Ping Yi
University of Houston,
5000 Gulf Freeway Bldg. 9,
Houston, TX 77204-0945
e-mail: pingyi1129@gmail.com
Aniket Kumar
Halliburton,
10200 Bellaire Blvd.,
Houston, TX 77072
e-mail: aniket.aniket@halliburton.com
Robeilo Samuel
Halliburton Fellow
Halliburton,
10200 Bellaire Blvd.,
Houston, TX 77072
e-mail: robello.samuel@halliburton.com
Realtime Rate of Penetration
Optimization Using the Shuffled
Frog Leaping Algorithm
The increasing complexities of wellbore geometry imply an increasing well cost. It has
become more important than ever to achieve an increased rate of penetration (ROP) and,
thus, reduced cost per foot. To achieve maximum ROP, an optimization of drilling param
eters is required as the well is drilled. While there are different optimization techniques,
there is no acceptable universal mathematical model that achieves maximum ROP accu
rately. Usually, conventional mathematical optimization techniques fail to accurately
predict optimal parameters owing to the complex nature of downhole conditions. To
account for these uncertainties, evolutionary-based algorithms can be used instead of
mathematical optimizations. To arrive at the optimum drilling parameters efficiently and
quickly, the metaheuristic evolutionary algorithm, called the ‘‘shuffled frog leaping algo
rithm, (SFLA) is used in this paper. It is a type of rising swarm-intelligence optimizer
that can optimize additional objectives, such as minimizing hydromechanical specific
energy. In this paper, realtime gamma ray data are used to compute values of rock
strength and bit-tooth wear. Variables used are weight on bit (WOB), bit rotation (N),
and flow rate (Q). Each variable represents a frog. The value of each frog is derived
based on the ROP models used individually or simultaneously through iteration. This
optimizer lets each frog (WOB, N, and Q) jump to the best value (ROP) automatically,
thus arriving at the near optimal solution. The method is also efficient in computing opti
mum drilling parameters for different formations in real time. The paper presents field
examples to predict and estimate the parameters and compares them to the actual real
time data. [DOI: 10.1115/1.4028696]
Introduction
It is well known that an increase in WOB and bit rotation will
result in an increase in ROP, assuming that the bit hydraulics level
is correct. In fact, if the drilling fluid flow rate is not high enough,
an increase in bit rotation and weigh on bit can result in decreased
ROP. Therefore, to compute the maximum ROP, the combination
of drilling parameters should be properly optimized. Various stud-
ies have been presented in literature over the years aiming to
apply different optimization methods to improve the overall dril-
ling performance.
Amadi [1] has analyzed the application of mechanical specific
energy optimization techniques to reduce costs by predicting the
optimum ROP using real-time drilling parameters. Monden and
Chia [2] have presented real-time drilling optimization and risk
mitigation methods by utilizing real-time operation support centres.
Mostofi et al. [3] have presented drilling optimization based on the
ROP model with the objective of minimizing the cost per foot for
Iranian oil fields. Rashidi et al. [4] have studied in the detail the
real-time bit wear optimization techniques using intelligent drilling
advisory system to try and improve the drilling performance.
Some authors have also focused on the underlying parameters
that affect the ROP like the rock strength, lithology changes, and
tooth wear during the drilling operation to improve performance.
Hoberock and Bratcher [5] have studied the dynamic differential
pressure effects for drilling of permeable formations and prove
that the appropriate differential pressure is a dynamic quantity
rather than a static value as assumed in current analyses. Ohno
et al. [6] have proposed practical methods to estimate rock strength
This paper was previously published at the 2014 SPE Intelligent Energy
Conference and Exhibition held in Utrecht. The Netherlands on 1-3 April, 2014.
Contributed by the Petroleum Division of ASME for publication in the Journal
of E nergy R esources T echnology . Manuscript received April 9, 2014; final
manuscript received July 7, 2014; published online October 21, 2014. Editor:
Hameed Metghalchi.
and tooth wear while drilling with roller cone bits. Nes et al. [7]
have presented a working methodology to minimize wellbore sta-
bility problems through an integrated rock mechanics analysis to
reduce the overall drilling time and improve performance. Reck-
mann et al. [8] have explored the use of dynamic measurements to
analyze formation changes while drilling and have successfully
used the data and models to simulate drilling dynamics.
However, one of the major drawbacks of these presently exist-
ing methods is that they are not able to provide optimization of
drilling parameters combined with dynamic real-time measure -
ments of the formation properties while drilling. This study pro-
poses a new optimization technique using the SFLA to overcome
these challenges. The proposed method utilizes the real-time
downhole gamma ray measurements to estimate rock properties
and uses them in analytical models for optimization on a real-time
basis to improve the ROP.
This is one of the first of its kind of optimization methods pro-
posed that utilizes real-time drilling data combined with engineer -
ing analysis to improve the drilling performance. The SFLA has
been successfully implemented before in various other industries
with very positive results. For example, Aghababa et al. [9] have
successfully applied a modified SFLA for robot optimal controller
design. However, this optimization technique has not been applied
before in the oil and gas industry for drilling engineering analysis.
In this method, like a mathematic model, all frogs are set at
the start points. Once optimizations are conducted, frogs can jump
toward different directions on their own. After the frogs have fin-
ished the loops of local explorations and global searches, the best
maximum ROP can be computed. In addition, at this moment, the
WOB, rotary speed, and flow rate are optimal, respectively.
SFLA
The SFLA is regarded as metaheuristic to implement an
informed heuristic search using a heuristic function to seek a
Journal of Energy Resources Technology Copyright © 2015 by ASME
MAY 2015, Vol. 137 / 032902-1