Vous êtes sur la page 1sur 14

IADC/SPE 27470

An Expert System for Drill Bit Selection


M.J. Fear, * BP Exploration; N.C. Meany, * * Hughes Christensen Co.; and J.M. Evans,
Data Sciences UK Ltd.
'IADC and SPE Member
"IADC Member

Copyright 1994, IADC/SPE Drilling Conference.

This paper was prepared for presentation at the 1994 IADC/SPE Drilling Conference heid in Dallas, Texas, 15-18 February 1994.

This paper was selectad for presentation by an IADC/SPE Program Committee following reviaw of information contained in an abstract submitted by the author(s). Contents of the paper,
as presented, have not been reviewed by the Society of Petroleum Engineers or the International Association of Drilling Contractors and are SUbject to correction by the author(sj. The
material, as presented, does not necessarily reflect any position of the IADC or SPE, their officers, or members. Papers presented at IADC/SPE meetings are subject to publication
raview by Editorial Committees of the IADC and SPE. Permission to copy is restricted to an abstract of not more than 300 words. Illustrations may not be copied. The abstract should
contain conspicuous acknowledgment of where and by whom the paper is presented. Write Librarian, SPE, P.O. Box 833836, Richardson, TX 75083-3836, U.S.A. Telex, 163245 SPEUT.

ABSTRACT cutter shape.


The sensitiVity of bit run performance to bit design
An expert system has been developed to assist with implies that good bit selection is crucial if bit-related
the selection of poly-crystalline diamond compact (PDC) costs are to be minimised. Selection of a bit design
drill bits. The system uses a knowledge base of bit which is inappropriate for the conditions encountered
selection rules to produce a generic description of the leads to low rate of penetration and/or short bit life, in
most suitable bit for a particular set of drilling and turn leading to excess time spent rotating and tripping.
geologic conditions. The rules are derived from rigorous As an example, consider the rate of penetration
study of foot-based drilling, geology and rock property achieved on the PDC bit runs listed in table 1. Two bit
data from past bit runs. Compared to conventional bit styles were used, though both have identical IADC codes
selection methods this approach makes more effective for hydraulic design and cutter size/density (the
use of past experience, leading to a higher and more characteristics in the IADC classification expected to be
uniform level of bit selection expertise in drilling most influential to penetration rate). Despite their
operations. apparent similarity, penetration rate from bit type 2 was
so poor compared to type 1 that the bit was pulled after
INTRODUCTION drilling only 41 feet. After careful scrutiny of foot-based
mud logging and wireline rock property data, the low
A worldwide review of BPX wells identified that 10- ROP was found to be an effect of cutter shape; bit type 2
40% of the dryhole cost of a well is bit related. On one had hemispherical PDC cutters which could not be
particular bit run, these costs are composed of the time- indented into the relatively hard rock with available
based operating cost of the drilling facility, the duration weight on bit. One and a half days of rig time was lost as
required to drill the bit run interval (Le. rate of a result of the low ROP and consequent trip, compared to
penetration), trip frequency, determined from the length drilling the interval with the first bit type.
of the interval over which the bit can be kept drilling (bit This example illustrates that subtle differences in bit
life), and the price of the bit itself. This means that design can be highly influential to drilling performance
drilling cost is a strong function of the rate of penetration and thus to well cost. In this case, the ROP reduction
(ROP) and life provided by the bit. This in turn means due to minor variation in the geometry of the cutter tip
that the suitability of a bit design for drilling in a given (backrake angle) led to a wholly uneconomic bit run.
interval is a major factor in bit run cost. On a PDC bit for A systematic approach to bit selection, which
example, bit life can vary dramatically with subtle identifies how a bit is expected to perform based on its
changes in cutter distribution or the type of gauge design characteristics and predicted geological and
protection materials used, while rate of penetration may drilling conditions, is a necessary step in optimisation of
vary substantially with changes in nozzle placement or bit performance. This paper proposes a means by which
this may be achieved.

