

In the discrete model (Figure 2), all fracture corridors are explicitly represented using high-resolution grid blocks. This approach enables detailed modeling of the complex flow behavior associated with high-conductivity corridors and the surrounding low-permeability matrix zones but comes at a significant computational cost. To reduce the complexity while retaining essential heterogeneity, a combined model is employed. In this simplified representation, only the local fracture corridor near the well is modeled explicitly, while all other distant corridors are incorporated into the surrounding matrix using an effective anisotropic permeability formulation.

The numerical model is built using a 223 × 223 × 1 Cartesian grid (49,729 total cells) without corner-point geometry, which provides sufficient spatial resolution for both local heterogeneity and radial flow symmetry. The fracture corridors are assigned a high permeability $K_f$, while the matrix zones are characterized by a homogenized radial permeability $K_r$, calculated from the directional permeabilities using an analytical expression (see Figure 3 and Equation (1)). For fracture wells, the well is fully completed within a fracture corridor, and $K_m$ is used in the surrounding matrix zones. For matrix wells, the well is located between two corridors and completed only within the diffusion zone, with $K_m$ assigned locally and $K_r$ elsewhere.

To simulate the wellbore storage, the volume of the grid cell containing the vertical well is increased. The reservoir is assumed to be single-phase oil and slightly compressible, with constant fluid properties. No capillary pressure, gravity, free gas, or bottom water is considered. Boundary conditions are set as no-flow at the outer edges of the domain, and the initial pressure is uniformly distributed. A constant production rate is applied to the well to initiate pressure drawdown. These assumptions are consistent with conventional pressure transient modeling and enable tractable, interpretable simulation results.

This modeling approach is based on several simplifying assumptions to enable interpretable analysis. The reservoir is assumed to be slightly compressible and single-phase, with constant fluid properties throughout the domain. Capillary and gravitational effects are neglected, and geomechanically influences such as stress sensitivity of permeability are not considered. Fracture corridors are explicitly modeled only near the well, while all other corridors are incorporated into the homogenized matrix system, which is treated as isotropic at the large scale. These assumptions allow for the derivation of analytical diagnostic behavior that captures key flow regimes without requiring full-resolution numerical models. The accuracy and practical relevance of the simplified model are supported by diagnostic consistency with known flow behaviors in fractured systems and that are further validated using numerical simulation in CMG. The resulting pressure responses align with expected flow regime transitions, confirming the model’s ability to differentiate matrix and fracture wells under typical conditions observed in corridor-type NFRs.

The radial permeability $K_r$ of the homogenized matrix system is calculated as the geometric mean of the effective permeabilities in the x- and y-directions (Figure 3). Sheng [15] proposed the corrected formulation for radial permeability, expressed as:

$$K_r=\sqrt{K_x{K}_y}=\sqrt{{\displaystyle \frac{1}{L_t}}\left[K_{fm\_x}{L}_f+K_m\left(L_t-L_f\right)\right]\times {\displaystyle \frac{L_t}{\frac{L_f}{K_{fm\_y}}+\frac{L_t-L_f}{K_m}}}}=\sqrt{{\displaystyle \frac{K_{fm\_x}{L}_f+K_m\left(L_t-L_f\right)}{\frac{L_f}{K_{fm\_y}}+\frac{L_t-L_f}{K_m}}}}$$

(1)

where $K_{f{m}_x}$ and $K_{f{m}_y}$ represent the permeability of fracture corridors in the x- and y-directions, respectively; $K_m$ is the matrix permeability; and $L_f$ and $L_t$ are the lengths of the fracture corridors and the total system, respectively. This treatment captures the anisotropic behavior while preserving computational simplicity.

Flow behavior in corridor-type naturally fractured reservoirs (NFRs) varies significantly depending on the location of the well relative to the fracture corridors. Pressure-transient data provides information for classifying wells as either fracture wells or matrix wells. This distinction is based on the characteristic sequence of flow regimes observed in log–log diagnostic plots, which reflect the dynamic interactions between high-conductivity fracture corridors and low-permeability matrix zones.

