Severe wellbore instability in a complex lithology formation necessitating casing while drilling and continuous circulation system

Abstract

Most of drilling hole problems are attributed to wellbore stability issues which adversely cause excessive lost time and cost millions of dollars. The past drilling experiences in Kupal oilfield showed excessive mud losses, kick flows, tight holes and pipe stuck leading to repeated reaming, fishing and sidetracking. Most of the drilling-associated problems in this field occurred during drilling the 12 ¼-in. hole, which is across the non-reservoir Gachsaran formation (consisting of anhydrite, gypsum and marl with thin limestone layers). Mainly due to the lack of required formation evaluation data, no geomechanical studies of this formation have been conducted to date. In this work, first, we constructed a geomechanical model to investigate the root of the problems. This is a pioneer wellbore stability work for such a complex lithology formation which included finding the equations best-matching with core data and field observations. Finally, to overcome the field challenges and hole problems, the study proposes some field remedial actions. The results of the geomechanical modeling show that the pore pressure, shear and tensile failure gradients are greatly variable with the safe mud weight window becoming excessively narrow at some intervals. This accounts for the encountered wellbore stability issues as managing the mud weight in these situations requires several casing strings. To mitigate the extent of the problem, this study proposes the application of innovative drilling technologies including casing while drilling to eliminate the casing running time with potential reduction in drilling time, and continuous circulation system to prevent cuttings settling and kick flows during connections. These technologies are capable of elimination of the geomechanical part of the drilling delay (30% of the average 77 drilling days) per well.

Introduction

Wellbore instability is one of the most critical challenges affecting not only the well construction phase, but the entire life cycle of a well. It is one of the major causes of non-productive time (NPT) by causing issues such as borehole collapse, lost circulation, stuck pipe, sand production and other related well failure events. The NPT is any event that interrupts the progress of a planned operation causing a time delay; it includes the total time needed to resolve the problem until the operation is resumed again from the point or the depth where the NPT event occurred.

The NPT typically accounts for up to 32% of drilling operations costs for deep-water wells (Halliburton 2016). Further, the geomechanical problems are associated with 40% of the drilling-related NPT in deep-water and other challenging environments (Schlumberger 2016). The total cost of geomechanics-related issues is multibillions of dollars (Aadnoy 2003). Fifty percent of this NPT is associated with geomechanics-related issues (stability, lost circulation, stuck pipe, etc.) accounting for nearly 11% of the drilling budget (Mody 2013).

In summary, the unexpected instability events increase risk, reduce safety, potentially harm crew and cause non-productive time. Moreover, they are costly and can easily lead to a cost overrun if they occur frequently. Successful construction of wells containing trouble zones depends on accurate analysis of all available well data to deliver the well and its objectives. Being familiar with the local drilling environment can substantially reduce risk. Unfortunately, the drilling personnel often ignore the data and the corresponding learning observations from previous well construction attempts. The next well design is left unchanged expecting different results than what obtained from the previous failed attempt. Although this approach is illogical, it has too often been the norm in many offshore environments as proven by the amount of money spent on avoiding issues in drilling known and expected trouble zones (York et al. 2009). From this perspective, in this study, we investigated the root causes of wellbore instability in a critical Middle Eastern oilfield where severe drilling-associated problems occurred. The end-of-well reports of this field demonstrate that multiple problems related to wellbore instability occurred while drilling the 12 ¼-in. hole section (across Gachsaran formation), and they include severe mud losses, pit gains, tight holes, etc. There are currently no investigations of the problems of this field available in the literature. Since most of these instability events stem from geomechanical reasons, analyzing the geomechanical situation can help gain knowledge about when and where instability could occur and how to mitigate it.

In order to investigate the problems, first, the paper discusses the records of the drilling and non-productive-time (NPT) issues encountered in the field. Next, to associate the hole problems with geomechanical reasons, the mechanical earth model (MEM) of Gachsaran formation will be built and presented. Using the results of the MEM, wellbore stability analysis will be conducted and the geomechanical origin of the borehole problems including the narrow safe mud weight window will be detected. This is a pioneer wellbore stability work for a Gachsaran non-reservoir formation (with a complex lithology of anhydrite, gypsum and marl, with thin limestone layers) which is responsible for most of the wellbore stability issues and lost times (NPT). However, previous wellbore stability works in the region such as Aslannezhad et al. (2016) focused on the carbonate reservoir formations (which is not responsible for much lost times). Other wellbore stability works focused on sandstones such as Kolawole et al. (2018). There are several wellbore stability works on shales in recent years (Wang et al. 2014).

Finally, to mitigate the borehole problems, some potential solutions will be presented and discussed. These consist of application of innovative drilling techniques casing while drilling (CWD) and the continuous circulation system (CCS) to merge with the conventional method.

Field of study

Figure 1 shows the map of the area where Kupal oilfield is located in the Zagros region in the Middle East. Severe drilling problems are encountered particularly during drilling wells in this field. The majority of hole problems in this field occur during drilling Gachsaran formation overlying the reservoir. This formation (which has a complex lithology consisting of anhydrite, gypsum, marl and thin limestone layers) is extremely abnormally pressured. Severe hole problems in drilling this formation are generally attributed to thick trouble zones, extremely high abnormal pore pressure (causing great differential pressure leading to differential pipe sticking). The necessity of overcoming extremely high pore pressure requires the use of greater percentage of weight materials (which in turn causes cuttings settling and stuck pipe). Furthermore, greater concentration of fractures than nearby fields causes more severe mud losses and lost circulation. The following subsections briefly review the commonly observed drilling problems (excessive mud losses and tight holes or stuck pipe) as observed in the field as well as indications of low-performance drilling (ROP and NPT).

Fig. 1

The location of Kupal oilfield in the Middle East

Excessive mud losses or possible gains

In drilling the 12 ¼-in. hole (Gachsaran formation), excessive mud losses and gains occur intermittently. Mud loss usually occurs due to induced-fracturing of the thin limestone layers interbedded in this formation. The pore pressure of Gachsaran formation, as will be estimated and discussed in the “Wellbore Stability Analysis” section, is excessively abnormal and close to 1 psi/ft (or greater). This is mainly attributed to the conversion of gypsum (CaSO₄·2H₂O) to anhydrite (CaSO₄) releasing reaction water. Having the reservoir in a closed basin, the released water which has moved to the interbedded limestone layers becomes excessively over-pressured during geologic time. Therefore, when mud weights are not selected and adjusted appropriately, kick flows may occur.

