MATEC Web of Conferences 148, 16008 (2018)

https://doi.org/10.1051/matecconf/201814816008
ICoEV 2017

Suppression of Drill-String Stick-Slip Vibration

Vahid Vaziri1 ,⋆ , Marcin Kapitaniak1 ,⋆⋆ , and Marian Wiercigroch1 ,⋆⋆⋆

1
Centre for Applied Dynamic Research, University of Aberdeen, UK

Abstract. We investigate experimentally and numerically suppression of drill-string tor-

sional vibration while drilling by using a sliding mode control. The experiments are con-
ducted on the novel experimental drilling rig developed at the University of Aberdeen [1]
and using PDC commercial drill-bits and real rock-samples. A mathematical model of
the experimental setup which takes into account the dynamics of the drill-string and the
driving motor, is proposed. Then a sliding mode control method is employed to suppress
stick-slip oscillations. The experimental and numerical results considering a time delay
of the actuator are in a close agreement. Stick-slip vibration is eliminated and significant
reduction in vibration amplitude has been observed when using the controller.

1 Introduction nents of the drill-bit rock interactions [4]. Then,

a sliding surface is defined and consequently a
During a downhole drilling process excessive vi- sliding mode controller is developed which ac-
brations can occur, which in most cases have a commodates the parameters uncertainties. The
negative impact on the effectiveness of the pro- experimental and numerical results demonstrate
cess and the drilling equipment [2]. Such vibra- the predictive capabilities of the mathematical
tion may lead to an accelerated wear and pre- models and the power of the controller to elimi-
mature damage of the expensive drilling equip- nate the stick-slip vibrations in the given condi-
ment. Often different dynamic effects such as tions.
bit-bounce, stick–slip, forward and backward
whirls may appear in the drilling process. Re-
2 Drill-String Experimental Stand
cently several attempts have been made to repli-
cate these vibrations in academic laboratories. A novel experimental drilling facility has been
However, in most studies the cutting process is designed and built at the University of Ab-
simulated by a friction between two disks [3]. erdeen, capable of reproducing all major types
In this work we focus on suppression of of drill-string vibration [4–7]. It allows to in-
stick-slip phenomenon while drilling, which na- vestigate nonlinear behaviour between the drill-
ture is still not well understood and can lead to bit and the formation and to introduce differ-
a catastrophic failure of the drill-stings. There- ent control methods of vibration suppression.
fore, firstly, a 2-DOF lumped mass model is de- Unlike typical other research experimental aca-
veloped using the torque-on-bit curves, which demic facilities, our rig is configured in such a
capture both the frictional and cutting compo- way that the cutting process is undertaken us-
⋆ e-mail: vahid.vaziri@abdn.ac.uk
⋆⋆ e-mail: marcin.kapitaniak@abdn.ac.uk
⋆⋆⋆ e-mail: m.wiercigroch@abdn.ac.uk

© The Authors, published by EDP Sciences. This is an open access article distributed under the terms of the Creative Commons Attribution
License 4.0 (http://creativecommons.org/licenses/by/4.0/).
MATEC Web of Conferences 148, 16008 (2018) https://doi.org/10.1051/matecconf/201814816008
ICoEV 2017

ing real commercial drill-bits and rock samples. then the drill-pipe twist can be calculated accu-
The rig shown in Fig.1 is equipped with a va- rately. Furthermore, the card high performance
riety of different sensors and transducers. An- allows high frequency data sampling up to 30
gular positions are measured by two quadra- kHz. This card is controlled by a LabVIEW pro-
ture encoders having 500 pulse per revolution, gramme with custom built graphical interface al-
where axial motion of the drill-bit is captured lowing to monitor responses of the system and
by a P1010 position transducer attached to the to present time histories of variables and phase
Bottom Hole Assembly. Horizontal and vertical portraits in real-time. This program also sends
forces as well as torque coming from the bit to the command to the top AC motor through the
the rock are measured by a 4-component Kistler NI card and the ABB frequency convertor. This
dynamometer placed under the rock sample. All convertor can work in speed or torque control
the voltage signals are sent to a NI PCIe-6321 modes which control the velocity or torque pro-
data acquisition card which has multiple ana- vided by the motor. Different speed or torque
logue input (16-Bit, 250 kS/s) and output (900 nonlinear control methods can be implemented
kS/s) channels and four 32bit counters/timers. in the LabVIEW program such as a sliding mode
Accessing to 4 counters, allows us to precisely control in Fig.1.
synchronise two encoders. This is important as