For a fracture well—defined as a well completed within a fracture corridor—the pressure response typically follows a distinct pattern (Figure 4). At early times, a short radial flow regime may be observed while the pressure transient is confined within the corridor. This may be followed by a linear flow regime if the fracture corridor has high conductivity and sufficient width. As the pressure front begins to interact with the surrounding low-permeability matrix, a bilinear flow regime may develop. During this period, fluid simultaneously flows linearly through both the high-conductivity fracture corridor and the lower-permeability matrix, resulting in a characteristic slope of 1/4 in the diagnostic plot. At later times, the system may transition into a second radial flow regime as the pressure disturbance expands further into the reservoir. Eventually, the pressure response may exhibit a boundary-dominated flow regime, represented by a unit slope on the log–log plot. It is important to note that not all flow regimes are always present in every test. For example, if the corridor is narrow, the initial radial or linear flow regimes may be too brief to be observed, and in cases with short corridors, the bilinear flow may also be absent.

In contrast, a matrix well—completed in the low-permeability rock between two fracture corridors—exhibits a different pattern, as shown in Figure 5. Initially, the pressure transient propagates through the matrix, producing a radial flow regime. As the pressure front reaches a nearby fracture corridor, the rate of the pressure drop slows, leading to a characteristic trough in the diagnostic curve. This response is due to the sudden transition from a low-permeability matrix to a high-permeability fracture corridor. Following this, bilinear flow may emerge, similar to that in fracture wells, as the pressure front expands concurrently through both domains. A second radial flow regime may follow as the system approaches late-time behavior. As with fracture wells, the exact sequence of regimes depends on the geometry and conductivity of the fracture corridor, the matrix properties, and the well’s distance from the corridor. If the matrix well is located too close to the corridor, the early-time radial regime may not be distinguishable. If the corridor is too short or poorly connected, the bilinear flow regime may not be fully developed.

Figure 4 and Figure 5 illustrate the theoretical diagnostic behavior expected for matrix and fracture wells under ideal flow conditions. These plots are schematic representations derived from analytical trends and are not based on numerical simulation or field data. The slight irregularities reflect regime transitions in conceptual models and are included to highlight key interpretive features. The diagnostic patterns discussed above provide critical insights into the spatial relationship between wells and fracture corridors. By recognizing the sequence and duration of flow regimes in pressure-transient data, operators can distinguish between matrix and fracture wells, estimate key reservoir parameters, and optimize future well-placement strategies.

2.3. Well-Test Analysis and Identification of Fracture Corridors

Well testing has become a primary tool for distinguishing between fracture wells and matrix wells and for characterizing the properties of fracture corridors. The analysis begins with the removal of wellbore storage effects to improve the accuracy of early-time pressure data. Wellbore storage is modeled by increasing the volume of the well block in the simulation software. The $\beta$ deconvolution method is used to correct the well-test data [5,16,17].

Analysis of the log–log diagnostic plots of pressure drawdown ($\Delta p$) versus time identifies the flow regimes associated with fracture wells and matrix wells. For fracture wells, a sequence of radial flow, linear flow, bilinear flow, and second radial flow is typically observed, while matrix wells generally show an initial radial flow, followed by a pressure dip indicative of the transition into the fracture corridor, then bilinear flow, and finally, a second radial flow regime.

Matrix permeability is estimated from the early radial flow stage as

$${\mathrm{K}}_{\mathrm{r}}={\displaystyle \frac{70.6\mathrm{q}{\mathsf{\mu }}_{\mathrm{o}}{\mathrm{B}}_{\mathrm{o}}}{\mathrm{h}{(\mathrm{t}\times ∆{\mathrm{p}}^{\prime })}_{\mathrm{R}1}}}$$

(2)

where $q$ is the flow rate, ${\mu }_o$ is the oil viscosity, $B_o$ is the formation volume factor, $h$ is the formation thickness, and $(t\Delta p^{\prime }{)}_{R1}$ is the derivative of pressure with respect to time during the first radial flow period.