This study chose 28 wells with reported mud losses for performing in-depth analysis (see Fig. 2). Several facts can be extracted from this figure, as follows:

Fig. 2

Mud loss magnitudes occurring in Gachsaran formation in different wells of Kupal field

• 1. The minimum mud loss occurred in well Z2 with 74 bbls, and the maximum mud loss occurred in well Z30 with 29,728 bbl.

• 2. Twenty-three wells (82% of the total wells) reported mud losses greater than 300 bbl.

• 3. It was observed that the magnitude of mud losses has dramatically increased during 2011–2017 (average of 6994 bbl per well) compared to the earlier time with the average loss of 1085 bbl per well.

In this field, several methods were tried aiming at mitigating the extreme mud losses. One of these methods was performed by maintaining the dynamic mud pressure, i.e., equivalent circulating density (ECD) greater than the formation pressure while the mud column pressure (hydrostatic pressure) is lower than the formation pore pressure. Finally, at the end of drilling, prior to tripping the drill string, they proceed to increase the (static) mud weight so that no kick flows may occur during tripping (although huge mud losses occurred and the well had to be filled-up). However, the main problem experienced in such instances was during connections when kick flows could occur, which made the crew expedite in making the connection (an unsafe practice). This study provides a solution to this practice, which can be successfully applied to most drilling operations in Kupal field.

In contrast, the pit gain, due to kick flow, is reported in only seven wells with the maximum value of 200 bbl in wells Z26, Z30 and Z32, as presented in Fig. 3. Similar to the mud loss, the pit gain shows a significant increase during 2011–2017 (arithmetic average of 60 bbl per well) compared to the average of 18.5 bbl per well prior to this period. In this period, the significant anomalous increase in mud losses and decrease in drilling performance occurred due to improper drilling equipment used recently. Generally, some cases of kick flows were not reported because they were underground blowout (not observed or noticed at the surface), or for other reasons. An underground blowout is an occurrence of flow from a high-pressured formation which loses to an underground thief zone (low-pressured), which may not necessarily be observed at the surface.

Fig. 3

Pit gains in Gachsaran formation in different wells of Kupal field

Review of the field reports indicates that in cases of mud loss where the driller lowered the mud weight for a small fraction of 0.025 ppg, it resulted in a noticeable kick flow. This suggests a greatly narrow safe mud weight window for drilling, justifying the initiative to conduct the geomechanical studies in this field. With the formation evaluation data becoming available from a recent well logging and coring into the Gachsaran formation, building the one-dimensional mechanical earth model (MEM) will now be possible.

Tight holes and stuck-pipe

In Kupal field, tight holes frequently occur in drilling the 12 ¼-in. hole, followed by stuck pipe in some cases. The crew periodically conducts wash-and-reaming following drilling operations. As another consequence, the casing string becomes stuck during running in the hole (RIH) in numerous cases. Stuck occurrences in drilling this formation has led to costly operations such as long fishing operations or even sidetracking in many cases.

Low drilling performance

The average ROP values (m/h) of drilling the 12 ¼-in. hole section for 35 wells are presented in Fig. 4. The following can be inferred from the figure:

Fig. 4

The average ROP of drilling the 12 ¼-in. hole section (across Gachsaran formation) in 35 wells of Kupal field

• 1. The average ROP of seven wells (i.e., 20% of the total number of wells) is greater than 7.25 m/h, whereas in 20 wells (i.e., 57% of the total number of wells) the ROP is lower than 5 m/h.

• 2. The ROP is lower than 1.5 m/h for six wells drilled during 2011–2017 period (17% of the total number of wells).

• 3. These data show the inefficiency in drilling due to non-optimized drilling parameters and methods, use of inefficient equipment and materials including drilling bits, and inappropriate mud pumps. However, the tectonic activity of the field requires a full geomechanics study to identify the technical parameters resulting in inefficient drilling. Therefore, the drilling parameters require improvement to yield optimal ROPs, and the drilling method should be modified (to a more suitable one).

• 4. Comparing the ROP chart with the mud loss chart in the 12.25 hole size across Gachsaran formation, it is seen that the lesser the mud losses/gains, the greater the ROP. Some examples of excessively low ROP are observed in wells Z14 (loss), Z26 (gain), Z30 (loss), Z32 (loss) and Z33 (loss). The maximum average ROP is in well Z4 with 10.6 m/h.

Drilling delays

Figure 5 shows the drilling delays/time deviations for 32 investigated wells (nearly half of the total drilled wells in Kupal oilfield). As the selected wells targetted Asmari reservoir formation below Gachsaran formation, the drilling delays are related to the hole problems during drilling Gachsaran formation (12 ¼-in. hole). The drilling delay for each well (in rig days) is found by subtracting the predicted rig days from the actual rig days. Table 1 shows the average delays for three periods of pre-1980, 1980–2011 and 2011–2017. As Fig. 5 and Table 1 show, a moderate increase in the delay is seen from 60 days for pre-1980 period to around 48 days for the 1980–2011 period. However, it abruptly increased to around 173 days for the period of 2011–2017 due to various drilling issues. The average drilling delay/deviation in this field is 77 days. About 30% of the drilling delays have occurred due to geomechanical reasons, but the rest was due to waiting for equipment or materials.

Fig. 5

Drilling delays/time deviations (for 32 wells in Kupal field)

Table 1 The average drilling delays/deviation in Kupal oilfield for the three periods of: (a) pre-1980, (b) 1980–2011 and (c) 2011–2017

In the following section, the study presents the wellbore stability analysis conducted in this field followed by proposed methods of drilling optimization.

Wellbore stability analysis

When a borehole is being drilled, the in situ stress state is disturbed. Depending on the field, it can potentially cause borehole stability issues (such as kick flows, tight hole, stuck pipe, breakouts and lost circulation), leading to NPT or drilling NPT. Using the data from well Z26 (the only well with data from Gachsaran formation), the 1-D MEM was constructed to evaluate the safe mud weight window. Figure 6 shows the schematic of well Z26. This is explained in detail in the following sections.

