SPE 23664

Directional Drilling Expert System

E.Martinez, Corpoven. S.A.

SPE Member

Copyright 1992, Society of Petroleum Engineers, Inc

SPE Society of Petroleum Engineers

This paper was presented at the Second Latin American Petroleum Engineering Conference, II LAPEC, of the Society of Petroleum Engineers held in Caracas, Venezuela, March 8-11,1992

This paper was selected for presentation by an SPE Program Committee following review 01 information contained in an abstract submitted by the author{s). Contents of the paper, as presented, have not been reviewed by the II LAPEC or the SPE and are subject 10 correction by the author(s). The material, as presented, does not necessarily reflect any position of the II LAPEC or the SPE, its officers, or members. Papers presented at SPE meetings are subject to pubication review by Editorial Committees of the Society of Petroleum Engineers. 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 Publications Manager, SPE, P.O.Box 833836 RiChardson TX 75083·3836 U.S.A. Telex 730989 SPEDAL.

ABSTRACT.

Minimizing the cost of drilling a directional well is a major concern for the drilling engineer. This computerized system was developed in order to provide the drilling engineer with an advisory tool which recom­mends changes in the Bottom Hole Assemblies (BHA). This results in better accuracy and faster decision making; hence, considerable savings in time and money.

The system involves five integrated modules:

A knowledge base where all the expertise concerning the subject is stored. .

A database with information from the BHAs used in drilling of previous directional wells and their statistical behavior.

A database with information' from the theoretical behavior of BHAs commonly used in directional drilling.

A database with information from directional surveys performed over the path of the bore hole.

Comp~terized applications developed to calculate and plot the path of the bore hole.

This system was developed using an expert systems technology, including a knowledge base and a inference engine. The computerized applications for plotting and calculating the course were developed in a conventional way using BASIC and PASCAL': Databases were all developed in a relational model.

References and graphics at end of paper

INTRODUCTION.

The use of tools that permit process optimization, reduce the time spent between operations, develop new techniques and provide information updated and precise, are of high priority in the Industry and are part of the strategic planning of many successful companies around the world.

Constant efforts are employed in incorporating new technologies in order to improve the processes, specially those including new computational applications, as well as those promoting the creative abilities of the personnel.

Among other innovative techniques, Expert Systems have drawn special attention. The most meaningful ambition is that these Intelligent Systems can keep the knowledge as part of the capital of the Corporations.

How can we define Expert systems? This question has generated great controversy, but scientific and engineering circles agreed that it can be an "AUTOMATIZED SYSTEM THAT USES ARTIFICIAL INTELLIGENCE, AND MIMICS THE RESPONSES OF A HUMAN EXPERT IN ONE SPECIFIC FIELD OF THE KNOWLEDGE, BEING CAPABLE OF SOLVING PROBLEMS IN THE ABSENCE OF THE HUMAN EXPERT" .

The main difference between an Expert System and a Conventional System is the capacity of the former to deal with real life problems, that presents challenges to the capacity of the brain of a human being, and to solve

223

2 ERNESTO R. MARTINEZ, CORPOVEN S.A. SPE 23664

conflictive tasks using processes that mimic the thinking capabilities of an expert.

The Latin American oil Industry has not been far behind other companies in this issue, many expert systems have been developed in the last three years, one of them is the principal protagonist of this article, and it was called SIEXPERD (Expert System on Directional Drilling) .

Directional drilling is a technique of directing a well bore along a predetermined course to a target located a given distance from the vertical axis of the superficial location. The Oil Industry has used this technique where there exist topographical obstacles (lakes or mountains) right up to the point where the expectation of finding hydro~ carbons is greater or for more urgent tasks like drilling a directional relief well so that mud and water could be pumped in to kill a wild and cratered well.

This assignment is not easy and requires a strict control over the drilling parameters, in order to maintain the bore hole between the allowable values that permit the drilling to reach the objective.

The Drilling Engineer must make decisions, based on.his experience, which are decisive in the success of ths well. Millions of dollars, even human lives, depend on his judgement.

Based upon that, and knowing that the experience acquired by a few engineers must be kept in the corporation, it was decided to develop an Expert system based on this background.

METHODOLOGY.

During the development of an Expert System the scope of the project, and the available resources, must be precisely structured in order to achieve success.

The basic structure of the knowledge must be built before the beginning of the information or knowledge gathering phase or Elicitation. This process of knowledge extraction must be cautiously driven in order to avoid rejection by the expert.

The analyst or knowledge engineer must formulate various models previous to beginning the compilation of knowledge. The modeis in reference are:

Behavioral Model: This reflects the procedure that the expert uses for solving a problem and the tools employed. (books, references to other expert, data anal¥sis, etc .. ). This model allows the knowledge gathering (elicitation) under a st.ructured scheme, getting this way a more ordered knowledge base, keeping out unnecessary elements.

