Tubular fracturing: Pinpointing the cause

Improper heat treatment can trigger temper embrittlement in oilfield tubulars  

By Srinivasa Koneti, Samit Gokhale, Thomas Wadsworth, T.H. Hill Associates

Figure 1 (left): Intergranular cracking, characterized by triple points, rock-candy or a faceted appearance, occurs at and along the grain boundaries of metal. Figure 2 (right): Transgranular cracking occurs through or across the crystals or metal grains and is characterized by cleavage steps, river patterns, feather markings and tongues. This shows an example of a transgranular fracture on the fracture surface of low-carbon steels.

Brittle fracture of oilfield tubular components can occur due to the material having low fracture toughness – such material often presents low Charpy V-notch (CVN) impact energy values – or from exposure of the material under load to certain corrosive operating environments. A brittle fracture can show characteristics of transgranular or intergranular cracking when analyzed through a scanning electron microscope (SEM).

Intergranular cracking is the cracking or fracture that occurs at and along the grain boundaries of a metal. It is characterized by triple points, rock-candy or a faceted appearance when the fracture is analyzed through SEM. Figure 1 shows a typical example of an intergranular fracture on the fracture surface of low carbon steels.

Transgranular cracking is the cracking or fracture that occurs through or across the crystals or metal grains. It is characterized by cleavage steps, river patterns, feather markings and tongues when the fracture is analyzed through a SEM. Figure 2 shows a typical example of a transgranular fracture on the fracture surface of low-carbon steels.

Intergranular cracking is often the mode of fracture that occurs when tubular components are exposed to environmental conditions that contain aqueous H2S. Such failures promulgate the notion that detection of intergranular cracking morphology on fracture surfaces is confirmation of failure through sulfide stress cracking (SSC) or hydrogen embrittlement, even when no evidence exists for exposure to H2S or a source of nascent hydrogen.

Study of intergranular cracking related failures has shown that such failures can occur not only when the component is exposed to nascent hydrogen but can also be caused by temper embrittlement of the material resulting from improper heat treatment.

Temper embrittlement typically occurs when carbon or low-alloy steels are held at or slowly cooled through the temperature range of 375°C (705°F) to 575°C (1,065°F) during the tempering process. If the steels are tempered or slowly cooled at these temperatures, the material shows brittle characteristics (loss of impact toughness).

Steels that have experienced temper embrittlement can be restored to their original or expected toughness by heating (tempering) to 600°C (1,100°F) or above, followed by rapid cooling to below approximately 300°C (570°F).The fracture surface of a material with low CVN impact energy values (brittle material) would normally show transgranular signatures when analyzed under a SEM, whereas a ductile material affected by environmental attack, such as hydrogen embrittlement, shows intergranular separation at grain boundaries.

However, brittle fracture of a material that undergoes temper embrittlement also shows signs of intergranular cracking. Examination of the fracture surface of a component that has undergone temper embrittlement can present intergranular or mixed mode of intergranular and transgranular fracture morphology.

n Case 1 from South Texas, the pin connection of a new saver sub failed. The drilling engineer recognized the failure as a brittle failure.

Case studies of Temper Embrittlement Failure

Case 1 – South Texas, onshore US

In March 2009, while making up the pin connection (6 5/8-in. reg) of a saver sub, the pin connection on the sub failed. The operator reported that the saver sub was procured new and was in service for three days before the failure occurred.

Based on the appearance of the fracture surface, the proximate cause of the failure was readily recognized by the drilling engineer as a brittle failure. To confirm the failure mechanism, the failed sub was sent for failure investigation.

In Case 2 from Oklahoma, the pin connection twisted off while making up the pin connection of a saver sub.

Case 2 – East Oklahoma, onshore US

In May 2010, while making up the pin connection (6 5/8-in. reg) of a saver

sub, the pin connection twisted off. Based on the fracture surface morphology, the failure mechanism was identified as a brittle fracture with rapid crack propagation. To confirm the cause of the failure, the failed sub was sent for further investigation.

Case 3 – Northeast Trinidad, offshore

In April 2010, the operator was in the final stage of drilling a horizontal well that entailed the pullback of the 36-in. production pipeline. While pulling back drill pipe joint No. 90, a 7 5/8-in. reg pin connection on a sub that fastened the 42-in. hole-opener to the 500-ton swivel failed downhole.

In Case 3 from Northeast Trinidad (lower left), a pin connection on a sub that fastened the hole-opener to the swivel failed downhole.

The attached fractured sub was pulled out, and the mating portion of the sub was not recovered by fishing and subsequently resulted in losing the well. Based on the appearance of the fracture surface, the proximate cause of the failure was identified as fatigue, followed by a brittle fracture. To confirm the failure mechanism, the failed sub was sent for investigation.

Metallurgical Analysis of Failed Subs

Metallurgical analysis of the fractured pin connections on the subs was performed to identify the cause of the failure and the factors that contributed to the failure. To differentiate the fractured pin connections of the subs, the subs will be referred to as:

C1: Sub from Case 1

C2: Sub from Case 2

C3: Sub from Case 3

The fractures of all the pin connections were located in the last engaged threads of the pin connections. The last engaged threads of a connection experiences higher stresses and stress concentrations compared with the rest of the connection, making these threads susceptible to cracking. The as-received condition of the failed subs is presented in Figure 3. The C3 sub was received after initial metallurgical testing was performed by another lab. A portion of the sample that was used for previous testing was missing.