Fig. 6

Schematic of well Z26 of Kupal oilfield (for the 12 ¼-in. hole across Gachsaran formation)

Input data and parameters

Table 2 shows the input data required for building the 1D-MEM. It includes well trajectory, petrophysical logs, drilling and mud parameters, core data and formation integrity data (leak-off or formation strength test). The interpreted petrophysical data are shown in Fig. 7. The core laboratory results available at three depths are provided in Table 3 with drilling and mud data presented in Table 4. The interpretation of the cuttings analysis is given in Table 5. The recorded data for the mud loss and kick flow occurrences are, respectively, listed in Tables 6 and 7. The image log of well Z26 (for the underlying reservoir formation) was available for investigating the directions of the horizontal in situ stresses. The breakouts (with an example in Fig. 8) indicate the direction of minimum horizontal stress of N75W (285 degrees from North), and the drilling-induced fractures indicate the maximum horizontal stress of N15E (15 degrees from North).

Table 2 Input data used for constructing the MEM in well Z26 and in which evaluations, they were used

Fig. 7

a Interpreted petrophysical data and mud weight required for MEM construction in well Z26 (2700–2950 m). The depth intervals of the members of Gachsaran formation have been indicated. b Interpreted petrophysical data and mud weight required for MEM construction in well Z26 (2950–3200 m)

Table 3 Core laboratory results (Gachsaran formation)

Table 4 Drilling and mud data corresponding to Gachsaran formation (in Kupal field) at some main depths

Table 5 The results of graphical well log (obtained from geologist’s cuttings analysis) of well Z26 (from top to bottom). It includes Gachsaran formation including its seven member/layers

Table 6 Mud loss occurrences in well Z26 of Kupal field (used for calibration of the $\sigma_{h}$ estimations by MEM). These specified depths can be indicated for calibration of the minimum horizontal stress

Table 7 Kick flow occurrence data (in member-5 of Gachsaran formation in well Z26)

Fig. 8

An example breakout indicating the direction of minimum horizontal stress of N75W

At some intervals, there were erroneous wireline well log data (e.g., due to malfunction of the centralizers at some intervals) which were filtered. These data were detected using the below detailed observations and replaced by average data of the overlying and underlying readings.

• 1. The data with extremely low Young’s modulus ($E$), Poisson’s ratio ($\nu$) and uniaxial compressive strength (UCS).

• 2. The data that resulted in tensile failure mud weight lower than shear failure mud weight.

MEM development workflow

Figure 9 and Table 8 show the workflow of construction of the MEM and determination of the mud weight windows. The process involves estimation of rock elastic moduli and strength properties as well as the pore pressure and in situ stresses. This is considered a pioneer work of MEM development for such a complex lithology formation (consisting of anhydrite, gypsum, marl and limestone) which included finding the equations best-matching with core data and field observations. The MEM workflow is discussed in the following subsections.

Fig. 9

Constructing the mechanical earth model (MEM) for WBS providing induced stresses and safe mud window

Table 8 Steps and parameters in the procedure for constructing the MEM

Rock moduli and strength

The dynamic Young’s modulus $E_{\text{d}}$ (GPa) is estimated using (Fjaer et al. 2008):

$$E_{\text{d}} = 10^{6} \left[ {\frac{{\rho_{\text{b}} }}{{\Delta t_{\text{s}}^{2} }}} \right]\left[ {\frac{{3\Delta t_{\text{s}}^{2} - 4\Delta t_{\text{c}}^{2} }}{{\Delta t_{\text{s}}^{2} - \Delta t_{\text{c}}^{2} }}} \right]$$

(1)

where $\rho_{\text{b}}$ is the bulk density (g/cm³); $\Delta t_{\text{c}}$ is the compressional wave slowness (μs/m); $\Delta t_{\text{s}}$ is the shear wave slowness (μs/m).

Based on van Lama and Vutukuri (1978), Barton (2006) and Holt (2013), the static moduli differ from the dynamic moduli. This is because the dynamic measurements are made in a certain (seismic/sonic/ultrasonic) frequency band, strain and stress amplitudes, whereas static moduli represent behavior at very low frequency band, 0.01 1/s or lower, and large amplitudes (Holt 2013). Based on Plona and Cook (1995) and TerraTek Inc. (1998), the dynamic Young’s modulus is often 2–3 times that of the static one. In this paper, for better matching of the static modulus with the core laboratory results, the dynamic Young’s modulus E_d is converted to the static one E_s using the following equation (which was modified from Wang 2000 or Plona and Cook 1995 to match with the core data):

$$E_{\text{s}} = 0.46E_{\text{d}}$$

(2)

The results of E_s were calibrated using the core test data at three different depths (see Table 3).

Next, the dynamic Poisson’s ratio $\nu_{\text{d}}$ is estimated using shear and compressional slowness times, $\Delta t_{\text{s}}$ and $\Delta t_{\text{c}}$ (Fjaer et al. 2008):

$$\nu_{\text{d}} = \frac{1}{2}\frac{{\left( {\frac{{\Delta t_{\text{s}} }}{{\Delta t_{\text{c}} }}} \right)^{2} - 2}}{{\left[ {\left( {\frac{{\Delta t_{\text{s}} }}{{\Delta t_{\text{c}} }}} \right)^{2} - 1} \right]}} = \frac{1}{2}\frac{{\Delta t_{\text{s}}^{2} - 2\Delta t_{\text{c}}^{2} }}{{\left[ {\Delta t_{\text{s}}^{2} - \Delta t_{\text{c}}^{2} } \right]}}$$

(3)

Assuming an elastic medium, based on Larsen et al. (2000), the non-elastic shear deformation coefficients vanish and thus the dynamic and static Poisson’s ratios are equal ($\nu_{\text{s}} = \nu_{\text{d}}$).

Different UCS correlations for the uniaxial compressive strength (UCS) listed in Chang et al. (2006) were checked. In this work, Militzer and Stoll’s correlation (1973) was found as the best-matching correlation with the core data:

$${\text{UCS}} = \frac{1}{145}\left( {\frac{7682}{{\Delta t_{\text{c}} }}} \right)^{1.82}$$

(4)

This equation was calibrated against core test results (see Table 3).

The tensile strength of essentially all rocks is quite low, on the order of just a few MPa (Lockner and Byerlee 1995). Therefore, in this work, it is considered 1/10 of the UCS (Jaeger et al. 2007).