The radius of investigation, $r_i$, representing the distance the pressure transient has traveled, is calculated as

$$D={\mathrm{r}}_{\mathrm{i}}=\sqrt{{\displaystyle \frac{{\mathrm{K}}_{\mathrm{m}}\mathrm{t}}{948{\mathrm{\varnothing }}_{\mathrm{m}}{\mathsf{\mu }}_{\mathrm{o}}{\left({\mathrm{C}}_{\mathrm{t}}\right)}_{\mathrm{m}}}}}$$

(3)

where $K_m$ is the matrix permeability, ${\varphi }_m$ is the matrix porosity, and $C_{t_m}$ is the total compressibility of the matrix. The minimum fracture corridor spacing can then be approximated as twice the radius of investigation.

In the case where the bilinear flow pattern is evident, it is possible to estimate the fracture corridor conductivity $\left(K_f{W}_f\right)$ as

$${\mathrm{K}}_{\mathrm{f}}{\mathrm{W}}_{\mathrm{f}}={\left[{\displaystyle \frac{44.1\mathrm{q}{\mathsf{\mu }}_{\mathrm{o}}{\mathrm{B}}_{\mathrm{o}}}{{\mathrm{m}}_{\mathrm{B}\mathrm{L}}\mathrm{h}{\left({\mathrm{\varnothing }}_{\mathrm{m}}{\mathsf{\mu }}_{\mathrm{o}}{\left({\mathrm{C}}_{\mathrm{t}}\right)}_{\mathrm{m}}{\mathrm{K}}_{\mathrm{m}}\right)}^{1/4}}}\right]}^2$$

(4)

where $q$ is the oil flow rate, ${\mu }_o$ is the oil viscosity, $B_o$ is the oil formation volume factor, $h$ is the reservoir thickness, ${\varphi }_m$ is the matrix porosity, $C_{t_m}$ is the matrix total compressibility, $K_m$ is the matrix permeability, and $m_{BL}$ is the slope of the bilinear portion of the log–log diagnostic plot.

The fracture corridor length, $L_f$, can be estimated depending on the flow regime observed. When the bilinear flow is fully developed and the dimensionless fracture conductivity, $C_{fD}$, is less than 1.6, the fracture corridor half-length can be determined using the bilinear flow time ($\left.t_{ebf}\right)$ as

$${\mathrm{L}}_{\mathrm{f}}=2{\mathrm{L}}_{\mathrm{h}\mathrm{f}}=2{\left({\displaystyle \frac{2.5}{4.55\sqrt{\frac{{\mathrm{K}}_{\mathrm{m}}}{{\mathrm{K}}_{\mathrm{f}}{\mathrm{W}}_{\mathrm{f}}}}\pm \sqrt[4]{\frac{{\mathrm{\varnothing }}_{\mathrm{m}}{\mathsf{\mu }}_{\mathrm{o}}{\left({\mathrm{C}}_{\mathrm{t}}\right)}_{\mathrm{m}}}{0.0002637{\mathrm{K}}_{\mathrm{m}}{\mathrm{t}}_{\mathrm{e}\mathrm{b}\mathrm{f}}}}}}\right)}^2$$

(5)

where $L_{hf}$ represents the half-length of the fracture corridor. This approach depends on the ability to detect the transition from bilinear to late radial flow in the pressure derivative curve.

Alternatively, if boundary-dominated flow is observed after radial flow, the fracture corridor length can be estimated using late-time production and pressure data. The corridor length under boundary-dominated flow conditions is calculated as

$${\mathrm{L}}_{\mathrm{f}}=2{\mathrm{L}}_{\mathrm{h}\mathrm{f}}=\sqrt{{\displaystyle \frac{34.2\mathrm{A}}{{\mathrm{C}}_{\mathrm{A}}{\mathrm{e}}^{\frac{({\mathrm{p}}_{\mathrm{R}}-{\mathrm{p}}_{\mathrm{w}\mathrm{f}}){\mathrm{K}}_{\mathrm{m}}\mathrm{h}}{70.6\mathrm{q}{\mathrm{B}}_{\mathrm{o}}{\mathsf{\mu }}_{\mathrm{o}}}}}}}$$

(6)

where $A$ is the drainage area, $C_A$, is the shape factor associated with the reservoir geometry, and $\left(p_R-p_{wf}\right)$ is the pressure difference between the initial reservoir pressure and the flowing bottomhole pressure. Additionally, the cross-sectional flow constant, $C_F$, is related to $C_A$ by

$${\mathrm{C}}_{\mathrm{F}}={\mathrm{C}}_{\mathrm{A}}/16$$

(7)

