Optimization of Rotary Steerable Drilling: Moustafa Elshafei
Optimization of Rotary Steerable Drilling: Moustafa Elshafei
Optimization of Rotary Steerable Drilling: Moustafa Elshafei
Moustafa Elshafei
King Fahd University of Petroleum and Minerals
Dhahran, Saudi Arabia
elshafei@kfupm.edu.sa
Mohammad Khamis
King Saud University
Riyadh, Saudi Arabia
mokhamis@ksu.edu.sa
Abdulaziz Al-Majed
King Fahd University of Petroleum and Minerals
Dhahran, Saudi Arabia
aamajed@kfupm.edu.sa
Abstract- Directional Drilling (DD) is the process of directing the wellbore along a desired trajectory to a
predetermined target, using different technologies, including Directional Steering Systems (DSS). DD achieves high
well production and improves accessibility of oil reservoirs in complex locations. This paper addresses the problem
of real time optimization of drilling parameters in directional drilling to minimize the drilling time, and the deviation
from the planned trajectory. This research considered a typical Rotary Steerable System (RSS) for directional
drilling with other drilling parameters such as rpm, torque, and weight-on-bit. We assumed no prior knowledge of
the formation properties. However, rock specific energy is considered as an unknown disturbance in the model. The
cost function parameters can be selected for the best compromise between trajectory tracking, energy consumption,
and rate of penetration.
Keywords: Directional drilling, Directional Steering, horizontal drilling, RSS, drilling optimization, ROP
optimization, real-time drilling.
1. Introduction
Directional Drilling is the process of directing the wellbore along some trajectory to a predetermined
target, Bourgoyne et al (1986). Directional Steering Systems (DSS) facilitate the accessibility of the oil
reservoirs if the reservoir is having large surface area and distributed over thin horizontal layer. Wells
could also be drilled directionally for several purposes like drilling the underside of an environmental
obstacle, or where multiple wells are drilled from one central surface location, such as an offshore
platform, Joshi (1991), Baker (2001), and Eustes (2007).
In 1999 a Rotary Steerable System (RSS) was introduced to directional drilling market, (Eustes,
2007). RSS allows three dimensional control of a bit without stopping the drill string rotation, and
increases the efficiency of directional drilling operations by reducing drilling time. It also provides better
borehole cleaning with fewer wiper trips, optimizes drilling parameters, and provides a higher rate of
penetration while drilling (Byliss and Matheus, 2008).
Unlike conventional drilling systems, the directional drilling requires sensors to provide estimations
of the azimuth ψ (deviation from the north direction in the horizontal plane), the inclination θ (deviation
from the vertical direction, or pitch angle), and the tool face angle ϕ (roll angle) of the drill bit, hence it is
called Measurement While Drilling (MWD). MWD systems includes a three-axis magnetometers and
167-1
three-axis accelerometers to determine the azimuth, the inclination and tool face angles, Thonhauser
(2004), and Dunlop (2011).
Modelling of the drilling operation for control and optimization is a challenging problem due to the
diversity of the factors affecting drilling as well as uncertainty in their determination. Among these
factors are the Bottom Hole Assembly (BHA) dynamics, torques and drags, formation properties, bit-
formation interaction and drilling fluid properties and its hydraulics, Bourgoyne and Young (1974).
At the while-drilling mode, the directional drilling system should try to coordinate various control
actions (RPM, WOB, mud properties, rate and hydraulic pressure, inclination actuators, azimuth
actuators, etc) to keep the down hole path close to the preplanned path trajectory. The control system
should be able to accommodate the large uncertainties in the formation properties, and stay within the
operational constraints. The main task in directional drilling is to properly orientate the down hole tool to
steer the well bore in a desired direction, and minimize the drilling time.
Bourgoyne and Young method is the most important drilling optimization method since it is based on
statistical analysis of the past drilling parameters. The model proposed by Bourgoyne and Young (1974)
derived equations to perform the ROP estimation using the available input data. This model is considered
as one of the complete mathematical drilling models in use of the industry.