References and figures at end of paper

359
2 AN EXPERT SYSTEM FOR DRILL BIT SELECTION IADC/SPE 27470

THE CURRENT APPROACH TO BIT SELECTION It is clear from simple drilling economics that the
incentive to develop and implement such a process
Given the obvious incentive to select the most depends upon the commercial return from improved
suitable bit design, tools should exist to assist drilling penetration rate and bit run duration. This commercial
engineers in bit selection. In fact, conventional bit incentive will be highest in regions where overall
selection methods typically lack access either to operating costs are high, and prevailing penetration rate
recorded bit selection expertise, to methods to identify and/or bit life is low. It is worthy of note however, that
appropriate bit characteristics for the expected drilling previous studies have shown that, on a worldwide
conditions, or to methods to predict performance for basis, the majority of drilling time is spent drilling at low
various bit designs. The standard of bit selection is penetration rate (1), so that the commercial incentive to
thus heavily dependent upon the expertise of the improve bit performance is both strong and
individual engineer, rather than being continually geographically widespread.
improved through a process of organisational learning. In response to this commercial incentive, an
There are weaknesses for example in the way improved bit selection process is being developed. The
historical data is used to review past bit performance. bit selection technique at the heart of this process is the
Table 2 shows a suite of data typically employed for main subject of this paper.
this purpose. The information is a bit record, where
details of the bits used and runs made are reduced to FORM OF THE BIT SELECTION GUIDELINES
one line of data per bit run. This simple report has the
advantage of being quick to retrieve and review, and is, The previous section highlighted the need to record
with minor variation in format and content, an industry bit selection and optimisation experience from past bit
standard drilling report. However, the brevity of the runs, and to transfer this expertise to operational
data mean that much valuable information on bit communities. Guidelines are needed that can be easily
performance and bit selection is missed if more updated with evolving knowledge, and that can deliver
comprehensive data is not utilised. advice based on the user's particular drilling conditions.
These limitations can be shown by reviewing table Industry feedback suggested that, in situations
2, bit run number 11. It is apparent from the bit record where capture and dissemination of knowledge is
that this was an unsuccessful run, where both ROP and required, and where the efficiency of decision-making
footage drilled were low, and the (expensive) bit fatally has significant cost implications, expert systems have
damaged by the run end. The weakness of the bit proven successful. Other advantages of an expert
record however is that the reasons for the poor system for this application included the capability to add
performance are not at all apparent. In contrast, figure on automated use of offset data or bit selection
1 shows a suite of foot-based mud logging and wireline methods based on risk analysis (2), and the tailoring of
data for the problem bit run. Annotated on the log are bit selection recommendations to the user's own drilling
the events which, from patterns in the drilling parameter and geologic conditions. A low cost prototype expert
relationships and their relationship to geology, explain system was therefore developed, and successfully
both the low penetration rate and severe bit damage. demonstrated methods by which bit selection expertise
In this case, the interpretation that bit balling caused the could be converted into "rules", which could be used to
low ROP is as expected from the bit design, which was make objective bit selection recommendations. The
heavily set with medium size PDC cutters in an attempt prototype was also used to assess receptiveness of the
to provide adequate bit life in the sands. The heavy intended operational users to a software-assisted bit
cutter density limited space for waterways on the bit selection process. Support was strong.
face, in turn limiting bit cleaning capacity. A decision was therefore made to develop an
These observations have great value for future bit operational expert system for bit selection. The
selection under comparible conditions, providing that decision was also taken to involve a major drill bit
guidelines exist into which this latest experience can be supplier in the development and implementation of the
inserted. Such a "knowledge base" is however rare in system. Such a joint development was expected to
bit selection, meaning that it is impossible to enhance the expertise contained within the system, and
demonstrate that bit selection decisions are made using to encourage other suppliers to develop a more
the best expertise and knowledge available in the systematic approach to bit selection.
organisation at that time.
A bit selection process is therefore required which OBJECTIVES AND GENERAL DESIGN OF THE
is more befitting of the associated drilling costs. In EXPERT SYSTEM
particular, such a process should provide (a) routine
use of foot based data, so that observations of use in At the outset, a number of functional objectives
bit selection and performance optimisation are were set for the expert system. First, the system would
identified and not missed, and (b) bit selection and initially target the selection of one class of bit only, to
optimisation guidelines, so that expertise can be minimise development cost and risk. PDC bits were
recorded, updated, and periodically issued to chosen, since they are a source of ROP improvement
operational staff (in essence, a system to facilitate over other classes of bit in many applications. Second,
organisational learning and performance improvement). bit selection recommendations from the system were to
relate to the user's specific drilling and geologic

