Sunteți pe pagina 1din 15

IADC/SPE 74488

Advanced Slip Crushing Considerations for Deepwater Drilling


U. B. Sathuvalli, Blade Energy Partners; M. L. Payne, BP; P. V. Suryanarayana, Blade Energy Partners;
J. Shepard, GlobalSantaFe

Copyright 2002, IADC/SPE Drilling Conference.


The seriousness of drill pipe slip crushing due to bi-axial
This paper was prepared for presentation at the IADC/SPE Drilling Conference held in Dallas, loading in the slips is well recognized by operators and
Texas, 26–28 February 2002.
drilling contractors. When the margin of overpull is small, the
This paper was selected for presentation by an IADC/SPE Program Committee following
review of information contained in an abstract submitted by the author(s). Contents of the
factor by which the tensile capacity of the drill string is
paper, as presented, have not been reviewed by the International Association of Drilling reduced by slip loading must be known. In particular, due to
Contractors or the Society of Petroleum Engineers and are subject to correction by the
author(s). The material, as presented, does not necessarily reflect any position of the IADC or the larger OD drilling and landing strings, and heavier
SPE, their officers, or members. Papers presented at the IADC/SPE meetings are subject to
publication review by Editorial Committees of the IADC and SPE. Electronic reproduction,
weights, there is some uncertainty in the prediction of slip
distribution, or storage of any part of this paper for commercial purposes without the written crushing loads, and in designing them with adequate safety
consent of the Society of Petroleum Engineers is prohibited. Permission to reproduce in print is
restricted to an abstract of not more than 300 words; illustrations may not be copied. The factors. This uncertainty leads to the consideration of alternate
abstract must contain conspicuous acknowledgment of where and by whom the paper was
presented. Write Librarian, SPE, P.O. Box 833836, Richardson, TX 75083-3836, U.S.A., fax
surface handling systems such as elevators and dual
01-972-952-9435. shouldered landing strings. While there is often a legitimate
need to adopt alternate surface handling systems, especially in
Abstract deep water (as evidenced by the horizons in Gulf of Mexico),
Deepwater drilling trends are moving towards deeper water there is a clear need to further understand the physics of slip
depths as well as deeper exploration horizons and reservoirs crushing. The current understanding of slip crushing is largely
below the mudline. These trends have created various based on a model developed in 19591 and tests reported in
operational and well design challenges. This paper addresses a 19622. While it is clear that manufacturers of slips and vendors
serious problem pertinent to the design of landing strings and of such equipment have a clearer understanding of the slip
drill strings for deepwater operations. These strings are crushing mechanisms, this knowledge is confined to
designed to run long and heavy casings, tiebacks, or liners unpublished internal company reports on tests performed for
(typically including deep intermediate and production specific clients.
tubulars). The total weight of these strings may approach or The purpose of this paper is to assemble all available data
exceed 1,000 klbf at the mudline. Adding in the additional regarding slip crushing and place them in the perspective of
weight of the landing string back to the rotary results in a current understanding. A secondary aim is to describe the
serious design problem regarding both the landing string itself physics of the slip-drill pipe interaction in the some detail, i.e
and the handling equipment. Slip-based handling systems extend the Reinhold-Spiri analysis1 to the next level.
work well in most instances, but lead to bi-axial loading from This paper begins with a review of published work on this
the tension and radial loads exerted by slip inserts. As a result problem and proceeds to describe the results of recent tests
of biaxial loads, the axial load rating of the landing string is made available to the authors by several organizations in the
reduced. The current understanding of slip crushing industry. These results are described, interpreted and discussed
phenomena is based on testing and modeling work dating to in the light of predictions of the Reinhold-Spiri model. Based
1959. This paper examines recent tests in the light of more on this discussion, currently used test procedures are
advanced models and presents an improved understanding of reviewed. Finally, a more detailed analysis (developed in
slip crushing loads. Appendix A) is used to illustrate the physics of slip-drill pipe
interaction. An outcome of this analysis, a modified formula
1) Introduction for slip crushing load is compared with existing data.
“As drilling depths continually increase and hydraulic
efficiency demands the use of 4 ½ or 5 in. OD drill pipe down 2) Review of Previous Work
to completion depth, the vastly greater hook loads now The failure of drill pipe in the region of contact between the
encountered are bringing attention to drill- pipe failures drill pipe and slips was first addressed by Reinhold and Spiri
occurring in the slip area.” These words form the opening in 19591. The chief contribution of this paper was the
statement of a 1962 paper by Vreeland2. If the diameters are recognition that drill pipe is subjected to bi-axial loading in
changed to 5 7/8 in., the preceding statement reflects the the slip contact area. By treating the slip as an immovable
situation today. wedge between a rigid bowl and the hanging drill pipe, a
2 SATHUVALLI, PAYNE, SURYANARAYANA, SHEPARD IADC/SPE 74488