Figure 4: The fracture on the Case 1 sub showed a grainy texture and “chevron marks” that point toward the initiation site, which is typical morphology for brittle cracking.

The overall appearance of the fracture surfaces on the subs was flat and oriented perpendicular to the sub axis. The fracture on C1 and C2 exhibited a grainy texture and “chevron marks” that point toward the initiation site. This is typical morphology for brittle cracking (Figure 4).

The fracture on C3 exhibited a small fatigue region (approximately 5%) that was followed by brittle fracture. The fracture surface also had the grainy appearance (Figure 5). All three fracture surfaces present a minuscule shear lip, which is also typical of a brittle fracture. Note that the missing material was used for testing during previous investigation.

A thread profile analysis of the failed pin connections of C1 and C2 was performed to check for stretched threads, but no signs of such were observed. This gave further evidence that the pin connections on C1 and C2 failed in a brittle manner and not through ductile torsional/tensile overload.

Figure 5: The fracture on C3 exhibited a small fatigue region that was followed by brittle fracture. The fracture surface had a grainy appearance and presented a minuscule shear lip, which is also typical of a brittle fracture.

Material Testing

Material testing of the failed subs was performed to verify compliance with API Specification 7-1 and Standard DS-1 and to determine if improper material properties contributed to the failures. Tensile tests, chemical analysis and CVN tests were performed on the failed pin connection material of the subs.

Tensile strength and yield strength met the minimum requirements specified in API Specification 7-1 and Standard DS-1 for sub material. However, the CVN impact energy values did not meet the minimum requirements specified.

Typically, for the type of chemistry used, CVN values correlate well with fracture toughness. Low fracture toughness makes the material notch sensitive and typically results in predominately brittle fracture.

This indicates that the heat treatment processes were not performed properly to achieve the correct mechanical properties on the subs.

With temper embrittlement, generally, there is no detectable drop in expected yield strength, tensile strength and percent elongation of the material. A drop in the CVN values is often experienced.

In cases of extreme embrittlement, there may be a drop in the percent reduction of area. The material test results obtained on the failed subs are similar to test results commonly observed on components that have experienced temper embrittlement.

Because there is no major change in the tensile properties of the material, hardness testing cannot be used to detect temper embrittlement. Performing CVN tests followed by examination of the fracture surface of the CVN samples under a SEM are necessary to ascertain failure through temper embrittlement.

Figure 6: No signs of stretched threads were observed after a thread profile analysis of the failed pin connections of C1 and C2 was performed.

Scanning Electron Microscopy Analysis

The fracture surfaces of the failed pin connections on C1, C2 and C3 were electrolytically cleaned to remove oxides, which mask the fracture signatures. The cleaned fracture surfaces were then observed through a SEM. The fracture examination on C1 and C2 revealed features typical of transgranular fracture. The examination also revealed signatures of intergranular cracking (Figure 7).

The presence of both intergranular and transgranular features indicates a mixed mode fracture morphology. As discussed, the presence of intergranular cracking is often considered proof of failure induced through environmentally assisted cracking, such as SSC or hydrogen embrittlement. However, a saver sub is unlikely to come into contact with downhole corrosive environment.

Moreover, review of the operating conditions and environment provided no evidence of a source of nascent hydrogen. In this instance, presence of a mixed mode of intergranular and transgranular morphology on the fracture surface, combined with the low CVN values, indicates that the failure is more likely associated with temper embrittlement of the material resulting from improper heat treatment of the component.

Figure 7: The fracture examination using a SEM on C1 and C2 revealed features typical of transgranular fracture (left and middle) and signatures of intergranular cracking (left and right). The presence of both intergranular and transgranular features indicates a mixed-mode fracture morphology.

SEM analysis of C3 was also performed. However, no signatures were observed as the fracture surface was too corroded for examination.

To confirm if the subs underwent temper embrittlement, the fracture surface of the CVN impact test samples were analyzed under a SEM. Typically, examination of the fracture surface of CVN samples from a ductile material can present portions of ductile dimples and transgranular “cleavage” cracking.

This morphology is also expected on CVN samples that are machined from an inherently brittle material or a ductile material that has fractured through SSC or hydrogen embrittlement. The fracture examination of the CVN samples from the failed subs revealed mixed mode of intergranular, transgranular and some ductile dimple features (Figure 8).

This mixed mode of intergranular and transgranular cracking indicates that the subs likely underwent temper embrittlement resulting from improper heat treatment. Hence, presence of intergranular cracking does not confirm environmental cracking. Instead, CVN testing should be performed to check the fracture toughness of the material.

Additionally, the fracture surface of the CVN samples should be analyzed under a SEM to verify the fracture mode. Temper embrittlement of material is a strong possibility if the material presents low CVN values along with presence of intergranular or mixed mode of intergranular and transgranular cracking signatures on the fracture surface of the CVN sample.

Figure 8: The fracture examination of the CVN samples from the failed subs C1 (left), C2 (middle) and C3 (right) revealed mixed mode of intergranular, transgranular and some ductile dimple features. This indicates the subs likely underwent temper embrittlement resulting from improper heat treatment.

Guidelines on Alloying Elements

Temper embrittlement is often associated with the concentration of certain trace alloying elements, such as arsenic, antimony, tin and especially phosphorus. These minor impurities segregate along the austenitic grain boundaries during the tempering process and cause cracking along the grain boundaries.