360
IADC/SPE 27470 FEAR, M.J., MEANY, N.C., EVANS, J.M. 3

conditions, rather than to generalised geological


scenarios. Third, system recommendations would The knowledge base has been constructed by first
define the acceptability of individual bit features for identifying which bit features, and facets of the
various drilling and geologic conditions, and would build operating conditions, are to be used to construct rules.
these into a generic description of a suitable bit. A constraint here is that using too many of these bit and
Classifying bits by the IADC classification was not used drilling conditions "attributes" will result in a very large
because this classification scheme is not based on and unwieldy rule base. In the current system,
expected performance, it is more descriptive of attributes used are restricted to those most influential to
materials used and bit geometry. Consideration of real PDC bit performance.
bit designs was also excluded, due to the speed with An example of where a rule requirement exists is
which PDC bit product lines change, and the the effect of open face (or "junk slot") volume on a PDC
consequent difficulty in updating the system. Users bit, when hydratable rock is being drilled with water
also preferred the educational value of a generic bit base mud. The open face volume influences the ease
description. Fourth, geologic information would be with which drilled material can be evacuated from
described by lithology· and wireline log measurements around the bit, and is a strong influence on ROP in
rather than local stratigraphy, so that bit feature sticky formations. The system therefore requires a rule
recommendations could be derived from, and applied, to stipulate appropriate values of open face volume for
in any geographic area where the relevant geologic a PDC bit under these conditions.
criteria were met. Fifth, the system would not attempt Similarly, certain facets of drilling and geologic
cost-based bit selection methods (e.g. by risk analysis), conditions may affect the suitability of a particular bit
since reliable prediction of penetration rate and run design. Carbide gauge protection may for example be
duration for a planned PDC bit run is not yet feasible. perfectly adequate when a sand section is to be drilled
Sixth, since individual bit features were to be related to with a rotary bottomhole assembly (BHA) at low
facets of drilling conditions, a large rule base was inclination, whereas a more abrasion-resistant diamond
foreseen. This implied use of a formal rule generation guage may give better performance if the same interval
and organisation process to ensure manageability of were to be drilled with a motor BHA in an inclined
the rule base. Last, the system was to be developed section. In this case, a rule would be required to relate
using hardware and software compatible with the preferred gauge protection to lithology, BHA type, and
drilling database "DEAP". DEAP is a platform for inclination.
drilling engineering applications in drilling operations Once the rule requirements have been defined,
(3). This stipulated that the system be developed on a individual rules can be constructed. One such rule is
Unix platform, in an X windows/Motif environment. shown in figure 3. This rule shows required values of
In summary, the objective of the expert system is to bit open face volume in formations where bit balling is a
identify appropriate bit characteristics for use in a risk (here simplified to "shale"), and when the drilling
particular set of geologic and drilling conditions. The fluid is water based. Note that the wireline resistivity
output would be a generic description of a suitable bit measurement is used as a simple indicator of the
for the interval to be drilled, which could then be degree of compaction of a shale, and by implication, its
compared to available products at the procurement tendency to exacerbate bit cleaning problems.
stage, or could be used to initiate design and This rule highlights a number of properties of the
manufacture of appropriate products by bit designers. expert system rules in general. First, the rules are
The expert system is therefore aimed at two main graphical, with the values of the bit feature as the y-
categories of user; those involved in rule development axis, and the operating conditions attribute as the x-
and maintenance of the knowledge base, and those axis. This prOVides a smooth relationship between the
using the system to generate bit selection or design bit and operating conditions attributes, which is
recommendations. This is reflected in the structure of essential when any value of the operating conditions
the system, which is divided into two modules; the rule attribute can be chosen by an operational user.
management facility, and the bit advisor module. Second, each bit feature is defined in terms of its
These modUles, and their component parts, are shown acceptability. This is divided into three categories;
schematically in figure 2. preferred, showing the range of values in the bit feature
In the rule management facility, functions exist to expected to give best bit performance; acceptable,
define those bit features and facets of the drilling and showing compromise bit feature values that may need
geologic conditions for which rules are required, to to be used because of other considerations; and
construct the rules, and to manage and update the rule rejected, showing bit feature values that would give
sets. In the bit advisor module, facilities exist to enter unacceptable performance.
data to describe expected drilling and geologic As part of the rule generation process, the context
conditions for the interval to be drilled, to generate within which each rule is to be applied must be defined.
assessments of suitable PDC bit features for use under In other words, the system must recognise from the
those conditions, and to review the rules responsible for expected drilling and geologic conditions, which rules
the recommendations. are to be applied for which intervals. In addition, this
process must recognise that while each rule relates one
THE KNOWLEDGE BASE AND RULE MANAGEMENT bit feature to one property of the downhole conditions,
FACILITY other conditions may negate the rule or change its