relationship between the axial force (Fa ) on the drill pipe and slip loads. The implications of the work by Vreeland and an
the transverse force (W) exerted on it by the slips was derived. improved model to explain the mechanics of slip crushing are
The ratio of the transverse force to the axial force is known as discussed in further detail in sections 4 and 5.
the “transverse load factor” or simply the “K-factor.” The Additional slip crushing work was performed by
average radial pressure on the drill pipe OD and the axial Hayatdavoudi3 in 1985. This paper deals with the development
stress in the pipe beneath the slip toe are estimated. This of theory to design slip inserts to reduce the probability of
“radial pressure” in turn is used to estimate the tangential yielding in the pipe. The work included representation of
stress at the drill pipe ID. Knowing the tangential and axial strain gage data from slip crush tests on 9 5/8 in., 53.5 ppf
stresses on the ID of the drill pipe, the von Mises equivalent casing. Unfortunately, due to lack of detail about the test
(VME) stress criterion was used to estimate the axial load at procedure and actual data, it is difficult to derive specific
which the drill pipe ID begins to yield. This analysis resulted conclusions from this paper. Nevertheless, the data presented
in the well known Reinhold-Spiri formula for the slip crushing in this paper indicates that the deformation of the casing ID
load of drill pipe, which is currently used in the above the toe of the slip is larger than the deformation near the
drilling industry. top of the slip.
Additionally, significant work on this problem was
performed by Vreeland through a set of carefully engineered 3) Discussion of Test Results
experiments2. In his experiments, Vreeland studied the As noted, despite the importance of slip-crushing limits in the
behavior of eight joints of 5 in., 19.5 ppf, Grade E drill pipe design of deepwater drilling and landing strings, limited data
under slip loads. The test fixture was designed such that the are available in the open literature. Frequently, tests have been
upper part of the test specimen (i.e. the drill pipe joint) was designed to verify performance of the slips rather than
placed in the slips while the lower part was connected to a accurately determine the slip-crushing limit of the pipe. In
piston that exerted axial load via an end piece. Standard and such tests, slip crushing loads have to be derived from
extended length slips in conjunction with standard API bowls inference rather than measurement. Even when the explicit
were used to hold the drill pipe in the upper fixture. Strain goal is the determination of slip-crushing loads, some tests
gages to record hoop strain were installed on the inside of each have been performed with limited attention to experimental
specimen at various distances from the bottom of the slip. In details such as dimensional measurements before and after
each case, the reduction of ID was plotted as a function of testing, mechanical properties, frictional conditions,
axial load. Prior to reaching the slip crushing load, the uncertainty assessment, calibration of sensors, etc. In this
reduction of ID is proportional to axial load. Since yielding light, the work of Vreeland is noteworthy for its attention to
begins on the ID of the drill pipe, the load at which the slope detail in the collection, reduction and interpretation of data.
of the axial load versus ID reduction curve changes is a Nevertheless, it is important to analyze all available data to
measure of the slip crushing load. Based on eight tests, the further our understanding of slip crushing limits.
following major conclusions were made: The following section describes the results of three sets of
1. Under similar conditions, pipe held in standard and tests performed by different organizations in the industry. To
extended length slips exhibits inelastic deformation at preserve anonymity, the organizations and the tests performed
nearly the same load; i.e. the slip crushing load is by them are simply designated by the letters A, B and C.
essentially independent of the slip contact length;
2. Friction between the slip and bowl reduces the transverse 3.1) Test Set A
load factor, i.e. increases the axial load required to initiate This test was performed in Fall 2000 on a single joint of 5 in.,
slip crushing, 34 ppf, S-135 drill pipe and a modified 14 in., 750 ton slip
3. The peak transverse stress occurs somewhere between 5 type elevator with extended arc slips. The slip contact length
in. and 10 in. from the toe of standard slips and between for this type of slip is 18.375 in. The Reinhold-Spiri equation
10 in. and 15 in. from the toe of extended length slips; predicts a slip crushing load of 1,162 klbf if a friction factor of
4. The Reinhold-Spiri formula is a reasonable prediction of 0.3 is assumed between the slip and bowl. This value reduces
the slip crushing load, i.e. the experimental points fall on to 1,108 klbf for a friction factor of 0.2. The test procedure
either side of the theoretical prediction. consisted of initially loading the specimen to 800 klbf.
The work by Vreeland2 is significant for several reasons. Subsequently, the specimen was loaded in predetermined
First, it represents the first and one of the only data sets increments to a maximum value of 1,200 klbf. The upper limit
obtained from carefully planned and well engineered on the axial load was based on an expected slip crushing load
experiments in the industry literature. The eight tests reported of 1,147 klbf. The test procedure and report did not contain
in this work, in the opinion of the authors, are still some of the information on the conditions at the slip-bowl interface, and it
best available test data on slip crushing loads. While slip crush is therefore difficult to estimate the friction factor. However, a
load tests have been conducted more recently, some of these slip crushing load of 1,147 klbf corresponds to a friction factor
tests are useful for qualitative assessments only, and may not of 0.27. At the end of each load increment, the specimen was
afford accurate quantitative conclusions. In the 40 years since drifted with a set of drift bars (with diameters ranging from
this work was published, there have been limited published 3.388 in. to 3.397 in.) to determine onset of yielding on the
papers that analyze the mechanics of pipe under the action of inner surface. (The nominal ID for this specimen is 3.5 in.)
IADC/SPE 74488 ADVANCED SLIP CRUSHING CONSIDERATIONS FOR DEEPWATER DRILLING 3

After the specimen was loaded to 1,200 klbf, the load was curve is expected to have a lesser slope. This is clearly an
cycled 100 times between 150 klbf and 1200 klbf during a 10 artifact of bias errors in the load and strain gage readings. In
hour period. The specimen was drifted after every 5 or 10 the absence of data about zero errors, strain gage locations,
cycles. Based on results of drifting, visual inspection, and etc., definitive conclusions cannot be made.
measurements of OD before and after the test, it was Figure 2 shows the hoop strain versus axial load in the mid
concluded that yielding had occurred on the ID of the pipe slip region for Test # 3. While strain gage 1 indicates a
around 950 klbf. Unfortunately, lack of data to determine the departure from linear behavior at 600 klbf, it is difficult to
deformation of the drill pipe ID as a function of load, and the determine a sharp break from linearity based on data from
absence of data regarding the condition of the slip-bowl gages # 2 and 3. Figure 3 shows the behavior of hoop strain at
interface do not permit more general conclusions. a different location in the mid slip region. As in test # 2, it is
In addition, two other tests were performed on 5 ½ in., 38 not possible to make general conclusions. However, assuming
ppf, S-135 drill pipe with proprietary slips. Unlike the 5 in. that the gages are located in the same cross sectional plane, the
sample, the 5 ½ in samples were strain gaged. While the data in Figure 3 seems to indicate progressive ovalization of
details of the testing procedure and loading sequence were not the drill pipe. Since the gages were mounted on the OD of the
available, results indicated that yielding occurred between drill pipe, departure from linearity cannot be interpreted, since
1,000 klbf and 1,200 klbf, which was less than the value of yielding begins on the ID and proceeds outward.
1,260 klbf predicted by the Reinhold-Spiri equation for this In summary, despite the availability of strain gage data,
combination of drill pipe and slip. Based on these three tests, due to lack of information about the location of strain gages,
and field observations, it was concluded that the slip crushing and slip geometry, it is difficult to make engineering
load is 83.3% of the value predicted by the Reinhold-Spiri conclusions from the data in these tests.
model.
3.3) Test Set C
3.2) Test Set B Table 2 shows the test matrix for this series of tests. A
This series of five tests was performed in Summer 2001 schematic of the test-setup is shown in Figure 4. The drill pipe
per Table 1. The purpose of the tests was primarily to verify is held within slips as shown, and two sets of ID sensors are
slip performance. The data gathered and the test report suggest mounted inside the drill pipe. In the discussion that follows,
that determination of slip crushing load may not have been an they are referred to as the top and bottom sensors. Each set of
explicit goal. This is indicated in tests 4 and 5 of Table 1, sensors consists of three individual radial displacement
where the tests were performed on solid 5 in. diameter sensors, mounted at 120º azimuthal intervals. Though the
steel bars. exact location of the sensors with respect to the geometry of
Strain gages were installed on specimens for tests 1 the slip was not recorded, discussion with personnel from the
through 3. The test report does not describe the location of test lab indicated that the bottom sensors were coincident with
strain gages. However, based on conversations with the toe of the slip. However, the top sensors were placed
participants who witnessed the test, the strain gages appear to below the top of the slip contact length. Load is applied on to
have been mounted on the OD of the specimens. The gages the drill pipe, and load and radial displacement are recorded as
were mounted so that the slips inflicted minimal damage. a function of time. Figure 5 shows the typical load sequence
Further all test specimens were covered with duct tape to (in particular, the figure shows the load sequence for the first
prevent damage from the slips. In each of the three tests, the three target loads on a 5.5 in. specimen). The test is conducted
sample was loaded in tension, and the load and strain gage at progressively increasing target axial loads. For each target
readings were noted. The data from the strain gages was load step, the load was slowly increased from zero to the
provided in the test report. target load, held at the target load for a short period of time
The data from Test #1 was invalidated since the slip toes (typically 2 to 4 mins), and then unloaded. During the entire
were in contact with the tool joint of the specimen, (i.e. the loading and unloading sequence, the load and radial
specimen was mounted upside down). Figure 1 shows selected displacement measured by both sets of sensors was recorded.
strain gage data from Test #2. The x-axis represents the Each curve represents the loading and unloading for one target
average of strain gage readings at a given location, while the load. The loading and unloading from a single test is shown in
y-axis represents the axial load. At approximately, 1,200 klbf, Figure 6 as a function of displacement. The hysteresis between
the strain gage readings indicate the beginning of yield in the loading and unloading is very clear. It is important to observe
mid-slip region. Not surprisingly, the section below the slip that this hysteresis is a consequence of friction and other non-
continues to exhibit a linear load versus deflection curve. If elastic effects on the displacement sensors, and not necessarily
the load versus strain gage readings for the “below slip” data, indicative of yielding in the material. Before interpreting the
are converted to stress versus strain data (using the geometry results of the tests, it is important to note that the loading
for a 5 in., 34 ppf drill pipe), the slope of the curve turns out to sequence described in Figure 5 roughly simulates the loading
be 3 x 107 psi. However, visual inspection of Figure 1 (and in the field as drilling proceeds, and joints are added to
calculations) indicate that the slope of the load versus strain the string.
curve for the mid-slip section is greater than the slope of the
curve for the section below the slips. Intuitively, the former
4 SATHUVALLI, PAYNE, SURYANARAYANA, SHEPARD IADC/SPE 74488