Molybdenum, tungsten and zirconium greatly reduce embrittlement, and nickel, titanium and vanadium slightly reduce the temper embrittlement effects.

API Specification 7-1 and Standard DS-1 do not have any requirements for chemistry on subs. However, API Specification 5DP and Standard DS-1 have chemistry requirements for drill pipe tube and tool joints for phosphorus (0.020% max) and sulfur (0.015% max).

The sulfur content obtained on all the three failures was above the maximum allowed for drill pipe tube. The phosphorus obtained on C3 was above the maximum requirement, while the content for C1 and C2 was near the maximum allowed. This provides basis for strict control on these elements to minimize the possibility of temper embrittlement problems.

If not already required in the governing standard, supplementary requirements on content of phosphorus (0.02% max) and sulfur (0.015% max) should be specified when tubular components are ordered.

To check if the failure mechanism of the failed sub C1 was temper embrittlement, the sub material was re-heat treated. The re-heat treatment was also performed to confirm whether the sub was improperly heat-treated at the mill.

Sections of the failed pin connection were re-heat treated with the following conditions:

Condition 1: Temper at 657°C (1,215°F) for 45 min, and cool.

Condition 2: Austenitize at 872°C  (1,602°F) for 55 min; water quench; temper at 1,215°F (657°C) for 45 min; and cool.

The heat treatment procedures listed in the material test report (MTR) were used for re-heat treatment of the sub material. These conditions were chosen because the tempering temperature listed in the MTR does not fall in the temper embrittlement range.

If the sub was heat-treated at the mill with the conditions indicated in the MTR, temper embrittlement likely would not have occurred. After re-heat treating the sections from the failed pin connection with the conditions listed above, CVN impact tests were performed.

Significant improvement in the CVN values was observed from the re-heat treated material under both conditions. The reason for higher CVN values obtained through Condition 1 compared with Condition 2 is that the Condition 1 material underwent a double tempering process at the mill, and again during the re-heat treatment process.

The minimum and average impact energy of the re-heat treated sections was greater than the minimum required value specified in API Specification 7-1 and Standard DS-1. This confirmed that the failed sub was not heat-treated to the parameters listed in the MTRs.

Figure 9: To check if temper embrittlement still existed after re-heat treatment, the fracture surfaces of the CVN samples were analyzed under a SEM. Microvoid coalescence, seen as ductile dimples, was observed, which is indicative of ductile overload of the material.

To check if temper embrittlement still existed, the fracture surfaces of the re-heat treated CVN samples were analyzed under a SEM. Figure 9 present the fracture surfaces of the re-heat treated CVN samples as seen under SEM.

Microvoid coalescence (seen as ductile dimples) was observed on the fracture surface of the CVN sample, which is indicative of ductile overload of the material. No features of intergranular or mixed mode of intergranular and transgranular cracking were observed.

Hence, temper embrittlement was eliminated by performing the re-heat treatment on the failed sub material. Temper embrittlement was eliminated with only tempering the sub material (Condition 1). This confirms that temper embrittlement can be reversed with a tempering process performed at the appropriate temperature.

1. Detection of intergranular cracking morphology on fracture surfaces of a failed component is often considered to be confirmation of failure through SSC or hydrogen embrittlement, even when no evidence exists for exposure to H2S or a source of nascent hydrogen.

Study of intergranular cracking related failures has shown that intergranular fractures can occur not only when the component is exposed to corrosive environment, such as aqueous H2S, but can also be caused by temper embrittlement of the material resulting from improper heat treatment.

2. Temper embrittlement typically occurs when carbon or low-alloy steels are held at or slowly cooled through the temperature range of 375°C (705°F) to 575°C (1,065°F) during the tempering process.

3. Temper embrittlement is often associated with the concentration of certain trace alloying elements, such as arsenic, antimony, tin and especially phosphorus. These minor impurities segregate along the austenitic grain boundaries during the tempering process and cause cracking along the grain boundaries. If not already required in the governing standard, supplementary requirements on content of phosphorus (0.02% max) and sulfur (0.015% max) should be specified when tubular components are ordered.

4. The fracture surface of a failed component that has experienced temper embrittlement can present intergranular or mixed mode of intergranular and transgranular fracture morphology when analyzed under a SEM.

5. Generally, temper embrittlement of a material, does not lead to a detectable drop in expected yield strength, tensile strength and percent elongation of the material. However, a drop in the CVN impact energy values is often experienced, and in cases of extreme embrittlement, there may be a drop in the percent reduction of area.

6. Since there is no major change in the tensile properties of the material, hardness testing cannot be used to detect temper embrittlement of a material. Performing CVN tests followed by examination of the fracture surface of the CVN sample under a SEM are necessary to ascertain failure through temper embrittlement.

7. If the fracture surface on the failed components presents signatures of intergranular fracture, then it should not be presumed that the failure is associated with environmental cracking like SSC. Instead, CVN testing should be performed to check if the material has low impact energy values.

Once tested, SEM analysis of the fracture surface of the CVN sample must be performed to check for intergranular or mixed mode of intergranular and transgranular cracking. Presence of intergranular or a mixed mode of intergranular and transgranular morphology on the fracture surface of the CVN samples, combined with low CVN values, indicates a failure more likely associated with temper embrittlement of the material.

8. If the component being tested, such as tubing, does not have sufficient thickness to machine minimum required size CVN samples according to the governing API specification (minimum size accepted by API is 10 mm x 5 mm), then CVN samples should be machined to 10 mm x 2.5 mm (¼-in.) size to perform CVN testing.