361
4 AN EXPERT SYSTEM FOR DRILL BIT SELECTION IADC/SPE 27470

form. The rule shown in figure 3 for example would be


different if the mud type was oil based. In such a case, In addition to bit feature recommendations, the
the mud prevents bit balling, leaving greater freedom in system contains a facility to make other
acceptable bit designs. recommendations that may aid bit performance. Such
Defining the context within which particular rules recommendations may be relevant only to certain types
are to be used has been handled in the system by use of drilling and geologic conditions, but may not be
of a "rule tree". This is shown in figure 4. The rule tree suitable for encoding into rules. Examples could
defines the scenario within which each rule is to be include advice to run surface drive system control
applied in terms of lithology and mud type. These two equipment to suppress torsional slip-stick vibration in
attributes were chosen since almost every rule has hard formations, or information on the most appropriate
some dependence on mud and/or rock type. One rule drilling parameter combinations for certain formations
tree is used for each combination of bit feature and or bits. This information is termed "operating practices"
attribute of the drilling/geological conditions. in the expert system.
The rule tree is used as follows; when the need for Activation of operating practices information is
a rule has been identified, the rule is constructed as triggered with a similar mechanism to that of the rules.
shown in figure 3. It is then placed in a blank rule tree, In other words, each item of operating practices
and linked to the mud and rock types within which the information is linked to facets of the drilling and
rule is valid. The rule tree then prompts for all other geologic conditions, and display of the information is
mud and rock "branches" to be terminated with made if the same conditions are recognised in the
variations on the rule, so that all possible mud/rock interval to be drilled.
combinations available to operational users will give
information on the selection of that bit feature. The rule GENERATING RULES FROM BIT RUN DATA
tree therefore helps to ensure that there are no gaps in
the rule base; a critical consideration for reliable system The expert system requires that objective bit
operation. selection rules be generated from observations of past
Between the individual rules and the lithology bit performance. Use of foot-based data is preferred for
section of the rule tree, further criteria to control the this purpose, so that observations of bit design effects
application of rules can be defined. These are termed are clear enough to be formulated into expert system
"activation criteria", and test another property of the rules. Integrated data sets of mud logging, wireline,
downhole conditions or bit features before applying a geology, survey and bit data are normally used
rule. This feature is especially useful when a suitable wherever available. Bit record data is inadequate for
range of a bit feature is dependent upon another bit this work.
feature. For example, the preferred number of cutters The main requirement from analysis of field data is
required to drill a particular interval will depend in part to isolate which features of a bit are responsible for
upon the predominant cutter diameter on the bit. The varying performance. For instance, from the data in
activation criteria order and handle this process. table 1, knOWing only that bit type 2 drills at low ROP in
In the current system, rule sets are split up by hole these formations is of little value to the expert system
size. This is because many of the values of knowledge base. What is important, and what really
acceptability for bit features are valid only for one hole improves future bit selection, is a knowledge of what it
size. For each hole size, the available rule set can be was about that bit that made it drill slowly. That
viewed through another rule management tool, the rule particular feature can then be avoided under similar
requirement matrix. This is shown in figure 5. The rule conditions in future, while other, perhaps beneficial,
requirement matrix shows which bit features and facets features of the same bit can continue to be safely
of the operating conditions are linked by rules, and exploited. Thus, the generic description of a suitable bit
whether the rules are currently enabled (individual rules can be built up.
can be disabled for troubleshooting purposes). One example of how field data can be analysed to
Properties of the downhole conditions can also be determine bit feature effects on performance relates to
deemed mandatory or optional, the former meaning the rule shown in figure 3. In this case, the rule
that the system will not run the rule base for a planned requires values of open face volume that have been
interval of drilling unless this information is provided. shown, by correlation to ROP in water based mud/shale
The rule requirement matrix also indicates whether applications, to give varying efficiency of bit cleaning.
rules need to be updated after changes to the To derive this rule, averages of foot-based (mUd
definitions of bit features or properties of the downhole logging) data were used to quantify the relationship of
conditions. ROP to other drilling variables. For a group of PDC bit
Facilities exist within the rule management module runs in water based mUd, bit hydraulic horsepower and
to expand the system to cater for other classes of bit bottomhole pressure were found by regression analysis
(roller cone, natural diamond etc). Appropriate to be the dominant effects on ROP. A simple empirical
flexibility exists in definition of new bit features or model developed to relate these three variables gave
properties of downhole conditions should these need to an r-squared value of 0.881, from 13 samples; Use of
be changed for other bit classes. this model to normalise ROP for these drilling
variables, within one lithology, duly revealed the bit
OPERATING PRACTICES INFORMATION design effect on ROP. This is shown in figure 6, as a

362
IADC/SPE 27470 FEAR, M.J., MEANY, N.C., EVANS, J.M. 5

plot of the ratio of actual to model ROP, versus the each item of data applies is known. For example, a
open face volume of the six bit designs in the data set. rock property would be deemed an attribute of a major
Incorporating this bit property into the ROP analysis or minor lithologic layer. This logic is used to divide the
improved the fit of the model; r-squared rose to 0.934. interval to be drilled into "zones", within which the rules
From this correlation between normalised ROP and can work to provide one recommended set of bit
the bit feature, a rule can be generated to stipulate (for features.
example) no less than 200 cubic inches of open face Once the interval to be drilled, has been described
volume, in order to maximise ROP. Other rules would and the zones defined, the rule base is run. A
temper this by stipulating more durability if more description of suitable bit features will be provided for
arduous fomations were to be drilled in conjunction with each zone within the interval. All bit features present
the shale, so that a greater cutter count would reduce on the vertical axis of the rule requirement matrix for
available open face volume. that hole size (fig. 5), and for which rules exist, will be
In this exercise, it is critical to consider properties of covered by the recommendations. The mechanism of
the bit that can feasibly affect performance. In the rule processing is described in the next section.
above example, open face volume was preferred over The rule recommendations are presented adjacent
junk slot area since the latter is less useful in terms of to the hole section details, and list the bit features most
expected performance. It is again worthy of note that suitable for use in each zone interval. The combination
the IADC classification for fixed cutter bits does not of these features provides the generic description of the
consider many bit features that are influential to suitable bit, in each zone. An example of the bit
performance. feature recommendations by zone is shown in figure 7.
Other rules in the system have, where possible, On the output screen, these are aligned to correspond
been developed with similar objectivity. For example, to the interval over which each generic bit description
the durability of different styles of PDC bit in highly applies. Flexibility exists to zoom in on particular
abrasive applications (e.g. motor drilling of hard sands) intervals for full display; otherwise a chosen sub-set of
has been found to depend heavily on the shoulder critical bit features will be displayed next to the hole
profile of the bit (cutter density), and cuttings removal section details. Figure 8 is an example of a display
capacity at the outer part of the bit (a function of nozzle combining hole section details and bit features, and
diversity and applied hydraulic energy). Rules showing the zone structure. Note how the sub-division
therefore exist to make stipulations on bit profile, cutter of major lithologic layers into zones is made on the
count, and nozzle count, in such applications. basis of changes in other important variables such as
Some examples of additional rules in the system mud type, BHA type or well profile.
include those to relate bit cutter count to lithology, At this point, some refinement to the bit
hardness and thickness, bit body material (erosion recommendations is required. It is obvious that the hole
resistance) to applied hydraulic energy and lithology, section will not be drilled with bit runs corresponding in
gauge protection materials to BHA type, inclination and depth to the zones defined by the expert system. This
lithology, and gauge length to required well profile reflects the common practice of selecting compromise
(steerability). bit designs to drill a number of intervals, thus
minimising trip time for bit changes. The expert system
GENERATING BIT RECOMMENDATIONS FROM THE caters for this by allowing definition of expected bit run
SYSTEM intervals, based on a comparison of generic bit
descriptions from the zones. In other words,
The bit advisor module is the part of the system commonality in bit descriptions between adjacent zones
where bit selection and optimisation recommendations means that some zones can be merged and a
are generated, based on a description of expected compromise bit sought.
geology and drilling conditions. Defining target bit runs in this way then triggers the
The operational user or bit designer is first required rule base to be re-run. The same processing logic as
to provide a suite of information on the interval to be applied to the zones now produces a generic bit
drilled (usually one hole section of constant size). This description for each (extended) bit run interval. An
"operating conditions" data includes descriptions of example of a partial set of recommendations for bit run
each major lithologic layer, minor lithologies, average intervals is shown in figure 9. Note that the bit run
rock properties for each lithology (using wireline intervals encompass a number of zones.
measurements), drilling fluid type and density, well Accompanying the bit feature recommendations is
profile information by depth (inclination, azimuth), the operating practices information that is designed to
expected bottomhole assembly types by depth (Le. aid bit performance once the bit has been selected.
motor type, or rotary), and other drilling variables. Figure 10 shows a window of operating practices
Some variables are then calculated from these data, information, for one zone. In this case, advice is
including expressions of required dog-legs, and provided on how to combat PDC cutter breakage and
overbalance pressure. Other variables exist to record vibration, both being anticipated in this zone.
the expected incidence of vibration or mechanically Following definition of the programme of generic
unstable formations. bits, the operational user can then compare each with
The system includes logic to sort these data for available products, and recommend the products with
processing by rules, so that the depth range over which closest resemblance. Alternatively, the bit designer can