3.4) Interpretation of data from Test Set C In this work, the point of departure from linearity is
As noted, the test procedure used sensors to measure radial interpreted as the onset of yield, and the axial load
displacement. Only the sensing ends of these sensors were in corresponding to this point is the “slip crushing” load. In
contact with the drill pipe ID. The other end of each sensor is, order to obtain a clearer comparison of the top and bottom
(though in principle stationary), likely to move. Thus the sensor measurements, the average displacement is normalized
accuracy and uncertainty of these sensors is likely to impact with respect to the first displacement measurement (for the
the quantitative evaluation of data. However, data from these first target load).
sensors can still be used in a qualitative sense to indicate The availability of bottom and top displacement
departure from expected trends, rather than provide measurements is very useful to estimate the stress distribution
quantitative estimates for radial displacement. It is important along the slip contact length. In particular, one goal of this
to recognize this at the outset, since the displacement exercise is to identify the load at which yielding occurs at the
resolution is of the order 10-3 inches (1 mil). Though the bottom and at the top locations. Figures 8 through 10 depict
standard practice to measure strains of these magnitudes is by the load-displacement results from the data on 5.5 in. tests.
the use of strain gage rosettes, it is clear that the small ID of Figure 8 shows the zero-corrected axial load vs. top and
the specimens makes the installation of strain gages on the ID bottom radial displacements from one top sensor and one
very difficult, if not impossible. Ensuring that the strain gage bottom sensor (sensor 1 in the raw data). Figure 9 shows the
is also located at the location of maximum stress is an even relationship of the zero-corrected axial load with the average
further challenge. top and bottom displacements. From Figure 8 we note that the
Figure 7 illustrates the raw displacement data as recorded displacements measured by the top sensor are about 25%
by each of the three bottom sensors, for a single target load greater in magnitude than that measured by the bottom sensor.
test. Although the trends are reasonable, the significant This is possibly due to the bias error in the sensors. As noted
difference between each of the displacement sensors is earlier, this prevents the use of absolute displacement values
possibly due to a combination of asymmetry in loading, to derive quantitative conclusions.
measurement errors of each sensor and aberrations in Figure 9 shows that the average displacements (average of
specimen geometry. Pre-test examination and measurements the three individual sensor measurements) at the bottom and
indicated that the ID was more of a hexagon with rounded top locations are closer to each other, as would be expected.
edges than a circle. (A similar observation was made in the (Unfortunately, this is not always the case). A more
report submitted for Test Set A). In addition, discussion with reasonable comparison can be made between top and bottom
personnel from the test lab indicated a certain amount displacements by normalizing average displacements as
eccentricity in the ID bore. described earlier. The zero-corrected axial (target) loads are
Though sensor # 2 typically gave negative readings (as plotted against the normalized average top and bottom
shown in the figure) for most of the tests, there was no pattern displacements in Figure 10. Despite the displacement
to the response and measurements of any of the sensors based inaccuracies, several important conclusions can be drawn from
on the data provided (in spite of the fact that the only change the figures.
between tests was (presumably) a change in target loads). The • The load-displacement behavior clearly indicates the
absolute sensor readings are unlikely to be useful in estimating point of departure from linearity.
actual deformations, especially in the absence of information • Departure from linearity is evident at the bottom location,
on the calibration, accuracy and uncertainty of each sensor. while it appears that the load-displacement relationship
However, the trends of displacement measurements can be continues to be linear at the top location up to the
very useful in assessing slip crushing loads. Since the load maximum target load tested.
measurements are likely to be more accurate, they can be • The axial load at which the departure occurs is the same
relied upon. The load measurements at the start of each regardless of whether the point of departure is determined
experiment, (i.e. the load sensor reading at no load) provide an from a single sensor measurement, average of three
indication of the zero bias error. This was used to obtain a sensors at each location, or the normalized average
corrected load. For each test, it can be noted that the target displacement.
load was held constant for several minutes. The load and the • According to Figure 10, for the tests on the 5.5 in. drill
displacement measured during this “constant load” time period pipe, departure from linearity occurs at an axial load of
are estimated from the data. The load is the average of the 1,058 klbf.
corrected load over the time period. Top and bottom If the peak target loads (see Figure 5 for definition of
displacement measurements over the time period are also target load) are plotted as a function of displacement, slip
averaged. Two averages are considered: the average of sensor crushing appears to commence at the bottom location (the
1 measurements over the time period and the average of all highest slip crushing loads are at the bottom, or near the toe of
three sensors over the time period. Pairs of load and the slips), at an axial load of 1,058 klbf for the 5.5 in. drill pipe
displacement are obtained this way for all target loads. Figure 11 shows the load versus normalized displacement for
The interpretive exercise is then one of identifying the the 5 in. drill pipe tests, and Figure 12 shows the results for the
axial load at which yielding just begins. This is indicated by a 6 5/8 in. drill pipe tests. Figure 11 clearly illustrates the onset
departure from elastic behavior in a load-displacement curve.
IADC/SPE 74488 ADVANCED SLIP CRUSHING CONSIDERATIONS FOR DEEPWATER DRILLING 5