top motor
(a) (b)
gearings
frequency convertor

top encoder frequency convertor

torque control

flexible shaft
Labview program

disk
interface

top speed BHA

lateral displacement
eddy current
probs
axial dispalacement
LVDT
bit speed encoder
drill-bit
rock sample

torque and weight on bit

load cell

Figure 1. (a) Schematic diagram of the experimental setup, (middle) photograph of the experimental rig and (b)
examples of rock samples and drill-bits used in the setup. Main components of the system are: sensors (top and
bottom encoders, eddy current probes, LVDT and 4-component load cell), electric motor, flexible or rigid shafts,
disks, the BHA, drill-bit and rock samples.

2
MATEC Web of Conferences 148, 16008 (2018) https://doi.org/10.1051/matecconf/201814816008
ICoEV 2017

(a) (b)
θ̇t , θ̇b [rad/s] 7

Tt

0 ct
motor
Jt
8 & gearing
T b [Nm]

θt

0 drill-pipe
17.5 t[s] 27.5

7 c, k
θ̇t , θ̇b [rad/s]

J BHA
0

8
drill-bit
T b [Nm]

θb
0
17.5 t[s] 27.5 Tb

Figure 2. (a) An example of stick-slip oscillations occurring in the experimental rig for Wb = 1.79 kN and
1.5 inch pre-buckled flexible shaft. The time histories of the angular velocities at the bit, θ̇b , and the surface, θ̇t
together with TOB obtained in experimental studies (top) and calculated using proposed model (bottom). (b) A
physical model of a 2-DOF lump mass torsional system. The viscous damping property of the motor and gearing
system and the visco-elasto properties of the pipe are given by ct , c and k, respectively. The reactive torque acting
on the system during drilling is represented by T b .

3 Mathematical Model and tion. The number of disks varies in those studies,
Sliding Model Control for example 2 disks [8], 3 disks [9], 4 disks [10]
and even 18 disks [11]. In the current work a
The most used class of models for capturing 2-disk model has been employed (see Fig.2(b))
uncoupled torsional vibration consists of sev- to model the CADR experimental drilling rig,
eral parallel disks, rotating around their common following the work presented in [8]. The vis-
axis and connected to each other by torsional cous damping property of the motor and gear-
spring and damper. Top disk in all these mod- ing system and the visco-elasto properties of the
els represents the rotary table and the bottom drill-pipe are given by ct , c and k, respectively.
disk represents the drill-bit. The bit-rock inter- The reactive torque acting on the system dur-
action is modeled by the velocity dependent re- ing drilling is represented by T b . In this model,
sistive torque acting on the bottom disk. This the frequency convertor is set in the torque con-
class of models has been widely employed in trol mode and the top velocity can be calculated
studies focusing in drill-string torsional vibra- from the equation of motion.

3
MATEC Web of Conferences 148, 16008 (2018) https://doi.org/10.1051/matecconf/201814816008
ICoEV 2017

(a) (b) (c)

7 7 7

3.5 3.5 3.5

θ̇b [rad/s]

θ̇b [rad/s]

θ̇b [rad/s]
0 0 0
−1 −1 −1
−0.5 0 2.5 -2.5 0 0.5 −1 0 1 2

θt − θb [rad] θt − θb [rad] θt − θb [rad]

(d) (e) (f)
9 7 9

4.5 3.5 4.5

θ̇b [rad/s]
θ̇b [rad/s]

θ̇b [rad/s]
0 0 0
−1 −1 −1
−0.5 0 2.5 −1 0 1 2 −0.5 0 2.5

θt − θb [rad] θt − θb [rad] θt − θb [rad]

Figure 3. Phase portraits of the drilling experiments using sliding-mode controller. The uncontrolled stick-slip
trajectories and the controlled limit cycles are shown in red and blue respectively. The controller achieves (a)
47.86% (b) 59.26% (c) 51.52% (d) 57.58% (e) 66.72% (f) 64.72% reduction in vibration.