363
6 AN EXPERT SYSTEM FOR DRILL BIT SELECTION IADC/SPE 27470

use the suite of features described as suitable by the that operational community in water based mud. The
system to initiate design of a new product. importance of an open-faced design for bit Cleaning,
and the differences in bit requirements between oil and
REVIEWING SYSTEM RECOMMENDATIONS water base mud, were not appreciated. Also, the
supplier did not have sufficiently close involvement to
To develop the confidence of operational users in prevent the run being made.
the expert system, and to aid trouble-shooting, facilities Under the same conditions as this unsuccessful
have been included to review system run, the expert system stipulates an open face volume
recommendations. These rule review facilities also of 602-885 cubic inches (which requires a bladed
show how the logic from many rules combines to design), with 33-70 3/4" PDC cutters. A bit meeting this
deliver a bit feature recommendation for one zone or bit generic description made a run in a similar geologic
run. interval in another well, the performance of which is
From the rule requirement matrix shown in figure 5, shown in table 3 (run 3). Actual design data for this bit
individual rules used can be identified and examined. type is shown in table 4, together with the expert
The rule shown in figure 3 has the value of the rock system stipulations. While not equalling the
property from the zone indicated. Where the marker performance of the run in oil base mud, performance is
crosses the preferred and acceptable regions defines significantly better than with bit type 1 in water based
the acceptable and preferred range of this bit feature, mud.
based on the one facet of the drilling and geologic The critical point here is that, because bit run
conditions only. experience existed elsewhere which made the poor
The combined effect of a group of rules produces performance of run 2 predictable, a better bit selection
the final feature recommendation for a zone or bit run, could have been made if that experience, and
after processing for all items of drilling and geologic guidelines developed from it, were available to this
data within the interval. This can be reviewed by use of particular group of operational users at that time. The
the tool shown in figure 11. In this example, the final expert system is the means by which this is made
recommendation for 40-44 cutters is seen to be the possible. It appears idealistic to expect the best bit
result of six rules, the sequential application of which selections always to be made without the assistance of
has progressively narrowed down the preferred number a guidelines mechanism. This is not surprising given
of cutters to this range. This process is repeated for all the complexity of the biVrock interaction.
bit features within each zone or bit run, culminating in
the generic bit description for that interval. CONCLUSIONS

COMPARISON OF SYSTEM RECOMMENDATIONS 1. In many areas, a concerted effort at optimisation of