of yield at the bottom location before the top location yields. of the values shown in Table 4 with those shown in Table 3
In this case, yielding seems to initiate at the bottom at an axial indicates that the test results according to column 2 of Table 3
load of 1,157 klbf. However, Figure 12, which shows the show good agreement with the Reinhold – Spiri predictions
results for 6 5/8 in. drill pipe, displays considerable scatter in for a slip to bowl friction factor friction factor of 0.2. On the
the data. It is not clear where yield initiates, although the other hand, if departure from linearity of the load
bottom location curve seems to display a change in slope at displacement curve is used as a criterion, the results shown in
around 1,000 klbf. column 3 of Table 3 (which are indicative of the slip crushing
From the preceding discussion, it is clear that the load), are lower than the values predicted by the Reinhold
uncertainty in the displacement measurements is responsible Spiri equations, by 30% for the 5 in . and 6 5/8 in. sizes, and
for much of the scatter in the curves. Also to be noted is the by 19% for the 5 ½ in. size. (These percentage reductions are
fact that the Figures 10 through 12 show the displacement at based on the calculated values for friction factor of 0.2).
the target load, as a function of changing target loads (i.e., Apart from the test results discussed so far, discussions
different data sets). Importantly, recall (see Figure 5) that each with drilling engineers in other organizations in the industry
loading sequence is followed by unloading to zero load. indicate that tests on other slip – drill pipe configurations have
Besides introducing uncertainty in the data (due to hysteresis, been performed. Unfortunately, the results of these tests have
see Figure 6), the loading and unloading has a significant not been published. Of these, the results of tests on 5 7/8 in.,
implication on the determination of the slip crush load, i.e., the 29.5 ppf, S-135 and 5 7/8 in., 26.68 ppf, S-135 drill pipe with
point at which yielding first begins. This is discussed in more slips of contact length, 16 in., indicated slip crushing at 550
detail in section 4. klbf and 630 klbf respectively. If a friction factor of 0.2 is
Recall that the main interest is the point at which the load- assumed, these test loads are 19.8% and 19.3% less than the
displacement curve departs from linearity (i.e., we are theoretical values respectively.
interested in the trend, and not in the magnitude of the Though it appears, that there is not enough reliable test
displacement). This information can be obtained with less data to clearly determine slip crushing loads, the available
exposure to sensor uncertainties between data sets by results, when collated together seem to indicate that theoretical
examining the load-displacement behavior while loading values of the slip crushing load (as predicted by the Reinhold-
during any given test. In order to capture possible yielding, Spiri model) should perhaps be de-rated by a factor of 20%. In
the data sets of most use are those that have high target loads order to conclusively establish, a de-rating/safety factor, the
(in excess of the possible yield initiation loads). This is shown need for experiments that account for statistical uncertainties,
in Figure 13. The figure plots applied axial load against minimize experimental errors, and use accurate procedures
normalized average radial displacement at the bottom location, is clear.
for all three drill pipe sizes, during the loading phase to the
predetermined target load. Data obtained while loading to 4) Test Procedures
maximum target load were chosen for this figure. A critical parameter in the Reinhold-Spiri formula that is
From the figure, the departure from linearity can be currently used to calculate the slip crushing load is the friction
estimated as shown in column 3 of Table 3. Column 2 of this factor at the slip-bowl interface. Though the commonly used
table also shows the results obtained from Figures 10 through value for the friction factor is 0.2 (which results in a K-factor
12. As the table illustrates, the yield initiation load is much of 2.64), the friction factor is known to vary widely depending
lower than is indicated by the more “averaged” load based on on the lubricant (pipe dope) and mechanical condition of the
progressively increasing target load tests. Indeed, identical mating surfaces of the slip and bowl. Discussions with slip
yield initiation loads (shown in column 3 of Table 3) are vendors indicate that, depending on the lubricant used friction
indicated when data from the loading phase of previous target factors can vary as follows4: 0.197 for a type of proprietary
loads are analyzed. pipe dope, 0.27 for preservative oils, 0.274 for greased
Figure 14 and Figure 15 show the load-displacement surfaces, 0.286 for lightly oiled surfaces and 0.29 for dry
behavior for the 5 in. and 5.5 in. drill pipe tests, respectively, mating surfaces. The tests reported by Vreeland2 were
during three different load cycles. The 5 in. test results clearly performed with dry and greased slips in API no. 3 bowls.
indicate the departure from linearity, in the neighborhood of However, the effect of variable friction factor in the tests
800 klbf in all cases. The 5.5 in. results are not as clearly discussed in section 3 is not apparent.
indicative of the departure, but the departure appears to be Secondly, the phenomenon of slip crushing results from bi-
between 850 and 900 klbf in all cases. Since yielding has axial loading of the drill pipe. Figure 16 compares the load-
already occurred by the time the first of the three load cycles deflection curves obtained when drill pipe is loaded in simple
depicted in the figures is completed, some form of strain tension and in slips. In simple tension, the behavior of the drill
hardening and yield modification is likely, although this pipe is represented by the curve OABD. The load is
cannot be indicated by the data being examined in this work. proportional to deflection until point B when the stress in the
Also noteworthy is the fact that in all the data analyzed, the drill pipe reaches the yield point. The drill pipe fails
top location does not approach yield. catastrophically at a slightly higher load (at point D, when the
Table 4 shows the predicted values of the slip crushing stress in the drill pipe cross section equals the tensile strength
loads according to the Reinhold-Spiri formula. A comparison of the material). Importantly, the margin between the point of
6 SATHUVALLI, PAYNE, SURYANARAYANA, SHEPARD IADC/SPE 74488

yield (point B) and point of catastrophic failure (point D) is and photoelastic methods to determine stress and strain
not very large (and is roughly the cross sectional area times profiles may be considered.
the difference in tensile and yield strengths of the material). • Most importantly tests should be designed to assess
On the other hand, when drill pipe is loaded in slips, the statistical uncertainty of results, and enable determination
region near the toe of the slip is subjected to both radial and of a safety factor that can be used to de-rate theoretical
axial stress. The radial stress due to slips at the drill pipe OD values.
manifests as a large hoop stress on the ID. In general, the hoop Conversations with manufacturers of slips and drill pipe
stress on the ID is magnified by a factor of [2OD2/(OD2-ID2)]. handling equipment indicate that they have each performed
Therefore, the maximum VME stress occurs on the ID, and at extensive internal DVTs to qualify their equipment. As noted,
the slip crushing load, yielding begins at the ID and migrates this information has not been published, or easily accessible. It
outward. The load-deflection curve is linear until the slip is hoped that the aggressive exploration of deeper horizons,
crushing load (represented by point A in Figure 16) is reached. and the need to optimize drill strings and surface handling
At higher loads, an elastic-plastic boundary migrates outward, equipment will spur the need to discuss and share data related
and the load-deflection curve resembles AC. Beyond point A, to this problem in published literature.
the behavior is determined by the material stress strain curve.
Due to gradual migration of the plastic zone from the ID to the 5) Theoretical Considerations
OD, the slope of the curve in section AC is higher than the As noted in section 2, Reinhold and Spiri1 presented a
slope of the curve in section BD. This feature is clearly relationship between the axial force (Fa ) on the drill pipe and
illustrated in the results of the eight tests reported by the transverse force (W) exerted on it by the slips, i.e.,
Vreeland2. Specifically, Vreeland notes that “An increase in W = KFa (1a)
load of 100,000 lb over the yield load produces less than where K is the transverse load factor given by
0.005 in. permanent deformation” with standard and extended 1 − µ tan α
length slips in lubricated bowls. This is also evident in the K= (1b)
results of Test Set C discussed in this paper (see Figures 13, µ + tan α
14, and 15). Therefore, exceeding the slip crushing load does α is the slip angle and µ is the coefficient of friction between
not cause immediate catastrophic failure. However, note that the lateral surface of the slip and the bowl. The transverse load
the load at which catastrophic failure occurs during biaxial factor described by Eq. (1b) is widely known as the “K-factor”
loading (represented by point C), is less than the load required in the drilling industry. The axial and average radial stresses,
to cause total failure in simple tension. When the drill pipe is σa and σr,av on the drill pipe can now be determined from axial
unloaded, it unloads along the dashed line in Figure 16, and it and transverse forces described by Eqs. (1), so that
is permanently deformed. As demonstrated in tests by F
Vreeland, deformation of the cross section may not visibly σa = a (2)
A1
manifest on the OD till significant plasticization has occurred.
Thus, unless the test procedure is designed suitably, it is W  K 
σ r ,av =
=  Fa (3)
difficult to determine onset of yielding and by extension, the AL  AL 
slip crushing load. In this respect, the experiments reported in where, A1 is the cross sectional area of the drill pipe and AL is
Test Set C are noteworthy, since they illustrate onset of non- the slip to drill pipe contact area. By using Lame’s equations
linearity in the load-deflection curve in all instances. for a thick walled cylinder, the axial and hoop stresses on the
The preceding discussion suggests a few guidelines for ID of the drill pipe can be estimated from the radial stress
conducting slip crush tests that produce results that can be acting on drill pipe OD. Since the radial stress is a function of
interpreted quantitatively: the axial load, the axial load at which the VME stress at the ID
• Since the physics of slip crushing does not change with of the drill pipe reaches the stress required to cause yielding
slip/tubular geometries, it is preferable to choose large can be found. This load is the well known “slip crushing load”
diameter tubulars, so that strain gages may be installed at derived by Reinhold and Spiri1:
different ID locations. Care must be taken however to 1
choose tubulars with OD/WT ratios less than 10.  2
• Test procedures must be supplemented by data that  
describe the stress strain behavior of tubular steels. An  2 
Fa , slip crush = σ yp A1  2 
estimate of actual versus nominal values of yield and   2d o
2
2
KA1   2d o
2
KA1  
tensile strengths of the specimen must be known while 1
  +  1 +  +   
d o − d i AL   d o − d i AL  
 