Next, as the internal friction angle $\varphi$ is required for determination of the shear failure pressure/equivalent mud weight, Plumb correlation (Plumb 1994) is commonly used to estimate it:

$$\varphi = 26.5 - 37.4\left( {1 - \emptyset - V_{\text{shale}} } \right) + 62.1\left( {1 - \emptyset - V_{\text{shale}} } \right)^{2}$$

(5)

In this correlation, $V_{\text{shale}}$ (shale volume) is found using the gamma log measurements ($V_{\text{shale}} = \frac{{{\text{GR}} - {\text{GR}}_{ \text{min} } }}{{{\text{GR}}_{ \text{max} } - {\text{GR}}_{ \text{min} } }}$); $\emptyset$ is neutron porosity which is estimated by the neutron log porosity (${\text{NPHI}}$). In case ${\text{NPHI}}$ is not available, it can be estimated using its correlation with UCS considering the lithology (Chang et al. 2006).

Based on the estimation for the Gachsaran formation, in this work, the angle $\varphi$ ranges from 21 to 43 degrees, and its average magnitude is 23 degrees.

The ranges of the estimated rock moduli and strength are given in Table 9.

Table 9 The ranges of estimated static elastic moduli (including the bulk modulus, $K_{b}$ and shear modulus $G$) and rock strength of Gachsaran formation in Kupal field (using MEM results of well Z26)

Pore pressure

Measurement of the pore pressure is typically conducted by repeat formation tester (RFT), modular dynamics testing (MDT), well test analysis, logging while drilling sensor, etc. However, as the pore pressure of Gachsaran formation is excessively abnormal and its permeability is so low at some intervals (with high percentage of anhydrite/gypsum or salt), direct measurement was not practically possible. Therefore, in this work, it was indirectly estimated using correlations, which was calibrated using real pore pressure value(s) during kick flows (also indicated by calcium ion increase in the mud). Generally, some pore pressure correlation methods include the drilling exponent, seismic data (Dugan and Flemings 1998), sonic log using Eaton’s or Bowers method (Bowers 1994; Zhang 2011), or resistivity log (Zhang 2011; Mouchet and Mitchell 1989). These methods are based on the fact that the porosity of shale is expected to decrease monotonically as the vertical effective stress ($\sigma v - \alpha P_{\text{p}}$) increases, which are susceptible to some complexities (Zoback 2010). Alternative methods are the modified-drilling exponent methods (Rehm and McClendon 1971; Zamora 1972; Bourgoyne et al. 1986). In this work, Zamora’s model showed best-matching results (with field observations data) and was used for estimation of pore pressure.

As the pore pressures derived from logs and other sources are not accurate, calibration of the pore pressure estimations is essential (Aadnoy and Looyeh 2010). Therefore, in this work, the kick flow data were used for calibration of the estimated pore pressures. For this purpose, at the depth of 3076 m, Table 7), the measured surface shut-in drill pipe pressure (SIDPP) was added to the old mud pressure to find the real pore pressure, which is used to calibrate the estimations.

Modified D-exponent method (Zamora 1972):

In this method, first, the d-exponent, $d_{\exp }$ is found as follows:

$$d_{\exp } = \frac{{\log \left\{ {\frac{\text{ROP}}{{60 {\text{RPM}}}}} \right\}}}{{\log \left\{ {\frac{{12 {\text{WOB}}}}{{1000d_{\text{B}} }}} \right\}}}$$

(6)

Next, the modified d-exponent $d_{\rm mod}$ is found:

$$d_{\bmod } = \frac{{\rho_{\text{n}} }}{{\rho_{\text{e}} }}d_{\exp }$$

(7)

In the above equations, ${\text{ROP}}$ is the rate of penetration (ft/h); ${\text{RPM}}$ is the revolutions per minute, ${\text{WOB}}$ is the weight on the bit (k.Ibf); $d_{\text{B}}$ is the bit size/diameter (in); $\rho_{\text{n}}$ is the normal pore pressure in the region (ppg); $\rho_{\text{e}}$ is the equivalent circulating density, or ECD (ppg).

Next, the normal trend-line of modified d-exponent ($d_{{{\rm mod},{\text{n}}}}$) can be established using the normal trend-line. However, as all the members of Gachsaran formation are already abnormally over-pressured, we cannot use the available data to plot the normal trend-line and find $d_{{{\rm mod},{\text{n}}}}$ curve. Thus, alternatively, Zamora’s relation between $d_{{{\rm mod},{\text{n}}}}$ and $d_{{{\rm mod}, 0}}$ is used (Bourgoyne et al. 1986):

$$d_{{{\rm mod},{\text{n}}}} = d_{{{\rm mod},0}} {\text{e}}^{mD}$$

(8)

where $d_{{{\rm mod}, 0}}$ is the modified d-exponent; D is the ${\text{TVD}}\,({\text{ft}})$; and “$m$” is a constant to be found for each region. Using the data in Table 4, we evaluated the $d_{\rm mod}$ value for each data set. Next, in order to evaluate $d_{{{\rm mod},{\text{n}}}}$, we considered the value of 3.9 × 10⁻⁵ for “$m$” in the above equation (considering Bourgoyne et al. 1986) and changed $d_{{{\rm mod}, 0}}$ until matching and calibration with the real pore pressure value at 3076 m was obtained. The value of 1.26 was found as the best fitting of “$d_{{{\rm mod}, 0}}$” value. The modified d-exponent curve and the normal trend-line, $d_{{{\rm mod},{\text{n}}}}$, are plotted in Fig. 10. Based on Table 7, at the depth of 3076 m, kick flow occurred and shut-in pressures were obtained for real pore pressure evaluation.

Fig. 10

Normal trend-line of the modified d-exponent (following fitting of Zamora’s model for Gachsaran formation) versus the modified d-exponent values

Finally, using Zamora’s Eq. (1972), considering the normal pore pressure gradient (${\text{PG}}_{\text{n}}$) equal to 0.445 psi/ft, the pore pressure gradients (${\text{PG}}_{\text{p}}$) were found using the following equation:

$${\text{PG}}_{\text{p}} = \frac{{d_{{{\rm mod},{\text{n}}}} }}{{d_{\rm mod} }}{\text{PG}}_{\text{n}}$$

(9)

As Table 10 shows, the abnormality of this formation in pore pressure is excessively high (ranging from 0.96 to 1.07, which is comparable with 1 psi/ft, i.e., the overburden pressure gradient in Gulf of Mexico). This abnormality is mainly attributed to the conversion of gypsum to anhydrite in a closed basin during geologic time.

