SPE

Society of Petroleum Engineers

SPE 19133
Enhanced Drilling Softwares and
Integrated Advisor Expert Systems
by R.G. Fenoul, Elf Aquitaine
SPE Member

Copyright 1989, Society of Petroleum Engineers, Inc.

rhis paper Was prepared for presentation at the SPE Petroleum Computer Conference held in San Antonio, Texas, June 26-28, 1989.
This paper was selected for presentation by an SPE Program Committee fOllowing review of information contained in an abstractsubtnltted by the author(s). Contents of the paper,
as presented, have not been reviewed by the Society of Petroleum Engineers and are subject to co.rrection by the author(s). The material, as presented, does not necessarily reflect
any position of the Society of Petroleum Engineers, its officers, or members. Papers presented at SPE meetings are subject to publication review by Editorial Cornmittees of the Society
of Petroleum Engineers. Permission to copy is restricted to an abstract of not more than 300 words. Illustrations rnay not be copied. The abstract should contain conspicuous acknowledgment
of where and by whom the paper is presented. Write Publications Manager, SPE, P.O. Box 833836, Richardson, TX 75083-3836. Telex, 730989 SPEDAL.

ABSTRACT

INTRODUCTION

This paper describes the evolution of scientific

softwares applied to ELF AQUITAINE's drilling
operations.

The record of the use of scientific methods for

optimizing drilling jobs goes back to several
dozen years. To illustrate this remark, one may
just
recall
LUBINSKI's
charts
concerning
directional behaviours of drill strings (see
re.1).

Starting from an historically scattered situation.

it points out ho~ strategic decisions taken by the
company have led to the development of a framework.

More recently, in the sixties, the main offices of

large Petroleum Companies were equipped with
powerful scientific calculators. In those days,
solving a problem meant filling in a form which
was eventually transformed into a set of cards.
These cards were punched, checked and f~d to the
computer. The results were printed on listings
which often were bulky and difficult to decode.
In the best case, if indeed there ~as no misprint
or mistake when inputing the parameters, such a
process would require, at the minimum, a whole day
before obtaining any results.

This frame~ork, called CALPIN, set up to receive

and
integrate
scientific
programs,
is
now
operational. It contains numerous softwares, including expert systems which are useful for drilling
planning on operational bases and also help
decision-making, directly on the rigsite.
The development guidelines as well as the solutions
adopted are described in detail. The modular structure of the development allows for easy maintenance
of the system and favours its evolution. In particular, i t is easy to integrate a new software in
the framework. Open to the outside, CALPIN enables
the user to be connected with nearby data bases and
to process his own data files by means of independent data-processing tools.

In the seventies, the first "Desktop computers"

appeared. These computers were generally bulky and
ran on a rudimentary BASIC. Their memory space
amounted to 10 to 20 K bytes. The programming rule
was the "Overlay". The first drilling scientific
calculations were translated or developed on this
type of hardware and the first complete "packages"
appeared. Though the formulae used were rudimen-tary, these softwares were comprehensive and dealt
with problems of trajectories, casing computation,
cementing, hydraulics and the resistance of simple
offshore structures.

References and illustrations at end of paper

..

123

ENHANCED DRITJLIUO SCFTWARES AND INTEGRATED AQVISOR EXPERT SYSTEMS

In the early eighties, the computers became faster,

more powerful and the languages were improved.
Numerous extensions to the BASIC language were
allowed.

SYSTEM OBJECTIVE
For each production specialty area, Drilling in
particular, it was necessary to rapidly set up a
scientific computation system with the following
characteristics :

At that time Drilling Research had improved and

more and more complex mathematical formulations of
phenomena
appeared.
Concurrently
with
these
improve~ents,
the operational user started to
discern that fine data-processing had a real
interest and required fro~ computer services the
development of management softwares or small data
bases.

- having the ability to replace

possible the existing softwares
-being consistent with
recommendations
- being

It is interesting to notice that several companies,

including ELF AQUITAINE,

application modules for
calculators (see ref.2).