2 2 2 2
interpreting results. Needless to add, pre-and post-test  
measurements of drill pipe and slip inserts should be .
recorded. ………. (4)
• While strain gages are by far the cheapest and reliable
transducers, other methods such as photosensitive paints, Apart from deriving Eq. (4), the work by Reinhold and Spiri1
suggested that the “maximum axial and transverse loads do
IADC/SPE 74488 ADVANCED SLIP CRUSHING CONSIDERATIONS FOR DEEPWATER DRILLING 7

not act at the same cross section in the slip area.” They also L
suggested qualitative distributions for the axial load and radial V3 = − (5j)
A1 E1
pressure on the drill pipe in the slip contact region (see
Figure 17). The theoretical basis for the above formula is discussed in
These profiles suggest that the axial force in the drill pipe detail in Appendix A. The symbol definitions and geometry
is a maximum at the toe of the slip and decreases towards the are defined in the nomenclature and Figure A-1. Importantly,
top where it vanishes entirely. On the other hand, the note the similarity of Eq. (5a) with Eq. (4). Equation (5a) is
suggested radial profile looks like a flattened parabola, with essentially identical to Eq. (4), with the factor (K/AL) replaced
zero radial stress at the extremities of the slip contact region. by Kmod. The term Kmod defined by Eqs. (5b) through (5i) are
The profiles described in Figure 17 were based on intuitive functions of the slip and drill pipe geometry alone. Also, note
reasoning rather than theoretical analysis. To the best of the that the definition for Kmod in Eq. (5b) incorporates the K-
authors’ knowledge, these profiles have only been partly factor, which is shown to be a consequence of the problem
confirmed by some of the experimental data reported by geometry in Appendix A.
Vreeland2. In the Appendix to this paper, the mechanics of Figure 18 shows the axial and radial stress profiles in the
slip loading is examined in some detail. An outcome of the slip contact region, as predicted by the current model (Eqs.
analysis is the distribution of axial and radial stresses in the (A-7)and (A-4b) in Appendix A). The radial stress as
slip contact region that is discussed ahead. In addition, the predicted by Reinhold and Spiri in Eq. (3) is also shown. Note
following formula is developed to estimate the slip that the axial stress and (therefore the axial force) is
crushing load. reasonably similar to the profile suggested in Figure 17. The
radial stress profile is however different, especially at the
extremities of the slip contact length, where the radial stress is
1
clearly not equal to zero. In fact, the maximum radial stress
 2 occurs at the toe of the slip. Figure 19 shows the stresses
  caused at the drill pipe ID due to the stresses at the slip-drill
Fa  2  pipe interface. The figure indicates that the VME stress at the
= 2 2 
σ yp
A1 
1 +  2d o
2
  2d o
2
  drill pipe ID is 135 kpsi at the toe of the slip, indicating the
  
  d 2 − d 2 K mod A1  + 1 + d 2 − d 2 K mod A1  
Mod Slip Crush

initiation of yielding. The discrepancy between the radial


  o i   o i  
profile suggested by Reinhold and Spiri1 and the profile in
(5a)
Figure 17 can be explained by noting the difference in the slip-
where
drill pipe geometry assumed to derive the model in Appendix
Go 1  A versus the actual geometry of the slip.
K mod =  − θL  K . (5b)
πd sb sec α  L  The equations developed by Reinhold and Spiri as well as
2 those developed in Appendix A of this work, assume the
1 + ν  d sb  1 − ν
Go = +  
 (5c) geometry shown in Figure A-1. This configuration assumes
2  do  2 that the support from the bowl extends all along the length of
β − V1 + V3 L the slip-drill pipe interface. In reality however (see Figure 20),
θ= (5d) the slip-drill pipe contact area extends above and below the
(β − V1 )L2 − (γ − V2 ) slip. Consequently, the reaction on the slip OD at the slip
extremities is zero, and this causes a reduction in the radial
1  1  d st 2 − d o 2  d sb  d st − d o d sb + d o  stress above the peak values. As a result, the radial stress
β=  ln 2 − 
 2d ln d − d d − d

 profile is likely to resemble the skewed blue line shown in
πE 2 tan 2 α  2  d sb − d o
2
 o  sb o st o  Figure 17. The peak radial stress, will however be expected to
be much closer to the bottom of the slip, rather than inside,
(5e) thus implying the initiation of yielding in the vicinity of the
1 toe of the slip.
γ = IA (5f) This argument indicates the possible reason for the higher
2πE 2 tan 3 α
than observed predictions of slip crushing loads for tests
discussed in section 3. The Reinhold-Spiri formula implicitly
d sb + d o2  d (d − d sb ) 
2
 d st 2 − d o 2  uses a larger slip contact area than is truly available, and thus
I A = (d st − d sb ) − tanh −1  o 2 st  − d sb ln 2 
2 
do  d o − d st d sb   d sb − d o  underestimates the radial stress on the drill pipe OD. In reality,
most of the transverse stress on the drill pipe is a result of the
(5g)
horizontal component of the of the bowl reaction (see Figure
L2 A-1). The axial stress in the drill pipe decays rapidly with
V1 = − (5h)
2 A1 E1 distance from the slip toe, and the drill pipe sheds load onto
L3 the slips. The slip is essentially an elastic foundation that
V2 = − − (5i) transmits the radial component of the bowl reaction to the drill
3 A1 E1
pipe. Since the slip has the least thickness at the toe, it stores
8 SATHUVALLI, PAYNE, SURYANARAYANA, SHEPARD IADC/SPE 74488