Table 10 Using Zamora’s d-exponent method to estimate pore pressure gradients of Gachsaran formation

In situ stresses

The overburden stress $\sigma_{v}$ is required to evaluate the horizontal stresses. Using the density log readings ($\rho_{b}$), we can estimate the overburden stress as follows:

$$\sigma_{v} = \mathop \int \limits_{0}^{z} \rho_{b} \left( z \right)g{\text{d}}z$$

(10)

The bulk density for the reservoir section has been evaluated by the well logs; however, for the sections where density log data are unavailable, the seismic velocity vp can be used for the estimation of overburden stress (Mavko et al. 1995). In this paper, for the non-logged interval (above 2700 m), the overburden pressure gradient of 1.13–1.15 psi/ft (which has been estimated for the Middle Eastern fields in the region) was used.

In a tectonically active basin, the practical way to estimate minimum ($\sigma_{h}$) and maximum ($\sigma_{H}$) horizontal stresses is by using the tectonic stress terms. The poro-elastic relations are commonly used to estimate horizontal stresses (Blanton and Olson 1997):

$$\sigma_{h} = \frac{\nu }{1 - \nu }\left( {\sigma_{V} - \alpha P_{\text{p}} } \right) + \alpha P_{\text{p}} + \sigma_{{{\text{tect}}h}}$$

(11)

$$\sigma_{H} = \frac{\nu }{1 - \nu }\left( {\sigma_{V} - \alpha P_{\text{p}} } \right) + \alpha P_{\text{p}} + \sigma_{{{\text{tect}}H}}$$

(12)

Here, $\nu$ is the Poisson’s ratio; $P_{\text{p}}$ is the pore pressure; $\sigma_{\text{tect}}$ is the tectonic stress; and $\alpha$ is Biot’s coefficient (with the value of 0.7 in this study, considering Zoback 2010).

Including the tectonic terms, it enables shifting the minimum horizontal stress to match the real measured data (typically leak-off-test data). This is called the strain-corrected method (Blanton and Olson 1997).

Replacing the equivalent equations for the tectonic stresses in the above equations (including the tectonic strain coefficients in the direction of minimum horizontal stress, $\varepsilon_{x}$, and in the direction of maximum horizontal stress, $\varepsilon_{y}$) gives (Blanton and Olson 1997):

$$\sigma_{h} = \frac{\nu }{1 - \nu }\left( {\sigma_{V} - \alpha P_{\text{p}} } \right) + \alpha P_{\text{p}} + \frac{{E\varepsilon_{x} }}{{1 - \nu^{2} }} + \frac{{\nu E\varepsilon_{y} }}{{1 - \nu^{2} }}$$

(13)

$$\sigma_{H} = \frac{\nu }{1 - \nu }\left( {\sigma_{V} - \alpha P_{\text{p}} } \right) + \alpha P_{\text{p}} + \frac{{E\varepsilon_{y} }}{{1 - \nu^{2} }} + \frac{{\nu E\varepsilon_{x} }}{{1 - \nu^{2} }}$$

(14)

Using the leak-off-test (LOT), mini-frac test results (or in this study, the mud pressure at the depths where mud loss occurred), we can calibrate the estimated $\sigma_{h}$ values (inferred from Zoback 2010), which is used to adjust $\varepsilon_{x}$ and $\varepsilon_{y}$. This method will inevitably calibrate the values of $\sigma_{H}$. It is noted that there is another method for correlating between $\sigma_{h}$ and $\sigma_{H}$ by Sinha et al. (2006) and Sayers et al. (2009) using the vertical and horizontal shear moduli in anisotropic formations.

Finally, following the calibration, the results of the evaluated rock moduli and strength, pore pressure, and in situ stresses can be shown in Fig. 11, as the MEM. The ranges of the in situ stresses of Gachsaran formation of Kupal field are listed in Table 11. Comparing the magnitudes of the stresses in Fig. 11, it is obvious that the stress regime in Gachsaran formation is dominantly strike slip fault (as $\sigma_{H} > \sigma_{v} > \sigma_{h}$₎, whereas at some depths the reverse fault regime (as $\sigma_{H} > \sigma_{h} > \sigma_{v}$) is observed.

Fig. 11

a The MEM geomechanical parameters for Gachsaran formation in well Z26 (2700–2950 m). b The MEM geomechanical parameters for well Z26 (2950–3200 m)

Table 11 The ranges of in situ stresses of Gachsaran formation (Kupal field) using the MEM results of well Z26

Drilling-induced stresses

Kirsch (1898) developed a set of equations for calculating the stresses acting in a (1) thick, (2) homogeneous, isotropic, (3) elastic plate (which can be assumed so when the formation is below half of its compressive strength, Goodman 1989), (4) containing a cylindrical hole, which is subject to effective minimum and maximum far field principal stresses. As well Z26 in this study is vertical, no stress transformations are required (Bradley 1979), and the three principal stresses are the effective radial stress ($\sigma_{rr}$), the effective tangential/circumferential stress ($\sigma_{\theta \theta }$) and the effective axial stress ($\sigma_{zz}$). We can estimate these induced stresses at the wellbore wall as follows (Jaeger and Cook 1979):

$$\sigma_{rr} = P_{\text{w}} - \left( {\alpha P_{\text{p}} } \right)$$

(15)

$$\sigma_{\theta \theta } = \left( {\sigma_{H} + \sigma_{h} } \right) - P_{\text{w}} - \left( {\alpha P_{\text{p}} } \right) - 2\left( {\sigma_{H} - \sigma_{h} } \right)\text{Cos} \left( {2\theta } \right)$$

(16)

$$\sigma_{zz} = \sigma_{v} - \left( {\alpha P_{\text{p}} } \right) - 2\upsilon \left( {\sigma_{H} - \sigma_{h} } \right)\text{Cos} \left( {2\theta } \right)$$

(17)

where $P_{\text{w}}$ is the wellbore pressure evaluated by the mud weight; $\alpha$ is the Biot’s coefficient which is considered equal to 0.7 in this work; $\theta$ is the angle from the direction of the $\sigma_{H}$.