A response is considered of the stick-slip the TOB formulation, with corresponding pa-
oscillations of the drill-bit, for WOB value of rameters as developed in the previous work [4].
Wb = 1.76 kN, and a 1.5 inch pre-buckled flexi- TOB recorded in the experiment and modelled
ble shaft and T t = 39.57 Nm torque generated by can be seen in Fig.2(b) (blue curves). There is
motor. Fig.2(a) presents its (top) experimental excellent agreement between the experimental
and (bottom) numerical time history. As can be observations and performed simulation.
seen, the top velocities in experiment and simu- After observing and modelling the stick-slip
lation (black curves) have sinusoidal vibrations, vibration, we focus on the suppression methods.
and the drill-bit velocities (red and green curves) To control stick-slip, one of the most commonly
experience stick-slip oscillations of almost con- encountered vibration while drilling for oil and
stant amplitude. The numerical results were ob- gas exploration, a sliding surface can be defined
tained using the model shown in Fig.2(b) with as
 t  t
parameters identified as k = 10.00 Nm/rad,
c = 0.005 Nms/rad, Jt = 13.93 kg m2 and s = (θ̇t −ωd ) +λ (θ̇t −ωd )dτ+λ (θ̇t − θ̇b )dτ,
0 0
ct = 11.38 Nms/rad. A TOB model is used for (1)

4
MATEC Web of Conferences 148, 16008 (2018) https://doi.org/10.1051/matecconf/201814816008
ICoEV 2017

where θt and θb are angular positions of the ro- a sliding-mode controller have been developed,
tary table and the drill-bit respectively, ωd is the which proved to predict accurately the responses
desired angular velocity and λ is a positive con- of the analyzed experimental setup and elimi-
trol parameter. Based on the defined surface, nate the stick-slip oscillations and reduce it to
then a new sliding-mode controller has been suc- torsional vibration in presence of delay in the ac-
cessfully designed and applied to lead the sys- tuator.
tem to the sliding surface (s=0), which leads the
system to its equilibrium (θ̇t = θ̇t = ωd ) [5].
In presence of the delay in the motor, the con- References
troller has been tested in several different cases. [1] M. Wiercigroch, Modelling and Analysis
Fig.3 presents phase portraits of the drilling ex- of BHA and Drill-string Vibrations R&D
periments using sliding-mode controller. The Project sponsored by the BG Group (2010)
uncontrolled stick-slip trajectories and the con- [2] P. Spanos, A. Chevallier, N. Politis,
trolled limit cycles are shown in red and blue M. Payne, The Shock and Vibration Digest
respectively. The controller achieves (a) 47.86%
35, 81 (2003)
(b) 59.26% (c) 51.52% (d) 57.58% (e) 66.72%
[3] P.A. Patil, C. Teodoriu, Journal of
(f) 64.72% reduction in vibration.
Petroleum Science and Engineering 112,
227 (2013)
4 Conclusion [4] M. Kapitaniak, V. Vaziri, J. Páez Chávez,
K. Nandakumar, M. Wiercigroch, Interna-
In this paper we have presented a lumped mass tional Journal of Mechanical Sciences 101-
model for the drilling rig shown in Fig.1. By
102, 324 (2015)
using identified model parameters in the pre-
[5] V. Vaziri, Ph.D. thesis, University of Ab-
formed simulation, an excellent match between
erdeen (2015)
experiment and simulation is achieved captur-
[6] M. Kapitaniak, Ph.D. thesis, University of
ing the stick-slip vibration. The next step of
this study was to adapt a sliding-mode control Aberdeen (2015)
method and applied to the obtained model in or- [7] M. Kapitaniak, V. Vaziri, J. Páez Chávez,
der to eliminate the stick-slip vibration observed M. Wiercigroch, Mechanical Systems and
in the drilling rig and simulation. The controller Signal Processing 100, 454 (2018)
implemented in the experiment is successful in [8] E. Navarro-López, R. Suárez, IEEE Con-
eliminating the stick-slip vibration in the pres- ference on Control Applications pp. 1454–
ence of a delay and dead-zone in the actuator. 1460 (September 2004)
Several experimental results were presented in [9] E. Navarro-López, E. Licéaga-Castro,
this paper with a variety of the estimated param- Chaos, Solitons and Fractals 41, 2035
eters applied to the controller. A significant re- (2009)
duction in vibration amplitude is observed when [10] Y. Liu, IET Control Theory & Applications
the controller is applied. 9, 91 (2015)
Taking into consideration the positive exper- [11] T. Vromen, C.H. Dai, N. Van De Wouw,
imental results reported in this paper, we can T. Oomen, P. Astrid, H. Nijmeijer, IFAC
conclude that a robust mathematical model and Proceedings Volumes 48, 266 (2015)