the least amount of strain energy at the toe, and thus transmits 5. Herein, best available analysis of nearly all existing data
the largest stress to the drill pipe. It is therefore, conservative has been made. Results show variations between
to neglect sections of contact area at the slip-drill pipe matching the historical Reinhold-Spiri model (with a K
interface that extend above and below the rotary bowl. When factor of 2.64) to results indicating a non-conservative
viewed in this light, conclusion # 1 (see section 2) obtained by margin of ~20% in some cases. Through the development
Vreeland2 which states that the crushing load is the same for of a more accurate slip crushing mechanics model, these
both standard and extended length slips can be justified, since trends can be accounted for, since it clearly supports a
the maximum stress is likely to occur at the point nonlinear stress distribution in the pipe, with the predicted
corresponding to the bottom of the bowl. location of peak stresses being consistent with general
Finally, Table 5 shows the slip crushing loads as calculated test observations.
by Eqs. (5). Due to reasons explained in the preceding 6. With better quantification of actual friction factors, load
paragraphs, a contact length of 12.75 in. was used in uniformity, and equipment tolerancing and fits, this new
estimating the values shown in this table. The values in this slip crushing model and its underlying approach provide
table are 25 to 33 % lower than the Reinhold-Spiri predictions an appropriate framework for much more accurate rating
for a friction factor of 0.2. However, the values shown in of tubulars for slip crushing performance. Until such time
Table 5 must be treated with caution due to the assumptions that these refinements are made and available, drilling
involved in the derivation. The model in Appendix A must be personnel should consider and implement appropriate
viewed only as a step ahead in understanding the mechanics of safety factors when dealing with marginal designs for slip
slip loading and not as a final predictive formulation. Such an crushing scenarios.
outcome is possible after more testing that is supplemented by 7. As a result of the complexities of this problem and the
further analysis, both finite element and otherwise. significant consequences of failure and/or equipment
damage, it is likely that alternative product and equipment
6) Conclusions and Summary Statements systems will gain in popularity. Such options currently
1. Deepwater drilling trends towards deeper wells and water include enhanced design slip systems, “slip-proof” drill
depths have created critical design issues around drill pipe5, dual-elevator (shuttle-based) handling systems,
strings, landing strings, and the associated slip crushing and machined dual shoulder drill pipe for elevator
limits on both. only handling.
2. Experimental study of slip crushing is complex.
Difficulties include: 7) Acknowledgments
• accurate measurement of ID stresses in drill pipe, The authors gratefully acknowledge the help of Dr Ed Bailey
• locating peak stresses, of Stress Engineering Services, Mr Michael Breaux of Frank’s
• accounting for asymmetric loads and stresses, Casing Crew, Ms Patricia Cimpson of BP, Mr Glenn
• system hysteresis, McFarland, Dr Kurt Vandervort and Mr Lane Wilson of H O
• nonlinear elastic-plastic behavior, Mohr Engineering, Mr Robert Chou, Mr Dave Mason and Mr
• and sensitivity to and variations of interface friction. Gregg Sentell of VARCO, Mr Joe Ramey of Access Oil
Due to these complexities, limited quantitative data can Tools and Mr Dave Carr of Houston Holloway for providing
be derived from some test data sets to refine current information used in preparing this paper.
understanding of the precise mechanics of the problem. The authors also acknowledge the management of BP,
3. Several slip crushing test projects have effectively served GlobalSantaFe and Blade Energy Partners for providing the
only as specific proof tests on actual equipment tested. materials, time and funding to write this paper.
Even normalization techniques, such as accounting for
variation between actual yield strength of tested drill pipe 8) Nomenclature
versus the minimum yield strength that would be present A1 Cross sectional area of drill pipe, [L]2, in2
in field joints, have not been consistently applied in these A2(z) Cross sectional area of slip at z, [L]2, in2
test programs. do Outer Diameter of drill pipe, [L], in
4. Various aspects of existing slip crushing test procedures di Outer Diameter of drill pipe, [L], in
have the potential to generate significant non-conservative dsb OD of slip at toe, [L], in
effects which would tend to overestimate the actual slip dst OD of slip at top, [L], in
crushing load. Fortunately, the mechanics of slip crushing ds(z) OD of slip at distance z from toe, [L], in
are such that the yield failure mode is somewhat gradual E Modulus of elasticity, [M][L]-1 [T]-2 , psi
and the system retains reasonable stiffness and stability Fa Axial force on the drill pipe, [M][L] [T]-2, lbf
after initial yield. The post-yield behavior of the drill pipe Go See Eq. (5c)
under slip crushing has provided a “safety net” to the G(z) See Eq. (A-5b)
industry and certainly prevented more serious problems in H1 See Eqs. (A-7) and (A-10a), [M][T]-2, lbf/in
borderline operations. H2 See Eqs. (A-7) and (A-10b), [M][L]-1[T]-2, psi
H3 See Eqs. (A-7) and (A-10c), [M][L][T]-2, lbf
IA See Eq. (5f), [L], in
IADC/SPE 74488 ADVANCED SLIP CRUSHING CONSIDERATIONS FOR DEEPWATER DRILLING 9

K Transverse Load Factor (K-factor), see Eq. (1b) 10) Appendix A: The mechanics of slip-drill pipe load
Kmod Modified K-factor, See Eq. (5b), [L]-2, in-2 transfer
L Slip-drill pipe contact length, [L], in Figure A-1 shows the geometry of the drill pipe and slip
N1(z) Force in drill pipe at z, [M][L] [T]-2, lbf interaction. The hanging weight of the drill pipe below the slip
N2(z) Force in slip at z, [M][L] [T]-2, lbf is transferred to the slips in the slip contact region AD. For the
R Reaction at slip bowl interface, [M][L] [T]-2, lbf purposes of this analysis, the bowl is assumed to be an integral
S Outer lateral surface area of the slip, [L]2, in2 part of the rig floor (and therefore a member of infinite
V1 See Eq. (5h), [M]-1[L][T]2, in2/lbf stiffness) that can provide the reactive loads necessary to
V2 See Eq. (5i), [M]-1[L]2[T]2, in3/lbf support the hanging weight of the drill string. Force balance
V3 See Eq. (5j), [M]-1[L]0 [T]2, in/lbf on the drill pipe and slip, at a cross section ZZ’ perpendicular
W Transverse force on the drill pipe due to slips, lbf to the drill pipe axis, requires that
z Distance from slip toe, [L], in F (a ) = N1 (z ) + N 2 (z ) , (A-1)
α Slip angle (see Figure 1), radian where N1(z) is the force in the drill pipe cross section at a
β See Eq. (5e), [M]-1[L][T]2, in2/lbf distance z from the slip toe, and N2(z) is the force at the same
γ See Eq. (5f), [M]-1[L]2[T]2, in3/lbf location in the slip. Since the drill pipe supports all axial load
µ Coefficient of friction between slip and bowl just below the toe of the slip and no axial load at the top of
ν Poisson’s ratio the slip,
θ See Eq. (5d), [L]-2, in- N1 (z = 0) = Fa (A-2a)
σa Axial stress in drill pipe, [M][L]-1 [T]-2 , psi N1 (z = L ) = 0 . (A-2b)
σR(z) Normal stress on slip OD, [M][L]-1 [T]-2 , psi
σR,r(z) See Fig. A-2 and Eq. (A-4), [M][L]-1 [T]-2 , psi Now consider the equilibrium of the slip, as shown in
σr,av Average radial stress on drill pipe in the slip contact Figure A-2. Note that the reaction force R and frictional force
area, , [M][L]-1 [T]-2 , psi µR at the slip bowl interface have been replaced by a normal
σw(z) Normal stress on drill pipe OD, [M][L]-1 [T]-2, psi stress σR(z) and a shear force τ(z), which act on the elemental
lateral surface B’C’. Also, note that the shear stress on the slip
9) References at the slip-drill pipe interface is replaced by equivalent forces
1. Reinhold, W. B.and Spiri, W. H., 1959, “Why does drill acting on the two cross sectional faces of the slip at z and z
pipe fail in the slip area?,” World Oil, October 1959, pp. +∆z. Vertical force equilibrium requires that
100-115 N 2 (z + ∆z ) − N 2 (z ) = [σ R (z ).∆S ]. sin α + µ [σ R (z ).∆S ]cos α .
2. Vreeland, Jr., T., 1962, Deformation of Drill Pipe Held in By expressing the lateral surface area (see Figure A-2) of the
Rotary Silps,” ASME paper No. 61-PET-20, Presented at slice in terms of the slip geometry, the above equation
Petroleum Mechanical Engineering Conference, Kansas simplifies to
City, MO, Spetember 24-27, 1961 of the American ∂N 2 (z )
Society of Mechanical Engineers. = σ R (z )[µ cos α + sin α ].πd s (z ). sec α .
∂z
3. Hayatdavoudi, A., 1985, “Elastic Yield of Casing Due to
Elevator/Spider System,” Paper SPE 13449, Presented at Recognizing that Eq. (A-1) implies
the 1985 SPE/IADC Conference, New Orleans, March ∂N 2 (z ) ∂N (z )
=− 1 ,
6-8. ∂z ∂z
4. VARCO, 2000, “Failed Slips cause Investigation Report,” the normal stress at the slip-bowl interface is obtained
Internal Company Report. as follows:
5. Wilson, G. E., 1996, “A New Drill Pipe Design Virtually 1 1 ∂N1 (z )
σ R (z ) = − . (A-3)
Eliminates Failures that result from Slip Damage,” Paper πd s (z ). sec α [µ cos α + sin α ] ∂z
SPE/IADC 35036, Presented at the 1996 SPE/IADC
Drilling Conference, New Orleans, LA, March 12-15. Having determined the normal stress at the slip bowl interface,
6. Timoshenko, S, and Goodier, J. N., 1963, Theory of the radial stress on the outer surface of slip σR,r(z) is obtained
Elasticity, 3rd Edition, McGraw Hill Book Company, by resolving the normal and shear forces shown in Figure A-2
New York. in the radial direction, so that
7. Roark, R. J. and Young, W. C., 1975, Formulas for Stress 1  cos α − µ sin α  ∂N 1 (z )
σ R ,r ( z ) = − . (A-4a)
and Strain, 5th Edition, McGraw Hill Book Company, πd s (z ). secα  µ cos α + sin α  ∂z
New York, pp. 470-471.
8. Ziegler, F., 1991, Mechanics of Solids and Fluids, Note that the term enclosed in square parentheses on the right
Sprnger Verlag, New York, pp. 242-243. hand side of the above equation is the K-factor obtained by
Spiri and Reinhold in Eq. (1b) in the body of the paper.
Therefore,
K ∂N1 (z )
σ R ,r ( z ) = − . (A-4b)
πd s (z ). secα ∂z
10 SATHUVALLI, PAYNE, SURYANARAYANA, SHEPARD IADC/SPE 74488