The greatest possibility for tensile failure at the wellbore wall is in the direction of $\sigma_{H}$ ($\theta = 0$), where the Kirsch’s equations are simplified to:

$$\sigma_{rr} = P_{\text{w}} - \left( {\alpha P_{\text{p}} } \right)$$

(18)

$$\sigma_{\theta \theta } = 3\sigma_{h} - \sigma_{H} - P_{\text{w}} - \left( {\alpha P_{\text{p}} } \right)$$

(19)

$$\sigma_{zz} = \sigma_{v} - \left( {\alpha P_{\text{p}} } \right) - 2\upsilon \left( {\sigma_{H} - \sigma_{h} } \right)$$

(20)

The greatest possibility for shear failure (breakout) at the wellbore wall is at $\theta = 90$, where the Kirsch’s equations are simplified to:

$$\sigma_{rr} = P_{\text{w}} - \left( {\alpha P_{\text{p}} } \right)$$

(21)

$$\sigma_{\theta \theta } = 3\sigma_{H} - \sigma_{h} - P_{\text{w}} - \left( {\alpha P_{\text{p}} } \right)$$

(22)

$$\sigma_{zz} = \sigma_{v} - \left( {\alpha P_{\text{p}} } \right) + 2\upsilon (\sigma_{H} - \sigma_{h} )$$

(23)

The induced stresses calculated for well Z26 are plotted in Fig. 12, and the ranges of induced stresses are provided in Table 12. Table 13 provides the order of the three induced stresses in the direction of minimum horizontal stress (which will be used to determine the potential shear failure) and also the order of the induced stresses in the direction of maximum horizontal stress (which will be considered in determination of the potential tensile failure pressure).

Fig. 12

a Induced stresses at the wellbore wall (in the direction of $\sigma_{h}$) in Gachsaran formation of well Z26 (2700–2950 m). b Induced stresses at the wellbore wall in Gachsaran formation of well Z26 (2950–3200 m)

Table 12 The ranges of induced stresses at the wellbore wall in the direction of minimum horizontal stress ($\theta = 90$) and in the direction of maximum horizontal stress ($\theta = 0$) for Gachsaran formation of Kupal field, using the MEM results of well Z26

Table 13 The order of induced wellbore stresses in Gachsaran formation of well Z26 (in the directions of horizontal stresses)

Safe mud weight window

Construction of safe mud weight window is a really significant part of well planning and operations either in overbalanced drilling, or underbalanced drilling (Guo and Ghalambor 2002). This window is the final result of the 1-D mechanical earth model. It combines the equivalent gradients of pore pressure, shear failure (breakout), mud loss and tensile failure (fracture) in a single figure. The safe mud window consists of a lower and an upper boundaries. The lower boundary is constituted by either the pore pressure or shear failure profile (whichever greater). The upper boundary is constituted by the mud lost profile or the tensile failure profile. The following section explains the steps taken to construct the safe mud weight window.

Pore pressure gradient

Having estimated the pore pressure using the modified d-exponent method, its equivalent mud pressure gradient (${\text{PG}}_{\text{p}}$) is calculated as:

$${\text{PG}}_{\text{p}} \ge \frac{{P_{\text{p}} }}{{\left( {9.81 \times {\text{TVD}}} \right)}}$$

(24)

where $P_{\text{p}}$ is the pore pressure in KPa and ${\text{TVD}}$ is true vertical depth in m.

Shear failure pressure gradient

The drilling-induced shear failure is the compressive shear failure of the wellbore wall, which makes the intersecting conjugate shear planes resulting in pieces of rock spalling off the wellbore wall and its ovalization (Bell and Gough 1979; Meyer 2002). The shear failure criteria considered in this work are Mohr-Coulomb (MC), Mogi-Coulomb (MG), and Modified Lade (ML). The pressure causing the shear failure depends on the shear failure mode, which is determined based on the order of induced/near wellbore stresses (Aadnoy and Looyeh 2010). In the case of the studied well, based on the order of the induced stresses in the direction of the minimum horizontal stress ($\sigma_{\theta \theta } > \sigma_{zz} > \sigma_{rr}$), the mode of failure is known as wide-breakout (Bowes and Procter 1997) and the minimum mud pressure to avoid breakout, $P_{\text{w,s}}$ is calculated as (Mohr–Coulomb criterion):

$$P_{{{\text{w}},{\text{s}}}} \ge \frac{{3\sigma_{H} - \sigma_{h} - UCS + \alpha P_{\text{p}} \left( {q - 1} \right)}}{{\left( {q + 1} \right)}}$$

(25)

where $q = \left( {1 + \sin \varphi } \right)/\left( {1 - \sin \varphi } \right)$.

The shear failure pressure gradient is:

$${\text{PG}}_{{{\text{w}},{\text{s}}}} \ge \frac{{{\text{PG}}_{{{\text{w}},{\text{s}}}}}}{{\left( {9.81 \times {\text{TVD}}} \right)}}$$

(26)

Mud loss pressure gradient

The pressure required for mud loss ($P_{{{\text{w}},{\text{l}}}}$ in gr/cc) is equal to $\sigma_{h}$ and its corresponding equivalent mud weight (${\text{PG}}_{{{\text{w}},{\text{l}}}}$) is calculated as:

$${\text{PG}}_{{{\text{w}},{\text{l}}}} \ge \frac{{\sigma_{h} }}{{\left( {9.81 \times {\text{TVD}}} \right)}}$$

(27)

where, $\sigma_{h}$ is in KPa and TVD is in m.

Tensile failure pressure gradient

This failure is the upper limit of the safe mud weight window. The Fracture initiates when the effective minimum principal stress ($\sigma_{3}$) at the wellbore wall (in the direction of $\sigma_{H}$ at $\theta = 0$) reaches or exceeds formation rock tensile strength, $T_{0}$ (Fjaer et al. 2008). The tensile failure pressure is:

$$P_{{{\text{w}},{\text{l}}}}\ge 3\sigma_{H}-\sigma_{H}-\alpha{P}_{p}$$

(28)

The tensile failure pressure gradient is:

$${\text{PG}}_{{{\text{w}},{\text{t}}}} \ge \frac{{3\sigma_{H}-\sigma_{H}-\alpha{P}_{p}}}{{\left( {9.81 \times {\text{TVD}}} \right)}}$$

(29)

The order of stresses in well Z26 is either ($\sigma_{rr} > \sigma_{zz} > \sigma_{\theta \theta }$) or ($\sigma_{\theta \theta } > \sigma_{rr} > \sigma_{zz}$); therefore, either $\sigma_{\theta \theta }$ or $\sigma_{zz}$ is the $\sigma_{3}$. In either mode, the tensile failure pressure ($P_{{{\text{w}},{\text{t}}}}$) is evaluated (as shown in Table 14).