SPE 19133

the

as

soon

Company's

suitable to integrate new

by the Research Centers

as

computer
softwares

d~signed

have developed drillitlg

programmable hand held

- being easy to maintain and require as little

work as possible in case of portability
~

ln the mid-eighties, the user was faced with

ever-growing computerization. Satisfactory on the
whole, it was composed of tools independent from
one another, each tool having been developed to
meet a very specifically defined need.

being open .to other systems, which means having

the ability to exchange data

- lastly, being user-friendly in order to be

operated by any user with no specific computer
knowledge.

PROBLEM DEFINITION
ADOPTED SOLUTION
Though the user was on the whole satisfied, some
problems started to arise :

All the softwares we had to gather dealt with

drilling operations but had very different characteristics. Thete were even expert systems to be
merged with algorithm programs.

- All these systems had been designed correctly at

first. But several times transported, improved,
and made user-friendly later on, they turn out to
be difficult to improve or even maintain.

When we started this work in 1986. we thought it

was necessary to come back to a .basic scientific
layout of the software :

- Since developed with the maximum language

possibilities, they use BASIC extensions which
are efficient but raise huge problems of
portability.

INPUT

... Developed at different times, the user's inter

face was not the same for all products.

COMPUTATION
OUTPUT

- All these systems were discrete and, when

the user carried out a study requiring several
programs, he had to input similar data several
times.

and apply to this layout

principles such as :

- Finally, with such a policy, the wish for adding

new softwares developed by the Research Centers
would have, in the long run, worsened the incipient difficulties.

77

is

compulsory

for

and

strict

* STANDARDIZATION
* MODULAR STRUCTURE
* POOLING OF RESOURCES
With these basic principles and some new ideas
described below, we have developed a framework.
This framework manages the data flows transiting
from one software to another and distributes the
necessary processing resources.

In 1985, aware of this problem, ELF AQUITAINE took

a series of standardization measures at the
company level and for each specialty area :
- The FORTRAN
calculations

sensible

scientific

We have named this framework CALPIN.

- UNIX is defined as the operating system
CALPlN has the main following characteristics
- A development
planned

line

for

hardware

equipment

- Input/Output soft tools are specified

manager and graphic software).

is
- All Input and Output files are considered
identical by the framework. An output file of

(screen

124

SPE 19133

Robert FENOUL

one software may be considered as an input file

for another software. This is the case for a
CASING file. It is the result of an optimization
calculation, but is considered as an input for a
friction calculation used to determine the surface parameters while running in (see fig.!).
- Each handled entity represents a complete logical
object, For example, a hole trajectory or a
description of a drill string composition.
All these objects are handled as a whole, which
means that such an object may be used for directional calculation as well as for hydraulic
optimization.
Such objects characterize
the hole trajectory
the drill string description
the casing string
the mud characteristics
the internal geometry of a well for a given
phase
the overall architect~re of the well structure
the surface equipment
the inside temperatures
the lithological and physical characteristics
of the formations gone through
Each object being at the same level, it is
possible -in any program- to select any object,
associated with any other one. Of course,
coherency tests avoid incompatibilities.

Y~IN

DEVELOPMENT STEPS

There were three main steps in the development of

this framework. The first one consisted in
analysing the existing softwares
and grading
them by an order of priority and establishing the
related specification. The second step was the
deve~opment of the framework as such. Finally, the
third one consisted in industrializing the
product, by having it tested by users, gathering
their comments and ;improving the programs. These
steps are detailed below.
Establishment of the specifications
Until 1985, a lot a work had been done and
undoubtedly the know-how acquired during those
achievements had to be saved and further used for
the develop~ent of the new framework. So, all
scientific softwares have been analysed regarding
their scientific content, effective rate of use,
usefulness
for
Drilling
operations,
user
friendliness and comments made by the users. All
these softwares have been graded by their
development priority order. All the inputs have
been analysed as well as the outputs (see fig.3).
~rom the analysis of data flows between softwares
a certain number of "logical objects" have been
deduced. Each object constitutes one indivisible
data set common to several softwares. Along with
internal data flows, data tranfers to or from the
system have been analysed.
From this work onwards, the architect~re of the
framework has become more clearly outlined, in
particu,lar,
the files associated with these
logical objects. It has therefore been possible to
define the formats for each of them.