Although the values obtained through testing cannot be compared against API specification requirements, the fracture surface of the CVN samples can still be analyzed under a SEM to check for intergranular or mixed mode of intergranular and transgranular cracking.

9. Temper embrittlement is a reversible process. Carbon and low-alloy steels that have experienced temper embrittlement can be restored to their original (or expected) toughness by heating (tempering) to 600°C (1,100°F) or above, followed by rapid cooling to below approximately 300°C (570°F). Material susceptibility to temper embrittlement can also be reduced by strict control and reduction of embrittling impurities, such as phosphorus.

This article is based on SPE/IADC 139762, “Intergranular Cracking of Oil Field Tubular Components Resulting from the Tempering Process,” SPE/IADC Drilling Conference & Exhibition, Amsterdam, The Netherlands, 1-3 March 2011.

References
1. API Specification 7-1, Specification for Rotary drill Stem Elements, first edition, American Petroleum Institute (March 2006),
Section 7.5, Page 26.
2. API Specification 5DP, Specification for Drillpipe, first edition, American Petroleum Institute (August 2009), Table C.4, Page 86.
3. Hill, T.H.: Drill String Design and Failure Prevention, T H Hill Associates, Inc. (September 2002).
4. Metals handbook, Volume 4, Heat treating, ninth edition, American society for metals (November 1981), Page 84.
5. Metals handbook, Volume 11, Heat treating, ninth edition, American society for metals (November 1981), Page 6, 11, 99.
6. Metals handbook, Volume 12, Heat treating, ninth edition, American society for metals (November 1981), Page 13, 174.
7. Standard DS-1® Volume 1: Drilling Tubular Product Specification, third edition, fourth printing, T H Hill Associates, Inc. (January
2004), Table 3.2.1, Page 20 and Table 3.1, Page 13.
8. William T. Becker. ASM International Course 0335, Principles of Failure Analysis, Lesson 3: Ductile and Brittle Fracture, Page
61, 62, 63.

View the original article here

Study simulates kick responses during MPD

Experimental wells confirm alternative well control procedures to be effective in range of well conditions

By J.E. Chirinos, J.R. Smith, D.A. Bourgoyne, Louisiana State University

Figure 1 shows the layout for experimental well LSU #2, where a gas kick was emulated by injecting gas in the 1 ¼-in. tubing until a desired pit gain was obtained. The well was used to evaluate two non-circulating kick responses and the pump startup procedure for kick circulation after a non-circulating response is applied. A computer simulation approach was also used to evaluate and confirm the applicability of these procedures.

According to IADC, managed pressure drilling (MPD) is defined as “an adaptive drilling process used to precisely control the annular pressure profile throughout the wellbore.” The technology uses different approaches to control and influence wellbore pressure. It is able to actively manipulate the wellbore pressure profile by controlling backpressure, drilling fluid proprieties and circulating friction; hence, a combination of tools is used to achieve MPD objectives to reduce nonproductive time (NPT) and mitigate drilling hazards.

MPD has been shown to be successful in wells where kicks, lost returns, ballooning, wellbore instability and/or differential sticking cause excessive NPT or inability to reach objectives using conventional drilling methods.

Although the main application of MPD has primarily been drilling in a narrow margin between pore pressure and fracture pressure, it’s increasingly realized that MPD can be applied anywhere where more precise control of wellbore pressure is an advantage. As a result, the industry has addressed a significant number of challenges by using MPD. Applications include: narrow drilling margin between pore pressure and fracture pressure, depleted formations, tight-gas sands, shallow gas hazards, wellbore stability problems, fractured carbonates, HPHT wells, H2S wells, slim-hole coiled-tubing drilling and casing drilling.

This article focuses on one variation of MPD – constant bottomhole pressure (CBHP) – which uses a combination of equipment to manipulate annular frictional pressure losses and casing pressure to keep wellbore pressure at a selected depth relatively constant. Although the CBHP method of MPD has better control of wellbore pressure while drilling, well control events can still occur because of uncertainty related to pore pressure and fracture pressure, human error or equipment failure.

Table 1 (left) provides an example pump startup schedule for routine operations. It steadily increases the pump rate to keep BHP approximately constant. CP* = desired casing pressure for routine operations. Table 2 (right): The post-kick pump startup schedule for MPD kick circulation is based on a stabilized shut-in casing pressure (SICP) and the routine pump startup ?P.

Several alternative well control procedures during the application of the CBHP method have been studied by industry-supported research to establish a basis for determining appropriate procedures. This research has defined two non-circulating responses as preferable among multiple alternative non-circulating responses. These two responses are described as a simple shut-in and as a MPD pump shutdown with a “choked flow check.” Both require a pump startup schedule to begin circulating out a kick. The goal of this article is to explain and document these procedures and to demonstrate by applying them in computer simulations and to gas kicks taken in a full-scale experimental well.

The CBHP method is the most common variation of MPD. During its application, annular pressure in the well is held constant or near constant at a selected depth. CBHP actively controls the surface pressure using a drilling choke to compensate for changes in frictional pressure losses (?PAF) during routine operations, such as making a connection.

An important characteristic of this method of MPD is the minimization of wellbore pressure variation to keep wellbore pressure within the drilling margin. Consequently, it allows drilling within a narrower window, or margin, between fracture and pore pressures than conventional drilling methods.