The reservoir pressure during boundary-dominated flow can be approximated by

$${\mathrm{p}}_{\mathrm{R}}={\mathrm{p}}_{\mathrm{i}}-{\displaystyle \frac{\mathrm{q}\mathrm{t}}{\mathrm{N}{\mathrm{C}}_{\mathrm{t}}}}$$

(8)

where $p_i$ is the initial reservoir pressure, $q$ is the production rate, $t$ is the time, $N$ is the volume factor, and $C_t$ is the total compressibility.

Nonlinear regression or history matching was not applied in this study because the objective was to qualitatively validate theoretical interpretations and to identify diagnostic flow regimes rather than to quantitatively match parameters. Given the model’s emphasis on the early recognition of fracture connectivity and reservoir heterogeneity, qualitative diagnostic tools provide a more practical and robust solution in the absence of sufficient data for stable nonlinear inversion.

3. Results

3.1. Verification of the Well-Testing Model

To verify the accuracy of the simplified (combined) well-testing model relative to the discrete model (as presented in Figure 2), well-test simulations are compared for both fracture wells and matrix wells using the CMG IMEX VERSION 2015. The reservoir properties used in the comparison are shown in

Table 2. Reservoir properties of an NFR with fracture corridors. Reproduced with permission from Yingying Guo and Andrew Wojtanowicz, ASME 2021 40th International Conference on Ocean, Offshore and Arctic Engineering; published by American Society of Mechanical Engineers, 2021 [5].

Parameter	Numeric Value	Unit
Reservoir Temperature	200	F
Reservoir Top	10,035	ft
Reservoir Thickness	30	ft
Average Reservoir Pressure	4362.66	psi
Matrix Compressibility	$4\times {10}^{-6}$	1/psi
Matrix Porosity	0.1	Fraction
Matrix Permeability	10	md
Radial Permeability *	10.566	md
Fracture Corridor Permeability	10,000	md
Reservoir Wettability	Oil wet
Oil Formation Volume factor	1.1	rb/stb
Oil Viscosity	7	cp
Oil Density	56	lb/ft3
Oil Compressibility	$1.5\times {10}^{-6}$	1/psi
Production Rate	10	stb/d
Reservoir Temperature	200	F

* For the simplified (combined) model in Figure 2.

.

Figure 6 and Figure 7 present the diagnostic plots for both types of wells in the two models. The pressure responses show a strong match, confirming the equivalence between the two models for the tested scenarios. This agreement is attributed to the close similarity between the radial permeability of the corridor-type NFR and the matrix permeability. As the fracture corridors occupy only a small fraction of the reservoir area and are widely spaced, their influence on global permeability remains minor. Thus, the simplified model proves to be sufficient for describing well performance in corridor-type NFRs.

3.2. Example Application of the Well-Testing Analysis

The application of the well-testing analysis method developed in this study is demonstrated using the corridor-type NFR configuration described in

Table 3. Properties of a corridor-type NFR. Reproduced with permission from Yingying Guo and Andrew Wojtanowicz, ASME 2021 40th International Conference on Ocean, Offshore and Arctic Engineering; published by American Society of Mechanical Engineers, 2021 [5].

Parameter	Value	Unit
Reservoir Temperature	200	F
Reservoir Top	10,035	ft
Grid Size in the Vertical Direction	30	ft
Grid Size in the Horizontal Direction	15, (100, 2000, 3000)	ft
Initial Average Reservoir Pressure	4362.66	psi
Rock Compressibility	$4\times {10}^{-6}$	1/psi
Water Compressibility	$3.2\times {10}^{-6}$	1/psi
Oil Compressibility	$1.5\times {10}^{-6}$	1/psi
Matrix Porosity	0.01	Fraction
Fracture Corridor Porosity	0.001	Fraction
Matrix Permeability	1	md
Radial Permeability	1.06	md
Fracture Corridor Permeability	10,000	md
Oil Formation Volume factor	1.1	rb/stb
Oil Viscosity	7	cp
Production Rate	10	stb/d
Fracture Corridor Spacing	200	ft
Fracture Corridor Length	1500	ft

.