WITH CURRENT PRACTICE bit performance is justified by large potential
savings in well cost. This is particUlarly so where
Current bit selection practices rely heavily on the time-based costs of the drilling operation are high,
expertise of individuals to identify appropriate bit styles or where penetration rate and/or bit life are low.
for given applications. This process is not foolproof. 2. The current approach to bit selection cannot
For example, in large organisations, learning which has demonstrate best use of past experience, and
benefitted performance in one area may not be relies too heavily on data that is conveniently
promptly and effectively communicated to other areas. available rather than being fit for purpose. These
In addition, the standard of bit selection may be uneven deficiencies can be rectified by development of bit
due to varying experience between individuals. The selection guidelines drawn from past bit runs,
expert system aims to eradicate these weaknessses accompanied by rigorous analysis of foot-based
and provide a uniform level of bit selection expertise data.
corresponding to the highest available within an 3. Greater access to, and use of, foot-based data is
organisation at anyone time. an essential component in a more rational
Consider the contribution of the expert system in approach to bit selection. Improving access to
the following example.. Table 3, run 1 lists data from a these data, and developing the expertise to exploit
highly succesful bit run at 17.1/2". Table 4 lists it, will be rewarded by an improved understanding
pertinent design data for the bit (labelled as bit type 1). of how to optimise bit performance.
Note that the run was made in oil base mud, so that the 4. An expert system provides a suitable means to
presence of shale did not cause bit cleaning problems ensure that learning from past bit runs improves bit
despite the use of a face-set, non-bladed bit design. selection.
The same bit design was later run in water based 5. Bit selection guidelines, in a format suitable for use
mud, in a section of similar geology. The performance in an expert system knowledge base, can be
is shown in table 3, run 2. Performance was poor and derived from analysis of foot-based drilling and
the bit pulled after only 8 feet. The face of the non- geologic data.
bladed bit was unable to kept clear of the hydratable 6. Design and construction of an expert system for bit
drilled shale, causing the bit to ball up. selection requires a systematic approach to
The reason why this run was attempted was that identification of rule requirements, construction of
little experience with PDC bit selection existed within rUles, and definition of how rules will be applied.

364
IADC/SPE 27470 FEAR, M.J., MEANY, N.C., EVANS, J.M.

Rule management tools can be developed to


assist these steps.

ACKNOWLEDGEMENTS

The authors wish to thank BP Exploration, Hughes


Christensen Company, and Data Sciences UK Limited
for permission to publish this paper.

REFERENCES

1. Brett, J.F., Warren, T.M., and Behr, S. : "Bit Whirl:


A New Theory of PDC Bit Failure," SPE Drilling
Engineering (December 1990) 275-281.
2. Fear, M.J. et al. : "Optimisation of Rock-Bit Life
Based on Bearing Failure Criteria," SPE Drilling
Engineering (September 1992) 163-167.
3. Sawaryn, S.J., and Grayson, H.L. : "Industry
Participation in the Development of an Integrated
Drilling Engineering Applications Platform," paper
23892 presented at the 1992 IADC/SPE drilling
conference, New Orleans, Louisiana, February 18-
21.

365
Table I Comparison of penetration rate with two bit types

Bit ROP Weight Rotary Flow Hydr.H Wellbore Sonic Gamma Bit
type rt/hr on bit Speed rate Power pressure tr.time ray wear
KLBS RPM GPM HPIIn2 osi micslft api state
I 5.9 24 135 579 0.58 10101 58 63 Worn

2 2.5 24 126 573 0.56 10124 60 56 New

I 9.1 17 127 564 0.39 10101 59 65 New

Table 2 Industry standard bit report

BIT IADCI JETS DEPTH HRS ROP WOB ROT FLOW MW GRADING
RUN DIAM InlFeet TOT rt/hr klbs RPM l!pm PPl! IADC8PT
4 M614 9X13 1925 55.6 36.0 5 150 1300 9.6 2121BU/A
17.5 4XI2 2001 62.1 30 180 10.0 XlIlWTIPR
5 S612 3XI6 3926 27.8 22.8 10 150 1300 10.5 818/LT/A
17.5 2XI8 633 31.4 30 190 11.0 XlllROIPR
IX20
6 321 3X22 4559 2.1 4.8 20 130 1300 11.0 0/4IBTIH
17.5 10 4.8 45 180 2I1IJDIPR
7 135 2X22 4569 12.0 9.2 20 180 1340 11.0 313/WT/A
17.5 IX24 110 12.3 45 FJIlAlPR
8 M433 4XI5 4679 15.0 5.5 10 180 1340 11.0 111 INOlA
17.5 3X16 83 15.9 30 200 XIlINOIPR
9 S613 7XI5 4762 14,1 47.7 5 200 1340 11.0 II1INO/A
17.5 673 16.1 16 XIlINOIPR
10 S613 7XI5 5435 16.1 7.4 10 200 1340 11.0 8nIBT/A
17.5 119 21.5 35 X/I/WTIPR
II M433 2XI5 5554 51.8 5.5 10 180 1340 11.0 8/3IBTIN
17.5 5XI4 284 55.7 35 200 Xl2lWTIPR
12 435 2X22 5838 58.5 5.8 10 100 1120 11.0 2I2IWT/A
17.5 IX24 337 64.5 35 150 1330 EI1IOIFM
13 S232 2XI4 6175 44.4 16.5 5 140 1120 11.0 I/IINO/A
17.5 3XI5 732 52.6 20 160 XI1/NOITW
2X16

Table 3 PDC performance variation due to mud and bit type

BIT RUN BIT TYPE ROP (rt/hr) FOOTAGE MUD TYPE BIT WEAR STATE
1 I 20.3 2914 OBM 1-2
2 I 2.2 8 WBM 3-2
3 2 29.1 742 WBM 2-3

Table 4 Bit designs from field runs versus an expert system recommendation for WBM

BIT TYPE OPEN-FACE CUTTER FACE CUTTERS FACE CONFIG.