Table 14 Tensile failure pressures for different tensile failure modes (Simangunsong et al. 2006; Zhang et al. 2006)

Figure 13 shows the safe mud weight window of Gachsaran formation (in well Z26) using three failure criteria (Mohr–Coulomb, MC; Mogi–Coulomb, MG and Modified Lade, ML). Table 15 shows the ranges of the pore pressure, shear failure, mud loss and tensile failure boundaries. Some differences in the shear failure boundary can be seen between the results of the different failure criteria.

Fig. 13

a Safe mud weight window for Gachsaran formation (Kupal field, well Z26) using several shear failure criteria: Mohr–Coulomb, MC; Mogi–Coulomb, MG; and Modified Lade, ML (2700–2950 m). CMW_Kick shows the mud weights region where kicks can occur; CMW_Min shows the mud weights region where shear failure can occur; CMW_Max shows the mud weights region where tensile failure can occur; and CMW_Loss shows the mud weights region where mud loss can occur. b Safe mud weight window for Gachsaran formation for several failure criteria (2950–3200 m)

Table 15 The ranges of boundaries of pore pressure, shear failure (for different failure criteria), mud loss and tensile failure (from the results of the safe mud weight window)

Results and discussion

Using the geomechanical model and safe mud window developed for Gachsaran formation of well Z-26 (Fig. 13), the field operational challenges are first inferred by considering the safe mud window and field observations of mud loss and kick. Finally, to overcome such hole problems in the future, two innovative drilling systems are proposed as the remedial actions.

Field operational challenges

In the 2700–2950 m interval of Fig. 13, it is observed that the mud weights are lower than shear failure boundary causing shear failure to occur, which is confirmed by numerous tight holes in this depth interval. Next, we can see that the pore pressure suddenly increases around the depth of 3050 m and even exceeds the shear failure pressure. Due to the increased abnormality of the pore pressure, it exceeds the mud column pressure a few meters deeper (3076 m), i.e., kick flow is shown (which was confirmed by the crew’s experience). It is also observed that the mud weight reaches the mud loss boundary causing mud losses, which is confirmed by the loss observations at the rig site. Above all, the safe mud weight window is narrow and maintaining drilling mud weights in this window is not practically possible. Therefore, some methods to mitigate this problem should be found.

According to the results, Table 16 displays the polar plots of the drilling mud weight (at four arbitrary depths) to prevent onset shear and tensile failures in Gachsaran formation. The plots show the optimum mud weight distribution for each inclination and azimuth. It is inferred from the shear failure polar plot that we require a lower drilling mud weight for deviated holes in the direction of maximum horizontal stress (particularly with the inclination angles greater than 40 degrees). Considering the tensile failure polar plot, the mud weight span is wider in the deviated holes to the maximum horizontal direction. Therefore, the mud weight can be safely selected high enough to overcome excessively high pore pressures of Gachsaran formation. It also signified that the shear failure would not be any issue for such deviated holes. However, practically, Gachsaran formation is preferred to be drilled vertically. This is because drilling of Gachsaran formation would add additional operational challenges such as solids settling. Next, the preference is to keep the 12 ¼” hole (across Gachsaran formation) vertical and then directionally open a window in the next hole section (8 ½”) which is across the Asmari formation reservoir.

Table 16 Polar plots of the mud weights for shear and tensile failures at four selected depths (using Mogi–Coulomb failure criterion)

Suggested field remedial actions

In order to mitigate the wellbore stability issues and associated hole problems, it was recommended to drill the formation as quickly as possible. Therefore, CWD and CCS systems are designed as follows.

Casing while drilling

The tripping time of the casing string to the casing point is considerable in conventional drilling. This is detrimental to wellbore stability in a geomechanically unstable formation such as Gachsaran formation. The tripping time can contribute to a large number of drilling issues such as stuck casing or barite sag (i.e., settling of the mud weighting materials and solids). Therefore, in addition to the drilling optimization, casing while drilling (CWD) system is proposed in order to eliminate this casing running time.

The CWD is a system which allows the casing to be installed in a well during the hole-making process. To make this happen, it uses the casing string as the drill string with the bit. This system contributes to elimination of the casing trip time. Thus, following drilling, the hole can be immediately cemented and properly sealed. In addition, it provides the plastering effect, i.e., strengthening of the wellbore and the mud cake, which contributes to mud loss mitigation (Woods 2003).

The high cost of CWD limits its use only in special cases when safe drilling using conventional systems is severely challenged due to the numerous drilling hazards. An example is drilling in formations with severe wellbore stability issues such as unexpected lost circulation and mud losses particularly with narrow and variable safe mud window, as is the case in this study. Another example can be drilling in formations with high-swelling shale, mainly caused by the time-dependent chemical wellbore instability. The unexpected drilling problems can largely be mitigated due to the advantages of CWD, including elimination of the trip time and for its plastering effect. Therefore, safe and optimized drilling can be viable in such formations.

CWD design

CWD is offered in two possible designs: (a) drillable bit and (b) retrievable dual-body bit, whereby the dual-body bit can be retrieved via wireline. In the 12 ¼-in. hole section of Kupal field, due to the frequent hole problems, an unprecedented pulling out of the hole may become essential before reaching the casing point (due to equipment failure, e.g., the bit can become dull especially during drilling anhydrite, or it may become balled-up by the marl). Therefore, the retrievable CWD option is proposed in our case as it provides the capability to replace failed equipment and it is a cheaper option (Woods 2003). In the retrievable CWD, the drill string consists of (see Figs. 14 and 15):

• 8 ½-in. pilot bit The pilot bit is the lower part of the dual-body bit system (at the bottom of the string).