3.2.1. Well-Placement Recognition

Figure 8 presents a diagnostic plot for a fracture well with and without wellbore storage effects. Figure 9 shows the equivalent plot for a matrix well. There is a very small effect of wellbore storage on the pressure transient in the fracture well due to the rapid pressure response in the high-conductivity fracture corridor. In contrast, for the matrix well, wellbore storage substantially distorts the early-time pressure response, as shown in Figure 9. Moreover, the matrix well plot clearly displays a U-shaped portion representing the pressure transient from the flow reaching the nearby fracture corridor. In addition, the pressure drawdown is much smaller in the fracture well than in the matrix well, highlighting the role of the fracture corridor in enhancing the flow capacity.

3.2.2. Estimation of Fracture Corridor Spacing

The shape of the plot in Figure 8 indicates the well placement inside the corridor. For the matrix well’s test shown in Figure 10, the distance is 96 feet, estimated from Equation (3) as the radius of investigation after 3.35 h of flow. The relative error of the estimation is 4.16 percent. The minimum fracture corridor spacing is 192 ft—i.e., twice the radius of investigation.

3.2.3. Conductivity of the Fracture Corridor

A complete bilinear flow stage is shown in the diagnostic plots in Figure 11 for a fracture well and two matrix wells located near fracture corridors, with the bilinear flow segment’s slope (Figure 12) equal to 113. Using the bilinear flow equation, the fracture corridor conductivity is calculated as 1617 md-ft. The result confirms that both well types (in the corridor or close to the corridor) can provide sufficient data to estimate the fracture corridor conductivity when bilinear flow is present and wellbore storage has been removed.

3.2.4. Estimation of the Fracture Corridor Length

The fracture corridor length is computed using the value of the slope of the bilinear flow stage plot. For two estimated values of the transition from bilinear to late radial flow, Equation (5) gives two results: 1775 feet and 1207.6 feet. The dimensionless fracture conductivities, C_fD, corresponding to each length are also calculated as 1.8 and 2.7, respectively [5,16]. The method is applicable for dimensionless conductivity values of C_fD ≤≤ 1.6. From Figure 12, the closest value of C_fD = 1.8 is selected. Thus, the best estimation of the corridor’s length is 1775 ft. Considering the actual value of the length 1550 ft (in Table 3), the estimation error is 15.5 percent. Clearly, the bilinear flow method provides only an approximate estimate of corridor length. A more accurate estimation would require prolonged well testing to reach the boundary-dominated flow stage in Figure 4. This approach has been used in another study with a reportedly small error of 1.14 percent [11].

3.3. Accuracy of Fracture Corridor Detection

The accuracy of the proposed well-testing method for finding the distance from a matrix well to the nearest fracture corridor is appraised by simulating a matrix well away from the fracture for various combinations of matrix permeability and fracture corridor length. The estimation’s accuracy is quantified with relative error $\delta$, defined as

$$\delta =\left|{\displaystyle \frac{v_A-v_E}{v_E}}\right|\times 100\%$$

(9)

where

• $v_A$ = the tested distance value from the well test,
• $v_E$ = the actual distance from the simulation.

The relative errors are grouped by their size $(\mathsf{\delta }\le 10\mathrm{\%}$, $10\%<\delta \le 20\%$, and $\delta >20\%$) using the cumulative logit model. The model predicts the probability of a certain-size error as a function of the reservoir matrix permeability $\left(K_m\right)$ and fracture corridor length ($L_f$). The resulting probabilities are

$$P\left(\mathsf{\delta }\le 10\mathrm{\%}\right)={\displaystyle \frac{\mathrm{e}\mathrm{x}\mathrm{p}(-0.378-0.254{ K}_m+0.001{ L}_f)}{1+\exp \left(-0.378-0.254{ K}_m+0.001{ L}_f\right)}}$$