William and Jeff (2005) showed how Mechanical Specific Energy (MSE) was implemented in a
drilling information system in real time on the rig and at remote monitoring locations. The study showed
that the use of MSE in real time is a useful tool for both drillers and drilling engineers. Conducting MSE
tests in real time is an effective way to develop an understanding of MSE behavior and contributes to
acceptance by rig personnel. The general practice of adjusting drilling parameters to minimize the value
of MSE is a good rule of thumb.
Alum and Egbon (2011) developed semi-analytical model for Rate of Penetration (ROP) based on the
original Bourgoyne and Young Model using real time bit records, obtained from wells drilled in Niger
Delta reservoirs. Simple regression analysis was applied on the equation on the parameters that contain
differential pressure to obtain regression constants, which were then used to generate mathematical
relationship between ROP and drilling fluid properties.
Rashidi et al. (2008) presented a new method to combine Mechanical Specific Energy (MSE) and
Rate of penetration (ROP) models to calculate real time bit wear which takes into consideration the
fundamental differences between MSE and ROP models and that the latter only takes into account the
effect of bit wear. Encouraging results have been obtained which shows a linear relationship between
MSE (Rock Energy) and rock drillability (Drilling Strength).
Rashidi et al. (2010) described the real-time application of a developed model for bit wear analysis.
The model was developed based on the difference between rock energy model, Mechanical Specific
Energy (MSE), and rock drillability from rate of penetration model. It has been modified and
implemented as an engineering module in the newly developed software, Intelligent Drilling Advisory
system (IDA’s), and used to estimate real-time bit wear for both rollercone and PDC bits.
Tuna and Evren (2010) developed a model to optimize drilling parameters during drilling operations
such as weight on bit, bit rotation speed in order to obtain maximum drilling rate and hence minimize the
cost per foot and the overall drilling cost. The model developed used actual field data collected through
modern well monitoring and data recording systems, which will be used in predicting the rate of drilling
penetration as a function of available parameters. The study demonstrated that drilling rate of penetration
could be predicted at relatively accurate levels, based on past drilling trend. The optimum weight on bit
and bit rotation speed could be determined in order to achieve minimum cost drilling.
Koederitz and Johnson (2011) described the development and field testing of an autonomous drilling
system. This system software uses a test process to evaluate and quantify the drilling performance for a
given set of target setpoints. The research method is used to identify these setpoints; its development was
based on early work in the application of real-time Mechanical Specific Energy (MSE) display. Overall,
the field testing results were favorable, displaying that the potential for autonomous drilling optimization
without drilling knowledge is practical, flexible, and economical, exhibiting promise in a range of cost-
effective applications.
167-2
In this paper we present a unified approach for optimization of the drilling parameters and directional
steering. In section 2, we present a simplified model for DSS. In Section 3 we provide the optimization
function and procedure. The simulation results are presented in Section 4.
XE
North
Magnetic
YE
Gravity
Toolface
Drill bits
ZE
Fig. 1. Illustration of the azimuth angle and the inclination angle.
The orientation of the BHA with respect to the inertia axis is described by the rotational matrix R,
which is given by equation 1.
The rotational matrix R defines the transformation of a point P from the body axis to the inertia axes
as shown in equations 2 and 3.
167-3
xE
PXYZ yE RPUVW (2)
z E
Where [ xE, yE, zE]’ is the position of BHA with respect to the inertia frame, and Puvw is the position
of the point P with respect to the body frame {Ub,Vb,Wb}.