Fundamental or Basic Model: This model associates the elements that are distinctive I of directional drilling, (path model, tools, special machines, etc .. ) and their relations. I This model is distinctive of the process, its behavior and its components.

cognoscitive Model which covers the experiences and the heuristics, gathered by the expert, related with his specific work area, and the problems that are supposed to be solved by the Expert System.

DEVELOPMENT.

The main requirement of the project was the development of a drilling engineer computerized assistant, with the capacity of making suggestions and solving problems related with the directional control of the bore hole.

The first task in the plan was to define the boundary of our System.It was determined that it would cover the following steps during the work of the Engineer:

1. Planning and Design 2. Control of drilling parameters 3. And finally statistical Analysis.

This. is the general structure of the Behavioral Model, based on these activities, experienced engineers were interviewed. Based on their response, a frame was completed where the tasks related with decision making during a Directional Drilling Procedure were identified.

Planning and Design is one of the most critical steps involved during directional drilling. During the planning the engineer selects the course pattern, maximum lead angle, kickoff and the drop off· points, the rate of angle increment and the rate of angle reduction, and if we do not choose the appropriate parameters it is almost certain that we will be in trouble during the drilling operation.

Designing consists of the selection of the drill pipe and casing diameters, weight, type, Bottom Hole Assemblies, drill bits, etc. The characteristics of these elements and the possibility of changing them by others with different attributes must correspond with the changes in the course selected during the planning.

In order to maintain control during the process of drilling an oil well, the engineer collects all data needed to get the position of the bit (under the uncertaintity parameters), makes computations and compares the results with the planned course.Using his judgement and experience he makes the decision needed to maintain or correct the course. One of the most important responsibilities during

SPE 23664 DIRECTIONAL DRILLING EXPERT SYSTEM , 3

this stage is the early detection of problems, the personnel on duty must be conscious of this.

One aspect strongly stated by the experts was the necessity of graphic representation of the course in order to give a quick report of the status of the well.

One of the most important parameters involved in directional drilling control, is the selection of the Bottom Hole Assembly. This consists of different elements combined to guide the drilling string in direction and slope. This was one of the greatest concerns for the experts, because of the lack of information found about the behavior of the BHAs in different oil fields and at differ.ent depths.

Depending on the type of profile the Basic: Model used for this study was split in four sections: J, Modified J, S and Modified S.

The S profile, for example, comprises six stages: initial vertical section, start of deviation, angle growth, maintenance of angle, angle fall and final vertical section. Each course section elements was studied isolately from the other sections. This model is analyzed and connected with the Behavioral Model to obtain a structure that eases the system development. The main purpose of this is to identify when the expert uses heuristic and when he uses a methodical strategy in order to get optimal results.

This is important because this systematic behavior can be easier and faster impl~mented and we could focus our efforts 1n the heuristic that usually are harder to express by the expert and even harder to represent by the analyst. On the other hand this analysis ensures that all parameters used by the expert are included in our design and irrelevant data is excluded.

The cognosc:i ti ve Model is a collection of rules or Heur ist ics, used by the expert to solve different situations found during the operations. These rules are gathered from the expert, from interviews, and guided tests followed by a validation. This represents the heart of our system.

This work focused the heuristics used by the expert in order to optimize the BHA's selection, and those that regulate the decision of pulling out the drill string instead of drilling ahead.

This Expert System mimics processes used by human experts, therefore it needs identical sources of data, in order to obtain similar results. Hence it was necessary to implement databases with BHAs, ,. their theoretical behavior, their actual behaviors in previous wells, well design data, and all measurements and calculations required to get the position

225

of the bit during the drilling process and its relative position with the planned route.

The expert also makes calculations outlines the course of the hole, so processes also should be covered by system.

and this the

Additionally this expert system has uncertainty management mechanisms, in order to get the most appropriate answer in an environment where many inputs are not facts only assumptions or good estimates. In thi~ system the uncertainty factor expresses the lack of confidence in the BHA's behavior. This uncertainty factor is inversely related to the amount of data, or experience gathered by the system, and directly related to the variance span of the directional behavior in previous drillings.

The application was developed using various Software tools: Data Base Managers Programming Languages, and Expert system~ Shells. The selection of these depends mainly on th~se ~actors: ease of use, portabil i ty, graph1cs a1ds, and size of the knowledge base. Another factor considered was the necessity that the system run at the well site as well as at the office.

In order to get those goals it was decided that the hardware must be a microcomputer under MS-DOS, and of course, software compatible with them: a data base manager commonly used by our applications, Pascal programming language for the graphics and calculations and a PC Expert Shell, with capacity to interface both of them and a graphic front end.