(10)

$$\begin{gathered} P\left(10\%<\delta \le 20\%\right)={\displaystyle \frac{\mathit{exp}\left(1.097-0.254 K_m+0.001 L_f\right)}{1+\mathit{exp}\left(1.097-0.254 K_m+0.001 L_f\right)}} \\ -{\displaystyle \frac{\mathit{exp}\left(-0.378-0.254 K_m+0.001 L_f\right)}{1+\mathit{exp}\left(-0.378-0.254{ K}_m+0.001 L_f\right)}} \end{gathered}$$

(11)

$$P\left(\delta >20\%\right)=1-{\displaystyle \frac{\mathit{exp}\left(1.0965-0.2543K_m+0.000945L_f\right)}{1+\mathit{exp}\left(1.0965-0.2543K_m+0.000945L_f\right)}}$$

(12)

4. Discussion

The study offers a simplified well-testing-based method for characterizing fracture corridors in corridor-type naturally fractured reservoirs (NFRs). It shows that accurate interpretation of pressure-transient data may distinguish fracture wells from matrix wells and enable the estimation of key reservoir parameters such as permeability and fracture corridor spacing, which could ultimately improve the recovery strategy.

The findings emphasize the importance of identifying fracture corridors early in the development process to decide well placements. As already demonstrated by simulations and field examples, reservoirs in which bottom-water wells do not intersect fracture corridors exhibit markedly better productivity than those completed in matrix zones, while bottom-water reservoirs require wells to be placed in the exclusion (matrix) zone. This observation aligns with previous studies, such as that of Singh et al. (2009), who highlighted the role of high-resolution seismic imaging in detecting fracture corridors and in guiding well placement in carbonate reservoirs [18].

Traditional development strategies based on uniform well spacing are insufficient in corridor-type systems. Instead, the results suggest that well placement should be guided by an understanding of fracture corridor geometry and spatial distribution. This requires a shift toward integrated, multidisciplinary approaches that combine geological, geophysical, and engineering data. For example, linking core analysis, seismic interpretation, and well-test results can provide a more complete picture of fracture connectivity, spacing, and orientation—factors that conventional modeling approaches may overlook.

While this integrated approach is strongly recommended, we acknowledge that implementing such multidisciplinary workflows can be resource-intensive and may not be practical in all field scenarios—particularly in early-stage developments or marginal fields. However, even partial integration of such methods, such as combining well testing with basic seismic interpretation, can significantly improve fracture corridor detection and reduce uncertainty. The flexibility to scale the approach based on available data and project scope allows for broader applicability in real-world settings.

One of the significant contributions of this study is the validation of a simplified combination model that accurately replicates the behavior of more complex discrete fracture systems. This model, when used in conjunction with pressure-transient analysis, enables a reliable estimation of matrix permeability, fracture corridor conductivity, and corridor spacing. Importantly, the results show that boundary-dominated flow regimes provide more accurate estimates of fracture corridor length than bilinear flow analysis, which may produce large errors if the corridor conductivity is over- or underestimated.

In a broader context, the ability to estimate the distance from a matrix well to the nearest fracture corridor—and to do so with reasonable accuracy—has practical implications for optimizing field development. The statistical analysis presented in this work confirms that the estimation accuracy improves when fracture corridors are long and the matrix permeability is low. This highlights the value of using cumulative logit models or other probabilistic tools to quantify uncertainty and to guide decision-making under geologic complexity.

The implications of these findings may extend beyond the initial well placement. They also inform the application of enhanced oil recovery (EOR) strategies in fractured systems. The success of waterflooding or gas injection depends strongly on how injected fluids interact with fracture corridors. If the fracture spacing is too wide or the matrix permeability too low, sweep efficiency may decline. Thus, tailoring EOR methods to the unique fracture architecture of corridor-type NFRs is essential for improving the recovery factor.

While this study highlights the importance of such interactions, it does not explicitly address multiphase flow behavior or EOR-specific dynamics. Future work should incorporate fluid-flow mechanisms—such as sweep efficiency, phase mobility, and injection response—to extend the diagnostic method’s applicability to EOR planning and reservoir performance forecasting.