The slip is a thick walled hollow frustrum of a cone subjected that the integral of the radial component of the stress acting at
to the loads described in Figure A-1. To obtain the radial the slip bowl interface reduces to Eq. (1a), i.e.
stress at the slip-drill pipe interface, the problem of a thick W = ∫ σ R ,r (z )∆S . = KFa (A-6b)
walled conical frustrum should be solved. The solution to the BC

behavior of thin walled conical shells under different loads can Note that the truth of Eqs.(A-6) is a result of imposing the
be found by using the theory of plates and shells6, 7. However, boundary conditions described by Eq. (A-2). Further, note that
the slip is a thick walled shell, and the solutions borrowed while the Reinhold-Spiri equations describe the essential
from thin walled shell theory are not applicable (apart from physics, they do not capture the distribution of stress in the
being cumbersome). It is therefore proposed that an slip contact area.
approximate solution be used, by assuming that the element of To calculate the axial force profile, assume that the axial
slip shown in Figure A-2 can be treated as a thick walled force distribution is represented by a quadratic polynomial,
cylinder in plane strain6. That this approach yields a as follows:
conservative (or upper bound) solution on the radial stress N1 (z ) = −(H 1 z + H 2 z 2 + H 3 ) (A-7)
distribution at the drill pipe-slip interface can be shown.
where H1, H2 and H3 are unknown constants to be determined.
Therefore, referring to Figure A-3, note that the problem is
A higher order polynomial can be assumed for the force
now governed by the Lame equations for a thick walled
profile. However, the number of equations that can be used to
cylinder. By knowing the radial stress on the outer boundary
determine the unknown coefficients is limited. The present
of the element of the slip in Figure A-3, the pressure on the
approach is frequently used in the theory of beam columns,
inner boundary can be determined. For present purposes, it is
where an algebraic polynomial is assumed to represent the
assumed that the bowl is infinitely stiff and zero radial
shape of the deflected beam. Since displacements/strains are
displacements at the OD of the slip are prescribed. The Lame
proportional to forces/stresses in the elastic range, the use of
equations show that the radial displacement at the outer
Eq. (A-7) is justified. Since the force profile in the drill pipe
boundary, u(b), of a thick walled cylinder (in plane stress)
should satisfy the boundary conditions at the toe and top of the
with inner radius a and outer radius b is given by
slip, Eqs. (A-2a) and (A-2b) provide two equations. The third
equation to determine the unknown constants is obtained by
1  a2
u (b ) = ( p − p )(1 + v ) +
pi a 2 − p o b 2
(1 −ν )b ensuring that the axial elongations of the drill pipe and slip are
 2
E b − a 2 i o
(b − a )
2 2
 equal, i.e.,
where pi and po are the internal and external pressures, and E L
N1 (z ) L
N 2 (z )
and ν denote the elastic modulus and Poisson’s ratio ∫z=0 A1 E1 dz = ∫z=0 A2 (z )E2 dz . (A-8)
respectively. By substituting, σw(z) for pi, σR,r(z) for po, do/2 A1 and A2(z) denote the cross sectional area of the drill pipe
for a and ds(z)/2 for b in the above equation and setting and the slip respectively. Also note that Eq. (A-8) has its
u(b)=0, the radial stress at the slip-drill pipe interface is theoretical basis in Castigliano’s theorem. It can be derived by
obtained as calculating the complementary energy stored in the drill pipe
σ w (z ) = G (z )σ R ,r (z ) (A-5a) and slip, and setting its derivative with respect to either N1(z)
where or N2(z) equal to zero, since the drill pipe and slip are at rest
1 + ν  d s (z )  1 −ν with respect to each other. This special case of Castigliano’s
2

G (z ) = +  (A-5b) theorem is sometimes known as Menabrea’s theorem8. Finally,


2  do  2 in Eq. (A-8), the drill pipe and slip areas are given by
If the distribution of the axial force in the drill pipe N1(z) is
known, the radial and axial stresses in the drill pipe in the slip A1 =
π 2
4
(
do − di
2
) (A-9a)
contact region can be found. Before determining the axial
force distribution, the validity of the solution obtained is
π
4
[
A2 (z ) = d s (z ) − d o
2
]
2
(A-9b)
verified by estimating the vertical component of the forces
acting on the lateral surface of the slip, i.e., the vertical where ds(z) is the OD of the slip which is calculated from the
component RV of the reaction from the bowl, slip geometry as shown in Figure A-1. When the integrations
indicated in Eq. (A-8) are performed, three linear equations
RV = ∫ σ R (z ).∆S .(µ cos α + sin α )
BC
in H1, H2 and H3 are obtained. The solution of these
equations yields,
∂N1 (z )
L

=−∫ dz . (A-6a) 1 
z =0 ∂z H 1 (z ) = Fa  − θL  (A-10a)
L 
= N1 (z = 0) − N1 (z = L )
H 2 = θFa (A-10b)
By virtue of Eqs. (A-2), the right hand side of the above
equation equals Fa, thus satisfying force equilibrium. The H = − Fa
3
(A-10c)
horizontal component of the bowl reaction is the transverse where θ is defined by Eq. (5d) in the body of the text.
force W, calculated by Reinhold and Spiri. It can be verified
IADC/SPE 74488 ADVANCED SLIP CRUSHING CONSIDERATIONS FOR DEEPWATER DRILLING 11

# Tubular Comments
1 5 ½ in., 24.7 ppf Grade unknown
2 5 in. 34.7 ppf Grade unknown
3 5 ½ in., 21.9 ppf Grade unknown
4 5 in. dia . mandrel Grade unknown
5 5 in. dia mandrel Grade unknown

Table 1 Test Matrix for tests by Organization B


0
(A proprietary slip with slip angle 9 27’ 45” was used. Slip
contact length is 16 in.)