VOLUME DIAMETER
(cubic inches) (inches)
I (from table 3) 215 3/4 82 Face-set
Expert system 602-885 3/4 33-70 -
2 (from table 3) 677 3/4 64 Bladed

366
Wi rei in. Data Penetratian Rate Weight on bit Tarque
GAMMA RAY ROP SURFACE WOB AVER SURFACE TORQUE
150 I ;O::""_....J.(..:.'1'-.:":.:.rJ.)_-..:.;l0~0::'1.2 ,i"2,.l
° (CAP' )

ROTARY SPEED PREDICTED TQRQUE


4-.9I,O=-- {'"'K""'..;.I~. ';.:.;.:0...

MINIMUM SURFACE TORQUE


) .;.30.;.1

.2 .i.r,£".l. __ 15,9 (KIl.I.o) 30 ° 9.••••__._••••••••t~!_~d.·.!l. ._••••••••.:lQ.


MAXIMUM SURFACE tORQUE
.2 i.K.!.I ... 'lol .}Q.
t :.. ~-:.~ ...
) ..........
..................1 ~~ . . . . . . _
\
I ..~--.,.......
'--
,:' «;' TORQUE/ROP DROP
I ,:... ~ NO CHANGE IN
\ ~ GAMMA RAY & NO
(
I
WOB DROP -
I INDICATES BIT
I BALLI NG
(

)
\
,
I

,
I
1
)
\ TORQUE LOW
J COMPARED TO
I PREDICTED - ALSO
\ SHOWS LOW ROP IS
.• I DUE TO BIT
BALLING

HIGHLY ERRATIC
TORQUE IN SANDS;
SUGGESTS
VIBRATION CAUSED
THE BIT FAILURE

....-::.......
-~
;.---l...::.:;;;.;......_~ _

Figure 1. Interpretation of Bit Response from Foot-based Data

Rule Editor ConftgMngt

Rule Management

-- _ _ Fa~iIIt~ _
Drill Bit
Advisor
-

Figure 2 Basic System Design


367
Drill Bit Advisor: Recommendations for Junk Slot Volume from Shale Resistivity: Rule 3
Version: IPlan I PDC I
Facilit;:y~:I====., Bit Class:
VVeu:I~ ~ Hole Section:[ll;li] lin 01 Zone: UJ
In ut:
Ka .-. 1
.$
Preferred: Ittt~i~tl
Acceptable: II1II
--:l
u
CIoI
E
:l

~
Rejected: _
Input Value: _
-oS
rI:l
,.:,c
..,c:
:l

o 50 100
Shale Resistivity (OHM-M)

I( Explain... ) ( About... ) ( Close ) ( Help... ) I


Figure 3 Rule for Bit Open Face Volume

Rule Management Facility: Rule Tree Display


Bit Class: PDC Version: 1.1 Hole Size: 12.25 in Status: Experimental

..-
Mud Type Rock Type Available Bit HSI Rules
Ka iZ!
Preferred: t"",:""",:",,1 10 -> 10


·r..
Acceptable:
Rejected:
Enabled:
--
--- J OBM
'h !I lie

Ois!lbl..d,
- 0-> 10 Rule Not Applicable I

..
/1
·h.... "/11
I>"rl, 'II

I Junk Slot Volume I J 0 ->3

~V
vs. I

I Shale Resistivity I :h .. l ..
1"lit..
1\

( Redraw )
1 WBM I

~n ..

- H
' 3 -> 10

0->10
I-

• Rule Not Applicable I


( Xref.•• ) ~" ./

I( Apply )[ Undo )[ Reset... )[ Close... )[ Help... )


I ( Add-> ) ( Edit... ] ( Link ) ( Override) ( Delete ) ( Save... )

Figure 4 Complete Rule Tree for One Bit Feature

368
Rule Management FaciUty: Rule Requirements Matrix
Bit Class: PDC Version: 1.1 Status: Experimental
ENVIRONMENT ATTRIBUTES
Hole Size: 112•25 In 0 I
M M 0 0 M MIM M MIM M

Ku 1:1
_ Enabled .2... .... ...
lIIIIIIIl1IIIl

Disabled
Rule
f
:2 ~
'"~ ~ ""5t rI.l
:: ...!
1:1
~
...l!: .......
~ Activation Criterion > ~ ~ :: Q
Jl .!l! .l!l 1:1 ~ !. == U

. =..... .!
rI.l
" Has Bit Feature Q 1:1 fl .5l ~ Jl ~
:: ~ .lI!
~ ClIl
[] Activation Criteria
Modltled 1:1 .¥
1:1
Q
ClIl
1:1 ! <
~