If we are interested only in the direction of the well bore, we may then ignore the roll rotation of the
BHA. In this case R is given by equation 3.
cs s cs
R sc c ss
s 0 c
(3)
2. 2. RSS equations
The most common method for directional steering is known as Rotary Steerable system (RSS),
Downton (2007). The rotary steering system has the ability to provide directional drilling control while
allowing continuous rotation of the drill-string. Consider the simple representation of a directional drilling
system shown in Fig. 2, where the steering actuator eccentrically displaces the centre line of the drilling
system away from the centre line of the hole by a controllable amount ecc in a given plane. The steering
actuator is placed at a distance L1 from the bit. At a distance L2 from the actuator, the collar is permitted
to pivot about a touch-point with the borehole called a stabilizer. The stabilizers, steering actuators, and
sensors are placed in non-rotating sleeve.
Let us assume now at the stabilizer point the BHA axis is aligned with the bore hole centreline. If the
drill bit is currently at location wb(k) (measured depth) then the stabilizer position is at [ wb (k ) L1 L2 ]
and the actuator position is at location [ wb ( k ) L1 ] . If the actuator creates an eccentricity between the
drill string and the hole (eccu, eccv), measured with respect the BHA axes, then using the small angles
approximation, the deviation angles of the drill bit from the stabilizer is given by
eccu
L2 (5)
eccv
L2
167-4
ecc
Os
Oh
Drilling
string
Stabilizer
L2
Actuator L1
Where eccu, eccv are the eccentricity components in the body coordinates Ub, Vb. respectively. The
deviation angles are usually about 1.5 degrees, with a maximum of 3-4 degrees. To simplify notations, let
us denote eccu, eccv .
At the point k+1 along the trajectory, the predicted drill position with respect to the body axis is
given by
The predicted location with respect to the inertia axes will be given by
ub (k 1)
PˆE (k 1) PE (k ) R vb (k 1) (7)
wb (k 1)
T Wob w
b Pin QF PoQm w
b Ah Ers (8)
167-5
Where w b is the ROP in m/sec, T is the drilling torque, ω is the angular velocity of the rotating
string, Wob is the weight on bit, QF the inlet drilling fluid flow rate, Qm, is the mud outlet flow rate, Pin, Po
are the surface inlet and outlet pressures respectively, Ah is the area of hole bore, and Ers is the rock
specific energy.
Qm QF w b Ah (9)
The w b Ah is the volume rate of the crushed rocks. This volume is usually less than 5% of QF. Hence
we will take Qm =QF. Then equation (9) can be simplified as
Since all the parameters, except Ers, are directly measurable online, Ers can be estimated. The
estimated value of Ers is then used during optimization to predict w b during the next control step of DSS,
as follows
T QF P
wˆ b (k 1) f (T , , DP ,Wob , QF )
Ah Eˆ rs P0 Ah Wob
(11)
3. Drilling Optimization
At each point k, we assume that we have a state vector X, and a target position Wbt(k+1), and
it is required to find the optimal control parameters
The state vector includes
X (k ) [ttotal , wb , x E , y E , z E , , , w b , Eˆ rs ] (12)
U [T , ,Wob , tk , , , QF , DP ] (13)
QF (k 1) QF 0 q Ah w
b (k ) (14)
167-6
T QF P
wˆ b (k 1) f (T , , DP ,Wob , QF ) (16)
Ah Eˆ rs P0 Ah Wob
wˆ b (k 1) wb (k ) t k f (T , , DP ,Wob , QF ) (17)
The optimal values of [T , ,Wob , tk , , ] are obtained by minimizing the cost function
4. Simulation
The proposed optimization method is applied to follow the drilling trajectory shown in Figure 3.
This section of a well has a total measured length of 1318 m, and true vertical depth TVD= 254
meters. The section starts by a 30 degree inclination, and proceeds approximately in two continuous build
zones to reach almost horizontal drilling (87.9 degrees inclination). The simulated RSE is shown in Fig.3.
RSE is assumed to be unknown, but estimated based on the achieved ROP during the previous control
step.
6
Target Trajectory x 10
12
11
0
Rock specific energy J/m3
-50 10
-100
9
-150
8
-200
-250
7
-300
0
-50
6
0
-100
-200
-150
-200 -400 5
-250 -600 0 200 400 600 800 1000 1200 1400
-300 -800 Measured depth
Fig. 3. (Left) the target well bore trajectory. (Right) the RSE.