CBHP uses a collection of tools to control wellbore pressure during drilling operation. The minimum equipment required to apply CBHP are a rotating control device (RCD), a drilling choke manifold and, typically, a non-return valve. The RCD keeps annular space closed and diverts flow to the drilling choke; it is equipped with a rotating packer that rotates and holds pressure in the well during drilling operations.

The drilling choke manifold helps manipulate and control surface pressure while drilling; it can be controlled manually, semi-automatically or automatically. A non-return valve or float valve is installed in the bottomhole assembly; it allows only downward flow of drilling fluids, which is necessary if the well will be statically underbalanced.

Other optional tools can complement CBHP operation to improve wellbore pressure management, such as coriolis meters (flow meter), continuous circulating systems, downhole deployment valves, backpressure pumps, surface multiphase separators, pressure-while-drilling tools (PWD) and hydraulic flow modeling.

Pump Shutdown, Pump Startup Schedule

Many authors have discussed the method used to transition from dynamic to static state during CBHP. Medley, et al, and Rehm, et al, described a method to achieve CBHP objectives when rig pumps are shut down. It relies on a hydraulic model to estimate annular friction losses and equivalent circulating density (ECD) at different pump rates. Then, the ECD is manipulated by adjusting the casing pressure to keep wellbore pressure constant when the pump rate is reduced.

Wellbore pressure and ECD estimates are made with hydraulic models and can be validated with PWD tools. For example, to make a connection, the choke opening is reduced to increase casing pressure to the desired pressure, then the pump rate is reduced. Thus, surface pressure increases as the frictional pressure loss decreases by an equal amount. This process continues stepwise until the annulus surface pressure is at the maximum calculated value and the pumps are stopped.

The final annulus surface pressure should be equal to the frictional annulus pressure losses in the well, plus any surface backpressure held during normal operations.

Table 3 defines the pump startup and pump shutdown schedule for normal conditions, and Figure 2 illustrates application of the schedule in well LSU #2. BHP was kept essentially constant during pump startup and shutdown.

Initial Responses during Well Control Operation

A few studies have been done in the area of initial responses to well control events for the CBHP method of MPD. Das (2007) documented the first research from the university-industry consortium related to initial responses to a kick taken during the CBHP method. He compared three initial responses by using computer simulation: shutting in the well conventionally, increasing choke pressure while keeping the same pump rate, and increasing pump rate while keeping choke pressure constant. The most important conclusions from this research were: a) no single response was identified as the best, b) circulating responses may stop the influx faster than non-circulating responses, and c) the increased choke pressure response leads to a lower shoe pressure than shut-in, thus it reduces the risk of lost returns at the shoe.

In 2009, Guner studied the most appropriate initial response and kick circulation method for an unexpected reduction of bottomhole pressure created by a surface equipment failure or unintended ECD reduction. The conclusions explained that shut-in was the initial response that is applicable for all kick scenarios; however, increasing choke pressure would generally be the most effective response when it was practical. For both responses, Guner recommended kick circulation at the normal drilling circulation rate.

In 2009, Davoudi documented a comprehensive investigation of alternative initial responses to gas kicks taken during drilling operations with the CBHP method. He studied nine responses, five of them non-circulating and four circulating responses. They were compared based on the ability to stop formation flow, minimizing the risk of lost returns and additional kick influx, and the reduction of pressure imposed at surface and the casing shoe.

Davoudi performed more than 150 simulations and found that no single best initial response to all kicks could be identified. However, three initial responses were demonstrated to have a broad application to different kick scenarios: a rapid increase of casing pressure until flow out equals flow in, a simple shut-in, and an adaptation of the MPD pump shutdown schedule that allowed confirmation of low rate kicks using a choked flow check. In addition, he concluded that the best initial response depended on well conditions and the equipment being used.

Based on the consortium research, Davoudi, et al (2010) presented a proposed approach for selecting initial responses during well control events for the CBHP method. They explained that one criteria in selecting the initial kick reaction must be the equipment available onsite, specifically whether flow-out metering was being used.

In addition, according to this approach, the selection of the initial response should consider the certainty of the well control event. Each initial response has key factors that need be considered to ensure applicability and efficiency during the well control operations.

Figure 3: Actual shut-in pressures in LSU #2 were recorded, with the stabilized SICP interpreted at 510 psi. Table 4: A post-kick pump startup schedule for the experimental well was made based on the schedule for normal conditions.

Research Method

This article focuses on evaluating proposed procedures for the two non-circulating kick responses described by Davoudi in 2010 and on the pump startup procedure for kick circulation after a non-circulating response is applied. Two approaches were used to evaluate and confirm the applicability of these procedures: computer simulations and full-scale experiments.

Non-Circulating Responses

Simple Shut in

This response is widely known and accepted in conventional operations. However, according to work by Gunner in 2009, the simple shut-in procedure can be applied in MPD operation where accurate flow metering is not available or where equipment failure endangers the operation. This procedure can be summarized as:

1. Stop drilling, i.e., pick up off bottom and stop rotating.

2. Shut down the pumps as quickly as practical.

3. Close the drilling choke as quickly as practical.

4. Record shut-in casing pressure (SICP) versus time.

Figure 4: In this plot of actual and simulation results for a simple shut-in and pump startup after the kick, Number 1 represents the moment when the valve from the gas source was opened to LSU #2, and BHP increased by more than 500 psi. Number 2 shows pressure in the gas source is reduced as gas was injected into LSU #2.