• Under-reamer The under-reamer is the upper part of the dual-body bit system. It is expanded to its full size (12 ¼-in.) as soon as mud circulation is started bottomhole (to enable enlarging the under-reamer size to 12 ¼-in.). At the end of drilling Gachsaran formation, prior to retrieval (with wireline), upon stop of the mud circulation, the under-reamer reverts back to its initial size (8 ½-in.).

• DLA (drill lock assembly) It works as the connector of the casing string to the under-reamer.

• 9 5/8-in. casing string: The 9 5/8-in. casing string acts as the drill pipe. As the casing joints needs to stand the torsion and tension during drilling, their thread type must be VAM-TOP (which have extra tensile and torsion strength).

• Stabilizers Stabilizers are used between the pilot bit and under-reamer and in the casing string for deviation control of a vertical well, especially when there are intervals containing soft layers (as is the case with some intervals of Gachsaran formation in Kupal field).

Fig. 14

a Conventional drill string, b non-retrievable CWD BHA and c retrievable CWD BHA for drilling the 12 ¼-in. hole

Fig. 15

Schematic of the drill string in casing while drilling (CWD) including the casing string (9 5/8-in.) and the BHA for deviation control in a vertical 12 ¼-in. hole of Kupal field (modified after Woods 2003; Mohammed et al. 2012)

Required equipment for retrievable CWD

There are several steps taken to make retrievable-CWD option possible. Therefore, first, the rig must appropriate specifications as stated below. Next, a wireline truck is required to retrieve the dual-body bit.

The drilling rig should have enough hook load capacity to be able to stand the weight of the casing string (used as the drill string). This should be met not only in buoyed condition (when the string is suspended), but also in dry condition (in case the string gets stuck and we need to apply overpull/tension to free it). Therefore, the maximum hook load must exceed the dry weight of the string. In Kupal field, as the maximum TVD in the 12 ¼-in. hole does not exceed 4000 m, with the assumed 53.5 Ib/ft weight (ID of 8.535-in.) of the casing string, the minimum load would be nearly 700 K-Ib. Therefore, the currently active drilling rig in Kupal field (which has the hook load capacity of 1000 K-Ib and power of 2000 hp drawworks) has adequate hook load capacity for the CWD operation. Second, as the casing string is used in lieu of the drill string in CWD (with greater OD than the conventional drill string), the rotary table or top drive should provide or stand greater torques.

As the retrievable CWD design is found suitable in this field, a wireline truck (with enough weight and tension capacity) must be available at the rigsite to retrieve the dual-body bit prior to running and cementing the casing.

Considerations for CWD Use

The main limitation of using CWD to drill Gachsaran formation is the use of super-heavy mud with solid percentage reaching 50% (to overcome the pore pressure gradients up to 22.32 ppg) which may cause possible stuck of the casing string during drilling. A possible measure to prevent this issue could be using an Expandable Tubular Technology. However, it is not possible to convert the 9 5/8-in. casing to a liner in Kupal field as the 9 5/8-in. tubular must ultimately be extended to the surface due to the active tectonic forces of the field. It is noted that in 2003, this technology was already applied in one of the wells of this field (Yousafi et al. 2008) for running the 7-in. liner in the 8 ½-in. hole (not the 9 5/8-in. casing). Therefore, the necessary measures to prevent the mentioned problem is the use of high-quality mud materials (including weight materials such as barite and hematite and viscosifiers or polymers) to maintain the mud rheology and also using continuous circulation system (CCS) to prevent solid settling and kick flows during making connections.

Continuous circulation system

Continuous circulation system (CCS) is a model of managed pressure drilling (MPD) system to maintain the mud circulation during making connections (Ayling et al. 2002) with the dynamic mud pressure exceeding the formation pressure. It signifies that the well conditions can be maintained dynamic even during connections.

Application of CCS is essential in this field for several reasons. First, it can contribute to prevention of settling of the mud solids during connections, which is greatly important for CWD. Next, due to excessively high mud losses occurring in drilling Gachsaran formation, using the CCS system, ECD drilling should be commonly practiced (to mitigate extremely high mud losses encountered). Using conventional drilling system, circulation of drilling mud to the borehole has to stop to add a stand or joint of drill pipe to the drill string. Therefore, with the static mud column pressure lower than the formation pressure, kick flows occur during connections. Although the crew expedites in making the connections in such cases, the well records indicate that kick flows with 20–30 bbls pit gains have occurred. Therefore, it is essential to address this risky problem using CCS. To do this, as Fig. 16 shows, CCS simply requires a coupler (including a triple ram body which is installed at the rig floor) to allow the mud flow inlet and circulation from the mud pumps to the well as the drill pipe joint is added to (or removed) the drill string.

Fig. 16

a Mud flow circuit using a coupler installed at the rig floor and b cross-section of the coupler (Ayling et al. 2002)

In some cases, mud pumps working with extremely high mud weights of up to 22.3 ppg may fail specially with in pumps with low-quality internal parts (such as seat and valve). It is noted that mud pump failure should be strictly prevented; otherwise, CWD or ECD drilling with CCS would fail and put the well at high risks.

Conclusions

In Kupal oilfield, drilling the cap rock of the oilfields faces numerous geomechanical challenges and lost times. To address the geomechanical challenges, this work has developed an MEM’s workflow for a complex lithology formation by adjusting the equations/correlations to match with the core points and observations. The following items summarize the findings:

• 1. This is a pioneer wellbore stability work for the non-reservoir Gachsaran formation (with anhydrite, gypsum, marl and thin limestone layers). Therefore, such workflow is unique as it considers specific geomechanical and operational issues for such a complex lithology which are not available in other published MEM workflows. Having selected available correlations which are mostly for sandstones and carbonates, the constructed MEM was successfully calibrated using the available core data and field observations.

• 2. The results of the wellbore stability analysis in Kupal field indicated an excessively narrow safe mud weight window and over-pressured formation being the main cause of severe geomechanical drilling problems.

• 3. To mitigate the borehole problems, using the casing while drilling (CWD) and continuous circulation system (CCS) was proposed as a remedial action in order to prevent hole problems eliminating the geomechanical aspect of the lost times (30% of the total lost times).

• 4. The CWD reduces the casing running time after drilling the hole section. The CCS enables mud circulation during pipe connections to prevent settling of super-heavy mud solids in CWD and prevent kick flows.