During simulation 1 is taken to be unit matrix. Two control strategies were simulated. In the first
case, 2 was taken to be zero, i.e., no cost for the drilling effort. The drilling time came to 11.3 hours. In
the second case the cost of drilling time considered. The Drilling time is reduced to 10.697 hours with
minor compromise in the tracking error. It was observed that increasing the cost of time beyond this value
did not appreciably reduce the drilling time, but compromises the tracking accuracy.
The minimization of the cost function in Eq. (18) is performed using Matlab constraint minimization
function “fmincon”, by the interior-point algorithm. The lower and upper bounds on the control values
are given in Table 1.
167-7
Table 1. Upper and lower bounds for the control parameters
Actual Trajectory
-50
-100
-150
-200
-250
-300
0
-50
0
-100
-150 -200
-200 -400
-250 -600
-300 -800
Table 2. Nomenclature
5. Conclusion
The paper presents an approach for real-time drilling optimization, which combines the conventional
drilling parameters as well as the directional steering control. The objective function compromises
between trajectory tracking accuracy, drilling effort, and drilling time. The optimization problem can be
solved subject to operations limits and constraints using constraint optimization techniques.
167-8
Acknowledgements
This work has been supported by National Science, Technology, and Innovation Projects under
contract number NSTIP 12-OIL3033-04. The authors would like also to thank King Fahd University of
Petroleum and Minerals for its support.
References
Baker, R. (2001). A primer of oil well drilling. 6th Ed., Publ. Petr oleum Extension Service, Austin
Texas.
Bayliss, M. and Matheus, J. (2008). “Directional drilling tool simulation and system design”. SAE Int J.
Mater. Manuf.
Bourgoyne, A.T., Chenevert, M.E., Millheimand, K.K., Young, F.S., “Applied Drilling
Engineering,”Society of Petroleum Engineering (SPE), 1986.
Bourgoyne A. T. and Young F. S.( 1974) A multiple regression approach to optimal drilling and
abnormal pressure detection. Society of Petroleum Engineers (SPE) Journal, pages 371–384, Aug.
Downton, G.C., (2007), Directional Drilling System Response and Stability, Control Applications CCA
2007.IEEE International Conference, pp.1543-1550, Singapore.
Tuna and Evren, M. Ozbayoglu ,” Real Time Optimization of Drilling Parameters during Drilling
Operations”, SPE Oil and Gas Conference and Exhibition, Mumbai, 20-22 January 2010.
Eustes, A.W., (2007). The Evolution of Automation in Drilling. SPE Annual Technical Conference,
California, Nov. 2007. SPE 111125.
Dunlop, J., Isangulov, R., Aldred, W., Arismendi Sanchez, H., Sanchez Flores, J.L., Alarcon Herdoiza, J.,
Belaskie, J., Luppens, J.C.,(2011). “Increased Rate of Penetration Through Automation”,
SPE/IADC, 139897.
Joshi, S. D. and Ding W., (1991), “The Cost Benefits of Horizontal Drilling,” American Gas Association,
Arlington, VA, USA.
Moses, A. Alum and Egbon, F., “Semi-Analytical Models on the Effect of Drilling Fluid Properties on
Rate of Penetration (ROP)”, Nigeria Annual International Conference and Exhibition,Society of
Petroleum Engineers, 30 July - 3 August 2011.
Rashidi, B., Hareland, G. and Nygaard.R, ”Real-Time Drill Bit Wear Prediction by Combining Rock
Energy and Drilling Strength Concepts,” Abu Dhabi international Petroleum Exhibition and
Conference, SPE117109, 2008.
Thonhauser, Gerhard, (2004), “Using Real-Time Data for Automated Drilling Performance Analysis”, Oil
Gas European Magazine 4/2004, OG 170-173.
167-9