MPD Pump Shutdown with Choked Flow Check and Shut-in

This method is based on the pump shutdown schedule for routine operations in CBHP. It can be used to check for flow when signals of a kick are not clear. The procedure can be outlined as:

1. Stop drilling, i.e., pick up off bottom and stop rotating.

2. Apply the regular pump shutdown schedule.

3. At the end of the schedule, attempt to hold the casing pressure constant for approximately two minutes by adjusting the drilling choke.

4. If it is necessary to bleed fluid from the well to maintain a constant casing pressure, shut in the well by closing the drilling choke and record SICP versus time.

5. If not (i.e., if the casing pressure does not increase above the final schedule pressure with the choke closed), resume drilling operations but continue monitoring for kick warning signs.

The application of these non-circulating responses is dependent on the RCD static pressure rating. If the expected shut-in casing pressure will exceed the RCD rating, close the annular preventer or pipe rams and open the choke line valve with the well control choke closed rather than closing the drilling choke.

Pump Startup Schedule after a Non-circulating Response to a Kick

Pump startup and pump shutdown procedures are routine MPD operations that are intended to keep wellbore pressure relatively constant during pump manipulations. When a non-circulating response is applied, a new pump startup schedule to start kick circulation is needed. This procedure should keep BHP relatively constant and above formation pressure.

The goal of this part of the research was to document and evaluate a simple method to start the kick circulation and keep bottomhole pressure constant after a non-circulating response. The method uses information available on the rig to define a pump startup schedule.

The proposed procedure is:

1. A routine pump startup schedule should already exist. Table 1 shows an example of a pump startup schedule with four steps of increasing pump rate, Q1 to Q4. As can be seen from the table, when the mud pumps are off, the casing pressure (CP) is equal to the annular friction pressure loss (?PAF) plus a desired casing pressure for routine operations (CP*). Notice that during the startup, CP is reduced by a constant pressure increment (?P) at each step until it is equal to the CP*.

This schedule keeps BHP approximately constant while the mud pump rate is increased. Note that a pump shutdown would use this same schedule beginning at Q4 and reducing the pump rate.

2. Once a potential kick has been recognized and the relevant non-circulating response applied, is recorded until a stabilized SICP can be interpreted.

3. Based on the stabilized SICP and the routine pump startup ?P defined in step 1, a post-kick pump startup schedule is defined for the MPD kick circulation. Table 2 illustrates a post-kick pump startup schedule equivalent to the routine schedule in Table 1. CP0 is selected as equal to SICP plus a desired safety overbalance factor (?POB). At each step in the schedule, the CP is reduced by the same ?P, and the flow rate is increased to the same Q as defined for the routine pump startup.

4. Once the post kick pump startup schedule has been applied, the pump rate stabilized at Q4 and the choke pressure stabilized at CP0 Kick – 4?P, the drill pipe pressure should also stabilize. The kick is then circulated out, keeping drill pipe pressure constant at that value and the pump rate constant at Q4, equivalent to the driller’s method of well control.

Figure 5: In the method of MPD pump shutdown with choked flow check and shut-in, although the casing pressure time response did not match exactly between simulation and experiment, the trends and magnitude of response to choke manipulation are similar.

Well Scenarios for Evaluating Procedures

LSU #2 is a 5,884-ft deep vertical well with 9 5/8-in. casing. Most of the full-scale results of this research were measured in and then compared with computer simulations of LSU #2. Two additional well geometries were used to perform simulations representative of other drilling environments suitable for MPD application: Well X is a 6-in. slim-hole directional well with a potential deep kick zone whereas Well Y is 12 ¼-in. straight hole with a potential high-pressure sand at the bottom.

Simulation Procedure

The computer simulations start during drilling operations just above the high-pressure sands in wells X and Y. For LSU well #2, a high-pressure sand was created in the simulator to emulate real conditions. The bit would drill into the high pressures, where a gas kick would be taken. Two initial responses were used to stop formation flow: simple shut-in and manual MPD pump shutdown with a choked flow check. If a simple shut-in response was used, the pumps were shut down, and then the choke was closed as fast as practical. However, if a manual MPD pump shutdown with choked flow check was performed, the pump shutdown schedule for routine operations was applied.

At the end of the schedule, the casing pressure that was required to compensate for the lost annular friction (?PAF) was held constant for approximately two minutes before shutting the well in, unless shutting in was required to maintain casing pressure. In both cases, SICP was recorded and used to create a pump startup schedule for kick circulation. All simulations were carried out until the gas was circulated completely out of the system; the data was recorded and analyzed using a spreadsheet.

Experimental Procedure

In the experimental well LSU #2, drilling fluid was circulated down the annulus between the 3 ½-in. and 1 ¼-in. tubing at the desired pump rate, and the flow returns were taken through 3 ½-in. by 9 5/8-in. annulus. The gas kick was emulated by injecting gas in the 1 ¼-in. tubing until the desired pit gain was obtained. Subsequently, the planned initial response was applied to stop formation flow and circulate out the kick. The influx was circulated out through a mud-gas separator, where the gas was directed to the flare and the drilling fluid was returned to the mud pits.

During the experiments, the drill pipe, casing and gas-injection pressures at the surface were continuously monitored and recorded, and subsequently analyzed.