- Each object has its own input, modification or

print out module. Basically, these transactions
are alphanumeric, This is done through screen
masks or tables. However, to inp~t drill string
data, we have developed a graphic interactive
module that enables the user to create its
Bottom Hole Assembly visually by means of icons
selecte~ with a mouse (see fig.2).

Thus, this phase

specificatione :

enabled

us

to

define

the

develop a framework gathering several softwares

exchange data by means ot working files
develop modular program~
develop specific input modules
develop specific output modules

- All these objects are managed by only one module

that enables the user to create, recall or store
the selected object. For the user, each object
is characterized by an associated comment.

Pevelopment of the framework

- Only the comments are displayed for the user who

does not have to deal with the computer concept
of files or with the concept of UNIX tree.

This development was carried out either by

translating the existing algorithms or by working
out new ones. An analysis has been carried out to
determine what would be the computation modules
common to the set of selected softwares. Each
module has been developed, later on, as a
sub-routine (in the FORTRAN meaning of the term)
with input and output parameters accurately
defined. Consequently, each time a software has to
carry out a particular computation or management
task, the corresponding module is available and
stored in a common library.

- At each screen ~ask, an aid screen is associated

which
allows
the
user
to
have
on-line
documentation available.
- All the necessary "data files" are located in one
single place. Thus all the users benefit from
their updating. This is the case for the
geometrical and mechanical characteristics of
casings.

Thus a ,series of soft tools are now available.

125

ENHANCED DRILLING SOFTWARES AND INTEGRATED ADVlSOR :EXPERT SYSTEMS

They concern the following areas :

SPE 19133

It has been decide<,i to translate CASES in FORTRAN,

upgrading it and to integrate it into the
framework. C!MENTEX will be integrated too in 1989,
COMPLEX. will be integrated later on.

input, print out or modification of the entered

data.
These data may come from an alphanumeric input,
from a graphic interactive input or may be extracted from an existing data base,

file management

Once the input and output para~neters were clearly

defined, the integration of these systems did not
raise special problems (see fig.5). The only
difference with other softwares is the way in
which the questions are asked by the inference
engine.

. common computation modules

Product industrialization

graphic or alphanumeric outputs of results

For example, the d:~;illed/vertical depths crossrelations for a given trajectory constitute a
sub-program, common to most of the softwares,

By the end of 1987, we had a prototype composed of

an elententary set of softwares usef\,ll to the
drilling engineer on a base. The system was therefore installed in a Dutch pilo~ subsidiary both on
the computers of the base and on the rig. It was
installed too in the company qeptral offices. The
prototype was left at the users'disposal for four
months. Then an enq\,liry allowed us to determine
the framework weak points. Apart from a few bugs
which have been found, three main remarks have
been voiced:

At the beginning, all these steps seemed long,

drudgery and slowly-going. In fact, they turned
out to be profitable and we now have a library of
very power~ul soft tools.
This enables us to develop a new software without
having to design a good many elementary func.tions.

the general arrangement of ~he progra~s had to

be improved
- some internal masks had to be re-defined
- the computation spe~d of some softwares was not
satisfactory

There. is another important advantage to this

modular system : Nuintenance. For example, if we
wish to modify the pressure losses computation
theory, we just have to modify the source program.
This modification will be passed on to all the
framework softwares through a simple link edition.
'l'his phase led us to define a soft architecture
which is shown in figure 4.

During the second half of 1988, CALPIN was upgraded by taking into account these remarks. The
softwares concu;rrent,ly developed by the Research
Centers have been also integrated.

Since 1982, ELF AQUITAINE has developed several

expert systems applied to Drilling area.