Figure 1 Results from Test #2 of Table 1

Tubular Load Range, klbf


1 5 in., 38.01 ppf, S-135 950 - 1400
2 5 ½ in., 34.01 ppf, S-135 950-1350
3 6 5/8 in. 32.67 ppf, S-135 900 - 1150

Table 2 Test Matrix for Tests by Organization C


0
(Extended length slip with slip angle 9 27’ 45” was used. Slip
contact length is 16 in.)

Yield Initiation Yield Initiation


Drill Pipe OD Load1 Load2
(in) (kips) (kips)
5.0 1058 731 - 760
5.5 1157 851 - 860
6.625 1000? 790 - 830

Table 3 Summary Results for test Matrix in Table 2 Figure 2 Hoop strain at location 1 in mid-slip region, Test #3 of
1 Table 1
Yield Initiation at bottom location based on Figures 10-12
2
Yield initiation load based on Figures 13-15.
?. Data does not provide clear indication of yield initiation load

klbf
Friction Factor Test 1 Test 2 Test 3
0 817 872 679
0.1 988 1070 859
0.2 1083 1183 966
0.3 1141 1253 1035

Table 4 Predicted Slip Crushing Loads for specimens in Table 2


based on Reinhold-Spiri formula, Eq. (4)

klbf
Friction Factor Test 1 Test 2 Test 3
0 421 502 428
0.1 597 704 601
0.2 730 853 729
0.3 830 963 824

Table 5 Predicted Slip Crushing Loads for specimens in Table 2 Figure 3 Hoop strain at location 2 in mid-slip region, Test #3 of
based on Eq.(5a) Table 1
12 SATHUVALLI, PAYNE, SURYANARAYANA, SHEPARD IADC/SPE 74488

Bottom Displacement While Loading

1200

1000

800

Load (kips)
600 Sensor 1
Sensor 2
400 Sensor 3

200

0
-5.00 0.00 5.00 10.00 15.00 20.00
-200
Displacement (mils)

Figure 7 Radial Displacement Measurements from Three


Displacement Sensors 120º Apart, at Bottom Location

5.5" Drill Pipe Load-Displacement Results


Figure 4 Schematic of test set up for matrix in Table 2
1500

1200 1400
Target Load = 950 kips

Axial Load (kips)


Target Load = 1000 kips
Target Load = 1050 kips 1300
1000

1200
Axial Load, Klbf (kips)

800
1100

600 1000

900
400
8 12 16 20
Displacement (mils)
200
Bottom Sensor 1 Top Sensor 1

Figure 8 Results for 5.5 in Drill Pipe Tests – Load vs Top and
0
10:27 10:30 10:33 10:36 10:39 10:42 10:45 10:48 10:50 10:53
Bottom Displacement from Single Sensor
Time, hh:mm

Figure 5 Typical Loading Sequence During Tests described in 5.5" Drill Pipe Load-Displacement Results
Table 2
1500
Loading and Unloading Cycle Illustration
1400
1200
Axial Load (kips)

1300
Target Load
1000
1200
800
Load (kips)

Loading 1100
600
1000
400
Unloading
900
200 8 12 16 20
Displacement (mils)
0
Bottom Sensors Average Top Sensors Average
0.00 2.00 4.00 6.00 8.00 10.00 12.00
Displacement (mils) Figure 9 – Results for 5.5” Drill Pipe Tests- Load vs AverageTop
and Bottom Radial Displacements – Non Normalized
Sensor 1, Bottom

Figure 6 Typical loading and unloading sequence illustrating


hysteresis
IADC/SPE 74488 ADVANCED SLIP CRUSHING CONSIDERATIONS FOR DEEPWATER DRILLING 13

1400
5.5" Drill Pipe Load-Displacement Results

1500 1200

1400
1000

Load (kips)
Axial Load (kips)

1300
800

1200
600
1100 5" test to 1350 kips
400 5.5" test to 1350 kips
1000 6 5/8" test to 1150 kips

200
900
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
0.00 0.20 0.40 0.60 0.80 1.00 1.20
Normalized Displacement at Bottom Location
Normalized Displacement

Bottom Sensors Average Top Sensors Average


Figure 13 Normalized displacement while loading – all three sizes
Figure 10 – Results for 5.5” Drill Pipe Tests- Load vs Normalized
Top and Bottom Average Radial Displacements
1400
1300
5" DP Load-Displacement Results 1200
1100
1500 1000
900
1400

Load (kips)
800
1300 700
600
Load (kips)

1200
500
5-in test to 1000 kips
1100 400
5-in test to 1200 kips
300 5-in test to 1350 kips
1000 200
100
900
0
800 0 0.5 1 1.5 2 2.5 3 3.5 4
0.00 0.20 0.40 0.60 0.80 1.00 1.20 1.40
Normalized Displacement at Bottom Location
Normalized Displacement

Bottom Sensor Average Top Sensor Average


Figure 14 Load-Displacement while loading, 5” drill pipe, different
Figure 11 – Results for 5” Drill Pipe Tests- Load vs Normalized load cycles
Top and Bottom Average Radial Displacements
1400
6 5/8" Drill Pipe Load-Displacement Results
1200
1200
1000
1150
Load (kips)

1100 800
Axial Load (kips)

1050
600
1000

950 400 5.5" test to 1200 kips


5.5" test to 1250 kips
900
200 5.5" test to 1350 kips
850

800 0
0 0.2 0.4 0.6 0.8 1 1.2 0 2 4 6 8 10
Normalized Displacement Normalized Displacement at Bottom Location

Bottom Sensors Average Top Sensors Average

Figure 12 – Results for 6 5/8” Drill Pipe Tests- Load vs Normalized Figure 15 Load-displacement while loading, 5.5” pipe, different
Top and Bottom Average Radial Displacements load cycles
14 SATHUVALLI, PAYNE, SURYANARAYANA, SHEPARD IADC/SPE 74488

Axial Load

D
B
C

O
Deflection

Figure 16 Typical Load -deflection curves for (uniaxial) tension Figure 19 Stresses at the drill pipe ID
loading (OABD) versus (biaxial) loading in slips (OAC)

Figure 17 Suggested distribution of force and stress in slip


1
contact area (after Reinhold and Spiri )
Figure 20 Typical Slip Geometry

d st
d s (z)

C D
N 1 (z) N 2 (z)
µR
Z Z' L

α R
z

B A 0
di r

do
d sb
Slip OD:
Figure 18 Stresses at the slip drill pipe interface d s (z)=d sb +2z tanα
Fa

Figure A-1 Slip and Drill Pipe Interaction


IADC/SPE 74488 ADVANCED SLIP CRUSHING CONSIDERATIONS FOR DEEPWATER DRILLING 15

z
N 2 (z+ ∆ z)

d s (z+ ∆ z)

z+ ∆ z τ (z)= µσ R (z) ∆ S
C'
α
σ w (z)
σ R (z)
z B'

do N 2 (z)

d s (z)
∆ S= π d s (z). ∆ z.sec α
r

Figure A-2 Free body diagram for an axisymmetric slic of the slip

Slip OD, d s (z )

Drill Pipe OD, d o

σ w (z )
σ R ,r ( z )

Figure A-3 Top view of the slice of slip shown in Figure A-2

S-ar putea să vă placă și