Pump Startup, Shutdown for Routine Operations

The method described was tested in a full-scale experiment in the LSU #2 well. The drilling fluid used in the experiment was water, and the kick fluid was natural gas. A restricted valve was used in a surface return line to simulate higher annulus frictional pressure losses. Conventional detection for kicks was used in this experiment. The pump startup and pump shutdown schedule for normal conditions is detailed in Table 3, and Figure 2 shows its application. Solid lines represent data from the actual well, and dashed lines correspond with data for the equivalent computer simulation. Notice that BHP was kept essentially constant during pump startup and pump shutdown.

Simple Shut-in and Pump Startup after the Kick

The full-scale experiment was performed according to the procedure explained above. Gas was injected into the well until a 10-bbl kick was recognized. Then the well was shut in, and SICP was recorded vs time. The initial circulating underbalance was approximately 240 psi. Figure 3 shows shut-in drill pipe (a non-return valve was not used) and casing pressure buildup; the stabilized SICP was interpreted to be 510 psi.

Table 4 illustrates the post-kick pump startup schedule based on the pump startup for normal conditions in Table 3 and the SICP. In this experiment, ?POB was assumed to be equal to zero.

Once the post-kick pump startup schedule was prepared, it was applied manually by two persons: one operated the pump and the other manipulated the choke to control casing pressure. Figure 4 shows the experimental results.

It can be observed from the graph that BHP was kept almost constant during the post-kick pump startup, which achieves the goal described in previous section. Notice that the drill pipe and bottomhole pressure results from the experiment (solid lines) and the simulation (dash lines) are not identical for this case, probably because it was difficult to get exactly the same casing pressure versus time in the simulation as in reality.

However, the similarity in behavior supports the relevance of using simulations for this study. In the plot, number 1 represents the moment when the valve from the gas source (a well that is essentially a gas storage bottle) was opened to the LSU #2 well; notice that BHP increased more than 500 psi. Number 2 in the plot shows how the pressure in the gas-storage well is reduced as the gas is injected into the LSU #2 well to simulate the gas kick. BHP was controlled when the shut-in procedure was applied.

MPD Pump ShutDown with Choked Flow Check, Shut-in

The second non-circulating response was applied to a low feed-in rate kick. The gas-storage well pressure was set to cause a 100-psi underbalance in LSU #2. A restricted flow path from the storage well was opened, a slight increase in return flow was detected as an indication of a possible kick, and the manual MPD pump shutdown with choked flow check was applied.

Figure 5 shows the experimental results. It can be seen that the pump shutdown schedule described in Table 3 was applied to keep BHP constant. The choked flow check was extended over a period of about five minutes to show that the choke had to be opened periodically to maintain the intended casing pressure at the end of the MPD shutdown schedule.

When the choke was subsequently closed, the casing pressure would build back up, indicating that the well had been underbalanced. The experimental results from the test well were also compared with a computer simulation for the same conditions, shown in the dashed lines. Although the time response of casing pressure in the simulation is not an exact match to experiment, both the trends and the magnitude of the response to choke manipulation are similar.

A pump startup schedule (Table 5) was defined for Well X to keep BHP constant during pump manipulations; the schedule is simulated in Figure 6.

Example Simulations of Shut-in and Pump Startup

Simulations were conducted to provide a basis for evaluating the application of pump startup schedules to a wide range of well conditions. This section describes an example of these simulations.

A pump startup schedule was defined to keep BHP constant during pump manipulations (Table 5) for Well X. Figure 6 shows the simulation of the pump startup and pump shutdown for routine operations for the well. It can be seen that BHP is kept almost constant while the pump is being started and shut down.

Notice that casing pressure is used to compensate the loss of friction in the well; it is increased when pump rate is reduced, and it is reduced when pump rate is increased.

Once the pump startup schedule for routine operations was defined, the kick simulation was run according to the procedure described in an earlier section. The well was drilled into a high-pressure sand with a circulating underbalance of 0.2 ppg. As a result, a 20-bbl kick was taken, the well was shut in, and SICP was recorded.

Based on the pump startup schedule for routine operations (Table 5) and the SICP, a new pump startup schedule (Table 6) for kick circulation was built, and the kick was circulated out successfully. The stabilized SICP was equal to 1,183 psi. A safety overbalance (?POB) of 100 psi was added to the SICP for determining the post-kick startup schedule.

Figure 7 shows the simulation results for this scenario. A kick was recognized by the increase of surface mud flow rate out (Qout) and pit gain, as shown by the red line and the green line respectively. Drilling was stopped, the mud pump was shut down, and the choke was closed with a pit gain of about 20 bbls. Casing pressure subsequently increased to balance the kick zone’s pressure at SICP = 1,183 psi. Hence, the BHP increased and stopped formation flow.

At this point, the post-kick pump startup schedule was prepared and applied. Figure 6 demonstrates how casing pressure (purple line) was decreased while pump rate (blue line) was increased according to the schedule in Table 6. Notice that BHP (light blue line) was kept almost constant during the application of this schedule.

Three well scenarios, LSU #2, Well X and Well Y, representing hole sizes from 6 in. to 12.25 in. were simulated. Kicks were simulated in each well configuration for total pit gains of 2, 10 and 20 bbl and for three levels of circulating underbalance. All simulations were run successfully and demonstrated that BHP was kept relatively constant using the proposed post-kick pump startup schedule and procedure.

Table 6: A pump startup schedule for kick circulation in Well X was built. Figure 7 shows the simulation results of shut-in and pump startup for Well X.