CALPIN is now an operational product. It has left

the Research Centers for the Operational services
which will install it and train the users.

The first one was SECOFOR (see ref.2). This system

gives a diagnosis of the type of sticking
occurring while drilling. It was followed by
FORCAST,
a
preventive
system,
derived
from
SECOFOR. Both systems have been developed in LISP
language on XEROX equipment with TECHNOWLEDGE S1
inference engine. In 1985, CASES, an aid sy~tc.::
for the determination of the composition of the
drilling strings was elaborated on a BASIC
calculator (see ref.4). PARSEC, the inference
engine used, is an "in house" product which has
been tranf;>lat.ed into FORTRAN (see ref.5). In 1988,
two new products have been developed :

Today, it is composed of n\,lmerous softwares useful

both on operational bases and rigsites. The
integrated softwares are listed in appendix.
ASSETS OF SUCH A FRAMEWORK
These assets concern both users and developers of
softwares.
- For the users
They deal with one single system displaying all
the available softwares. The displays are
homoge~eous and identic,:al. All input data and
objects thus created are then available and the
user can create its introduction environment as
he wishes. For example, by changing the "object
trajectory", he can pass from a vertical well to
a deviated one, all the other data remaining
unchanged.

CIMEN'l'EX

enables a complete cementing job

sta:J;"ting from the Q.etepnination of
the position of the top-cements to the fitting
out of the casing string with centralizers and
scratchers, without forgetting the procedure for
the injection of cement slurry. 'l'his system runs
on UNIX and uses PARSEC.

preparation~

Moreover, the framework being open, the ~ser can

extract data from nearby data bases. He can also
compute the content of in.put/output files with
tools that are independent from the framework.
For example, it is very easy to transfer data

COMPLEX has been designed to determine a well

completion system. It runs on UNIX and uses the
S1 inference engine.

126

SPE 19133

Robert FENOUL

resulting from a trajectory computa,tion into a

personal computer table and handle or visualize
them later on.
- For the computer specialists
They are either in charge of the system
maintenance or of developing new softwares.
In all cases, the modular design ensures very
eesy maintenance. The pooling of files (such as :
casing tube characteristics) allows for an easy
update available to all users straight on. The
library of soft tools thus developed also enables
easier development of new softwares (see fig.6).
Lastly, the fact that all data are systematically
going through standard files ensures that in case
of portability the problems are widely reduced.
We have experienced this with ORPHEE 3d, a directional behaviour software (see ref. 6), when we
carried it into a personal computer environn1ent
under DOS. Apart from the fact that we have had
to cut some computation subs which were too long,
the only additional work has been to change the
screen management software with the associated
modifications in the data input programs.

Now containing a standard set of programs, it has

been designed to integrate easily new softwares,
including knowledge-based systems.
Open to the outside, it has the ability to exchange data. In particular, it will be able, on
request, to catch information from on-line acquisition systems. The processing of this information
will help the user in decision-making on the rigsite.
ACKNOWLEDGEMENTS
I wish to thank ELF AQUITAINE for permission to
publish this paper. I also particularly thank
. M. CARREDA, C. (CISI INGENIERIE) and M. PERUSAT,
Y., for their technical contribution.
REFERENCES
1.

Lubinski, A.
and Woods, H.
"Factors
affecting Angle of Inclination and Doglegging
in
Rotary
Boreholes"
Drill.
and
Prod.
Practise, API, (1953) 222-250.

FUTURE EXTENSIONS

2.

We do not consider that this system is complete.

First, it will have to live and be upgraded to
keep pace with technology, hardware and users
requirements. Then it will be necessary to improve
its content and its user's interface.

Aalund, L.R. : "Computer systems software" Oil

and Gas Journal, January 11, 1988, pages
33-42.

3.

Courteille, J.M., Fabre, M., Hollander, C.R. :

"An Advanced Solution" : The Drilling Advisor
SECOFOR", Paper SPE 12072 presented at the
58th
Annual
Technical
Conference
and
Exhibition, San Francisco, CA; October 5-8,
1983.