This study assumes constant permeability and porosity during the well-testing analysis. However, it is well recognized that these properties may exhibit stress sensitivity, particularly in tight or unconsolidated formations. Ignoring this behavior could lead to deviations in the pressure response, especially under high drawdown conditions, where a reduction in permeability may dampen or delay flow regime transitions. Although stress effects are beyond the current model’s scope, acknowledging their presence is important. Future work could incorporate stress-dependent reservoir properties to improve the diagnostic accuracy in geomechanically sensitive environments.

The effectiveness of the proposed diagnostic technique may vary depending on reservoir properties such as fracture spacing and matrix permeability. When fracture corridors are widely spaced or the matrix permeability is high, the pressure signal may weaken or distort, making flow regime identification more difficult. These limitations should be considered when applying this method in the field, especially in structurally complex or heterogeneous reservoirs.

It is important to note that in practical field applications, complete flow regimes may not always be observed due to constraints such as limited test duration, well interference, or complex reservoir boundaries. Even when full regime transitions are not clearly visible, partial features in pressure and derivative plots—such as early-time slopes or transitional inflections—can still offer valuable diagnostic insights. The proposed method is therefore intended as an initial screening tool rather than a definitive mapping technique. Its effectiveness lies in identifying relative flow regime transitions that suggest proximity to fracture corridors. However, in geologically complex reservoirs, this approach should be complemented by additional data sources such as seismic interpretation, core analysis, and image logs to reduce uncertainty. When integrated with other forms of geological and geophysical evidence, the method can support more reliable well-placement decisions and early-stage reservoir evaluation.

Future work could focus on validating the proposed methodology using field data from multiple reservoir types and geological settings. In particular, extending the statistical framework to incorporate the uncertainty from seismic interpretations, petrophysical measurements, and dynamic production data would enhance its reliability in operational decision-making. Advanced techniques such as Bayesian inference, Monte Carlo simulation, or bootstrapping could be employed to quantify the parameter variability and improve probabilistic predictions. Additionally, analyzing the interaction of multiple wells within fracture-rich environments may offer further insights into optimizing well-placement strategies in structurally complex reservoirs. Additionally, investigating the interaction of multiple wells in fracture-rich environments could further refine placement strategies in complex structural settings.

In summary, this study provides a framework for characterizing corridor-type NFRs through well testing, with clear implications for development planning, reservoir modeling, and recovery optimization. As reservoirs become more geologically complex and operators face higher development costs, the ability to extract meaningful information from early well data becomes increasingly valuable.

5. Conclusions

Corridor-type naturally fractured reservoirs (NFRs) represent a unique flow system dominated by discrete, high-conductivity fracture corridors embedded within a low-permeability matrix. This study demonstrates that the development of such reservoirs requires different strategies from those applied in conventional NFRs, where pervasive fracture networks are more uniformly distributed. The conclusions are as follows:

• Strategic well placement, rather than uniform well spacing, is the most critical factor for maximizing productivity in corridor-type systems. In reservoirs with no bottom water, wells must be intentionally located to intersect fracture corridors, as those drilled in matrix zones typically suffer from low deliverability due to limited permeability and delayed pressure support. In contrast, in bottom-water reservoirs, wells should be located in the exclusion (matrix) zone.
• Using the simplified model for pressure-transient analysis, it is possible to distinguish between fracture and matrix wells, estimate the matrix permeability, assess the corridor spacing, and approximate the corridor conductivity and length. The findings further show that boundary-dominated flow analysis offers a more accurate method for estimating the fracture corridor length than bilinear flow diagnostics.
• Maximizing the full productivity potential of corridor-type NFRs requires the step-wise development of a well-placement strategy based on the well’s pressure testing on a well-by-well basis. Since the fracture corridor spacing is distributed throughout the reservoir, the exact location of the wells is not known in advance and can only be determined in a step-wise fashion using well testing. Hence, continuous improvement of the next-well placement, based upon the knowledge gained from the previous wells’ testing—would be the best strategy for maximizing the overall recovery from a corridor-type NFR.

While this study focuses on corridor-type NFRs, the proposed methodology should be further validated using field data from reservoirs with varying structural and geological characteristics. Future work will explore the method’s adaptability and reliability in more complex settings, including reservoirs with irregular fracture networks, aquifer support, or multiphase flow conditions.