Once you are in the system it shows five options: PLANNING, CALCULATIONS, GRAPHICS, CONSULTATION, DATABASES, REPORTS AND HELP. Those options are activated by means of a mouse. At the bottom the user can see a window that shows messages like those sent by the help procedure or by the consultation.

The PLANNING option gives the engineer the possibility of studying diverse alternatives: profiles "s" or "J", changes in the Kickoff point or Drop Off Point, in the Build Up or Drop Off Rates, allowing a thoroughful analysis.

The GRAPHICS option permits the comparison between the planned course and the actual course, it can zoom in the area near the bottom of the hole.

The CALCULATIONS option giv.es the user the distance between the actual and the planned courses, the angle and direction and their trends.

The DATABASES option gives access to all the data. Here the data can be consulted,

4 ERNESTO R. MARTINEZ, CORPOVEN S.A. SPE 23664

modified, deleted or included. The databases contains the following data:

BHAs used in drilling of previous directional wells and their statistical behavior.

BHA's theoretical behavior.

Directional surveys performed over the path of the bore hole.

The CONSULTATION option starts the inference engine, making the analysis of the data using the knowledge base rules; once it finds an answer for the situation, it makes a suggestion. If this recommendation includes BHA's change the system examines the databases and suggests the most suitable.

The REPORTS options enable the user to print the data, and calculations.

The HELP option provides assistance to the user during the consultation.

CONCLUSIONS.

This system has been used at various wells in different fields in eastern Venezuela operated by CORPOVEN. The experience o~tained by this activities have been valuable, first because the system has gathered data related with the actual behavior of' the BHAs used in those areas, and this results in less uncertaintity, and less time spent in the selection of BHAs and the optimization of the course and the pullout string times; second more personnel have received training in the process, and are more conscious about the parameters involved in the determination of the adequate action for rectifying a deviated course.

Other accomplishments of this assignment were the experience and confidence acquired by the data management group in expert systems development, and the backing of the users and management in present and future projects.

The use of a systemic approach and modelling the behavior of the expert and the 'phisical phenomenon together with the heuristics, provides a well structured frame. This ensures the quality of the product and its fast development.

This work has been advantageous to all people involved and a great experience.

ACKNOWLEDGEMENT.

The author thanks PDVSA and its SUbsidiary CORPOVEN S.A. for allOWing to publish this work.

226

REFERENCES.

Crafting Knowledge Based Systems Wiley-Interscience 1990

Rolston, David Inteligencia Artificial y Sistemas Expertos McGraw Hill 1990

Zambrano, Alexander sistemas Expertos UDO 1990

Mishkoff, Henry Understanding Artificial Intelligence Texas Instruments 1986

Chorafas, Dimitri Appliying Expert Systems in Business McGraw Hill 1987

Harnell, Tim Sistemas Expertos Anaya Multimedia 1986

Harmon, Paul; King, David Expert Systems John Wiley & Sons, Inc 1985

Wong, V.G. Conocimiento Experto U.L.A. 1989

Jackson, Peter Introduction to Expert Systems Addison Wesley 1986

Klahr, Philip;Watherman, Donald Expert Systems: Techniques, Tools and Applica­tions Addison Wesley 1986

I

C

0

U

R

S

E

MODELS

I BEHAVIORAL I MODEL

I I BASIC

I ORGANIZATION I MODEL

I TOOLS I I BOTTOM HOLE ASSEMB. I

I COOHOSCITIVE I I TRAJECTORY I I STEPS OR I MODEL MODEL ZONES

I HEURISTIC I

FIGURE 1. EXPERT SYSTEM MODEL FRAMEWORK

TYPE ·S" PR.OFILE

FIGURE 2. BASIC MODEL REPRESENTATION EXAMPLE: 'S' PROFILE

C

0

N

T

R

0

L

BEHAVIORAL

DATA OATHEAINCl

CALCULATIONS

TRAJECTORIES

C()lflw.tPAAISON

EVALUA.TION

SOLUTIONS

SOLUTION SELECT'ON

MODEL

WELL DATA

THEORETICAL DATA

t:'lr;IlDP'\ RFHAVln12Al MnnPl I'=XAMl>1 t:'

227

SPE 2366 4

•

•

•

KNOWLeOGe SOVACI!

USER IHTERFASE

SPE 23664

EXPERT SYSTEM • APUCATION

FIGURE 4. EXPERT SYSTEM KNOWLEDGE ACQUIsmON

SIEXPERD

REAL DATA

THEORETICAL DATA

REAL BHA'S

THEORETICAl. 8HA'S

FIGURE 5. DIRECTIONAL DRILLING EXPERT SYSTEM OPERATIONAL STRUCTURE

228