4.

Marion, P., Huynh, C.T., Fenoul, R. : "CASES :

CASING Expert System", Paper presented to the
Offshore Conference and Exhibition, Aberdeen,
UK, october 8-10, 1985.

5.

Giard, D. , Roger, C.
"PARSEC
Programme
Alliant Representation Symbo1ique Et Calcul",
Inference
Engine
presented
at
the
6th
International Workshop on Expert Systems and
Their Application, Avignon, France, April
28-30' 1986.

6.

Birades, M. : "ORPHEE 3d ; Static and Dynaii1ic

Tridimensional BHA Computer Models", Paper SPE
15466 presented at . the 61th Annual Technical
Conference and Exhibition, New Orleans, LA
October 5-8, 1986.

To achieve
planned :

this,

the

following

actions

are

- improve the mathematical modeling of certain

phenomena, particularly in well design area
(casing composition optitnisation in complex
cases, two~phase flow phenomena -with or without
dissolution-, directional behaviour of drill
strings . )
- increase the number of softwares built on
knowledge-based system models.
Such systems
associated with data bases are very convenient
for taking into account the know-how which goes
with the use of a software
- prepare on-line supply for both types of
softwares previously described. We think that
this action is by far the most interesting and
promising one. Such pilot softwares are being
worked out in our company.
These actions are part of an overall futureorientated Research and Development policy and
also calls upon the know-how acquired on the past
achievements.
CONCLUSION
CALPIN is now an operational framework. It has been
tested in central offices, base offices and directly on rigsites. Its aim is to assist the operational user in decision-making on the rig and it is
also helpful for well planning.
127

ENHANCED DRILLING SOFTWARES AND IN'+EGR,ATED ADVISOR EXPERT SYSTEMS

AP~ENDIX

: List of the integrated softwares

* Drill

* Hydraulics

str~ng behaviour
- ORPHEE 3d : directional BHA behaviour
- Friction computation

- pressure losses calculations

- U tube effect
- thermal and dynamic effects of mud
circulation

* Trajectory
-

SPE 19133

calclulations
single path calculation
previsional trajectory design system
distances between well paths
measurement errors areas
hori~ontal projection of several wells
graphic path comparison

Gas kick
- monophasic gas kick simulator
- diphasic gas kick simulator (by the end of
89).

* Casing

Rock mechanics
- bore hole stability diagram

- casing string design exper~ system

- cementing job preparation expert system

Parameters
.__r_ra_je_c_to_rv_

_.l ' \

Hook load

Hole

Constraints

Torque

Mud

FIG 1

OUTPUT FILES MAY ALso BEEN CONSIDERED AS INPUT FILES

128

..... ~

'

'(,.~i.;.<

leE l91JJ

CODE ..............................,..................... 1
NUMBERS OF ELEMENTS .......

ST

D ~

TOTAL LENGHT .............................. 1 21.92 I

OD BLADES .......................................112

1/41

OD BODY .............................................I

ID BODY ............................................... J2

13/16J

D
0

FIG 2 : INTERACTIVE BOTTOM HOLE ASSEMBL V INPUT

LOGICAL OBJECTS

PROGRAMS

( Directional behaviour)

I
I
I
I

Drill string

Trajectory computations

Trajectory

( Casing computations )

Casing
Hydraulic

Gas kicks

Rock mechanics

Mud

Hole geometry
Well architecture

calculations )
129

CAL PIN

Workstation
management

Masks
files

Shell
scripts

FIG 4 : USER SYSTEM ARCHITECTURE

B
I

b~~~as

I I
Users

/1~
I I 1. ____I

User 1

User 2

User X

Knowledge
data base

FIG 5 : EXPERT SYSTEMS

INTEGRATION

Specific library :
- Input files management
- Output files management
- Computing subs

i
FIG 6 : SOFTWARE DEVELOPMENT

130