~
..
~
.!l!
A~ ! 'iii
.c
rI.l .... I
y
1:1 :a ~
:: .I
~
...
<
1:1
Q
rI.l
BIT FEATURES ==
Bit Body Material
~r
Pred. Cutter Diameter ~
Pred. Cutter Type
~
~
•• • ~ ~

No. of Primary Cutters ""


Pred. Backrake Angle ~ III • •• • ~
~

Bit Face Volume


No. of Nozzles
Gauge Prot'n Mat'l ~

• •• II
~
Ai1 III
[4(

Gauge Pad Design


Gauge Pad Length • • ~
Bit Proflle
Cutter Back-up Features •••
..... • II ~ ~ ~ II
Back Reaming Cutters

I( • •I
Edit... ) ( Disable ) ( Copy ) ( Paste )( Delete)
II

I( Save... ) ( Undo ) ( Reset... ) [ Close... ) [ Help... ) I


Figure 5 Rule Requirement Matrix
c.. 1.4

.... •
0
a:
..J 1.2
LU
Cl 1.0
0
~

c..
0
0.8
0.6

a:
..J
<
::::>
0.4
0.2

I-
0
< 0
0 100 200 300 400
OPEN FACE VOLUME (cu. in.)

Figure 6 Bit Design Effect on ROP


369
Zone Bit Feature Units Preferred Acceptable
1 OK: v J
Bit Bodv Material· N/A Matrix Matrix. Steel
Pred. Cutter Diameter: Imm ~I IF 116,19 IF 113, 16, 19, 24
Pred. Cutter Type: N/A IFEI Standard IF·· 1Standard, Dome
No. of Primary Cutters: N/A 1B§130 -> SO 1~120->80
Pred. Backrake Anele: Idel! ~I 25,30 15,20,25,30
Bit Face Volume: Iin3 ~I ~30->SO ~20->80
No. of Nozzles: N/A mS->7 ~3-> 10
Gaul!e Prot. Materials: N/A ND Flush Set ND Flush Set. Carbide
Gaul!e Desil!D: N/A Standard Standard
Gauge Pad Length: '1m ~I 2 ->4 1-> 6
Bit Profile: N/A Short. Medium Short. MedIUm
Cutter Backup Features: N/A Hybrid Posts Hybrid Posts
Back Reamine Cutters: N/A No No
2 OK: v J
Bit Bodv Material: N/A Steel Matrix Steel
Pred. Cutter Diameter: Imm ~I ~16 1F¥113,16

Figure 7 Bit Feature Recommendations by Zone

TVDMD Lithology Geometry BRA Bit Run Zone Cal Bit Body Material Pred. Cutter Diameter
Key••• AZ Key•••
13
5 P Matrix 16, 19
A Matrix, Steel 9, 13, 16, 19,24
p Steel 19,24

J
A Steel 16,19,24
etc etc etc

! 2

Figure 8 Combined Hole Section and Bit Recommendations Display

370
TVDMD Lithology Geometry BHA Bit Run Zone Cat Bit Body Material Pred. Cutter Diameter
Key•••
p Matrix 19
A Matrix, Steel 16,19,24

) 1
2

3
4
5 P Steel 19,24
A Steel 16,19,24
6
7

2
8

Figure 9 Bit Recommendations by Bit Run

Drill Bit Advisor: Operating Practice Recommendations


Facility: I I Well: I I I
Version: Plan I
Hole Section: 112.2511 in 01 Bit Class: IPDC 1
Zone: 1ill:Lllf1
The Following Operating Practices are Recommended:
~ Practice
1 Cutter Size and Breakage
If cutter breakage (IADC Code BT, CT) has been seen on
offset bit runs in this formation, using smaller cutters
and/or higher backrake angle may reduce breakage, if
compatible with other rule recommendations.

56 Soft Torque System


A soft torque system should improve bit life in this
formation, as slip-stick vibration is expected.

I [Explain... ) ( Close 11 Help... 11


Figure 10 Operating Practices Information

371
Drill Bit Advisor: Recommendations for No. of Primary Cutters
Facility: 1 Version: 1Plan 1 Bit Class: IPDC 1
Well: 1 Hole Section: 112.2511 in 01 Zone: [1J
ENVIRONMENTAL ATTRIBUTES
Ka c
.2
.-
Preferred: mill cu ~

Acceptable:

..
1111
"Cl
c<I.l
S
S ~ 0 ....
~
~

~ ....0>
...
-....... ~
CZl
=c ~

....~/Xl...
~
Rejected:
~ ....ur::: r::: '" /Xl

Attr. Missing: ~
~ 0
Q)

9 ~ E-<
~
.~
Q)
-- B
til
Q)

.... ~...
.0

Constraint: T
..!><:
CZl

~
'"
g "0r::: ..r::: ;j
til
CZl
- -<
Q)
~

CZl /Xl
~
~
0
~
til
~
>
Q)

.....
0

~
::::::::::IQ~R::mypi;::~::I:I::

~
I [EXPlain... ) ( About... ) [ Close ) [ Help... ) I
Figure 11 Stipulation of Bit Requirement from a Group of Rules

372

Vous aimerez peut-être aussi