Conclusions

• Simple shut-in is a procedure that can be applied during MPD operations to stop formation flow. Previous work has shown that it is the preferred response to a kick if accurate flow-out metering is not available or if a circulating response is impractical due to an equipment failure. Simple shut-in can be applied rapidly, typically in less than one minute, and it can provide a SICP as a basis for a pump startup schedule.

• The MPD pump shutdown with choked flow check response allows checking for flow during MPD operations without letting bottomhole pressure drop significantly below the intended pressure. Consequently, this response can be used to detect or confirm, and then shut in, low feed-in rate kicks that cannot be detected conclusively during circulation.

• The pump startup schedule method described successfully maintains bottomhole pressure relatively constant during pump startups for kick circulation. It is applicable after all non-circulating responses.

• Applicability of these methods was confirmed with both full-scale experiments and simulations covering a wide range of well conditions.

This article is based on IADC/SPE 143094, “Alternative Shut-In and Pump Start-Up Procedures for Kicks Taken During MPD Operations,” IADC/SPE Managed Pressure Drilling and Underbalanced Operations Conference & Exhibition, 5-6 April 2011, Denver, Colo.

Acknowledgements

The authors wish to acknowledge all of the previous researchers of the MPD consortium, especially Majid Davoudi who completed the study of the best initial responses to kicks taken in MPD operations, his contribution was essential to the conclusions reached in this work and are very much appreciated.

We thank SPT Group for providing licenses and technical support for the Ubitts and Dynaflodrill simulators, which were used extensively for this study. We also thank the consortium members: Chevron Energy Technology Corporation, Total E & P, ConocoPhillips, Shell E & P Company (SEPCO), At Balance™, Secure Drilling™, and Blade Energy Partners for their financial and technical support for this research. Note that their participation in the consortium does not indicate endorsement of this work or the conclusions reached.

Finally, we thank the faculty and staff of the Craft and Hawkins Department of Petroleum Engineering, especially PERTT Lab personnel, for their assistance in this research.

References

Arnone, M. A., and Vieira, P. 2009. “Drilling Wells With Narrow Operating Windows Applying the MPD Constant Bottom Hole Pressure Technology—How Much the Temperature and Pressure Affects the Operation’s Design”. Paper SPE/IADC 119882 presented at the SPE/IADC Drilling Conference and Exhibition, Amsterdam, The Netherlands, 03/17/2009. doi: 10.2118/119882-MS.

Das, A. K., Smith, J. R., and Frink, P. J. 2008. “Simulations Comparing Different Initial Responses to Kicks Taken During Managed Pressure Drilling”. Paper IADC/SPE 112761-MS presented at the IADC/SPE Drilling Conference, Orlando, Florida, USA, 03/04/2008. doi: 10.2118/112761-MS.

Das, Asis Kumar. 2007. Simulation study evaluating alternative initial responses to formation fluid influx during managed pressure drilling. MS Thesis, Craft & Hawkins Department of Petroleum Engineering, Louisiana State University, Baton Rouge, LA.

Davoudi, M., Smith, J. R., Patel, B. M., and Chirinos, J.E. 2010. “Evaluation of Alternative Initial Responses to Kicks Taken during Managed Pressure Drilling”. Paper IADC/SPE 128424 presented at the IADC/SPE Drilling Conference and Exhibition, New Orleans, LA, USA, 2-4 February. doi: 10.2118/128424-MS.

Davoudi, Majid. 2009. A Simulation-based evaluation of Alternative Initial Responses to Gas Kicks During Managed Pressure Drilling Operations. MS Thesis, Craft & Hawkins Department of Petroleum Engineering, Louisiana State University, Baton Rouge.

Guner, H. 2009. Simulation Study of Emerging Well Control Methods for Influxes Caused by Bottomhole Pressure Fluctuations during Managed Pressure Drilling. MS Thesis, Craft & Hawkins Department of Petroleum Engineering, Louisiana State University, Baton Rouge.

Hannegan, Don M. 2006. “Case Studies–Offshore Managed Pressure Drilling”. Paper SPE 101855 presented at the SPE Annual Technical Conference and Exhibition, San Antonio, Texas, USA. doi: 10.2118/101855-MS.

IADC. 2009. Glossary of MPD and UBD 2009. Available on http://www.iadc.org.

Malloy, K. P., Stone, R., Medley, G. Harold, Hannegan, D. M., Coker, Ol. D., Reitsma, D., Santos, H., Kinder, J., Eck-Olsen, J., McCaskill, J. Walton, May, J., Smith, K. L, and Sonnemann, P. 2009. “Managed-Pressure Drilling: What It Is and What It Is Not”. Paper IADC/SPE 12228 presented at the IADC/SPE Managed Pressure Drilling and Underbalanced Operations Conference & Exhibition, San Antonio, Texas, 02/12/2009. doi: 10.2118/122281-MS.

Malloy, Kenneth P. 2008. A Probabilistic Approach to Risk Assessment of Managed Pressure Drilling in Offshore Applications. Joint Industry Project DEA155: U.S Department of the Interior Minerals Management Service.

Medley, G., Moore, D., and Nauduri, A.S. 2008. “Simplifying MPD – Lessons Learned”. Paper SPE/IADC 113689-MS presented at the SPE/IADC Managed Pressure Drilling and Underbalanced Operations Conference and Exhibition, Abu Dhabi, UAE, 01/28/2008. doi: 10.2118/113689-MS.

Rehm, Bill. 2009. Managed pressure drilling, Gulf drilling series. Houston, TX: Gulf Pub.

View the original article here