Artificial Lift Selection Methods in Conventional

See discussions, stats, and author profiles for this publication at: https://www.researchgate.
net/publication/379719724
Artiﬁcial Lift Selection Methods in Conventional and Unconventional Wells: A

Summary and Review from Old Techniques to Machine Learning Applications
Article in International Journal of Innovative Science and Research Technology · April 2024
DOI: 10.38124/ijisrt/IJISRT24MAR2108
CITATION READS
1 87
3 authors:
Mohaned Mahdi Mohamed Amish

Robert Gordon University Robert Gordon University
3 PUBLICATIONS 1 CITATION 20 PUBLICATIONS 20 CITATIONS
SEE PROFILE SEE PROFILE
Gbenga Oluyemi
Robert Gordon University
101 PUBLICATIONS 561 CITATIONS
SEE PROFILE
All content following this page was uploaded by Mohaned Mahdi on 27 April 2024.
The user has requested enhancement of the downloaded file.

Volume 9, Issue 3, March – 2024 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165 https://doi.org/10.38124/ijisrt/IJISRT24MAR2108
Artificial Lift Selection Methods in Conventional and

Unconventional Wells: A Summary and Review from
Old Techniques to Machine Learning Applications
Mohaned Alhaj A. Mahdi1,2; M. Amish1, G. Oluyemi1
¹School of Engineering, Garthdee Road, Robert Gordon University, Aberdeen, AB10 7GJ, UK.
²SLB
Abstract:- Artificial lift (AL) selection is an important II. AL SELECTION IN CONVENTIONALS

process in enhancing oil and gas production from
reservoirs. This article explores the old and current states At the early 1980s, Neely et al. (1981) summarised the
of AL selection in conventional and unconventional wells, criteria for selecting four methods of lifting, namely gas lift
identifying the challenges faced in the process. The role of (GL), sucker rod pump (SRP), electrical submersible pump
various factors such as production and reservoir data and (ESP), and hydraulic pump (HP), by examining their
economic and environmental considerations is advantages and disadvantages in relation to reservoir and well
highlighted. The article also examines the use of machine properties. They found that SRPs are suitable for low volumes
learning (ML) techniques in the AL selection process, but not recommended for offshore or residential areas, or
emphasising their potential to increase the accuracy of wells prone to sand production. Continuous gas lift (CGL) is
selection and reduce data analysis time. The findings of suitable for high volumes, high bottom hole pressure (BHP),
this article provide valuable insights for researchers and and handling solids and sand; however, it is limited by back
practitioners in the oil and gas industry, as well as for pressure and high costs. Intermittent gas lift (IGL) is less
those interested in the development of AL selection expensive than continuous gas lift (CGL) but yields lower
methods. volumes. ESP is suitable for high volumes and confined
spaces, such as offshore platforms, and can tolerate deviations
Keywords:- Artificial Lift, Selection, Conventionals, of up to 80°. However, major drawbacks of ESP include sand
Unconventionals, Machine Learning. production, workover costs, and inefficiency at rates below
150 B/D. HPs (piston pump (HPP) and jet pump (HJP)) are
I. INTRODUCTION suitable for deep wells and can deliver up to 17,000 B/D.
HJPs are effective for sand production due to their lack of
The selection of the optimum Artificial Lift (AL) to moving parts, while HPPs are efficient with highly viscous
achieve the highest recovery is a real challenge in the fluids; however, they require more maintenance and have a
petroleum industry. Optimal selection is a requirement for shorter lifespan compared to jets and submersibles. HJP
obtaining the maximum profit from an oil well (Bucaram and cannot operate at BHP below 1000 psi, whereas HPP can
Patterson 1994). Several factors determine the selection operate at 0 psi. The aforementioned criteria and findings bear
process: depth, rates, reservoir and fluid properties, initial and resemblance to the AL selection decision tree outlined by
operating cost, and geographical and environmental aspects. Heinze et al. (1995), where 50% of the tree's selection criteria
Special AL selection techniques are required to cope with relied on productivity index (PI) and the Inflow performance
different reservoir, well, and field conditions, for instance: relationship (IPR). Brown (1982) introduced a selection
high-viscosity oil, high water cut, sand, gas, low reservoir methodology aligned with AL to aid engineers in their AL
pressures, high temperatures, low-productivity wells, surface selection process. In 1986, Blais (1986) developed selection
facilities, as well as human interference. Historically, the AL charts in 1986 to delineate the operational parameters for AL
selection process generally begins by studying the advantages methods. These charts served as a prominent reference for
and disadvantages of each method. Then the elimination selection during that period, alongside elementary computer
depends on the engineers’ decision based on their analysis of programs employed as supplementary tools. Table 1 provides
the AL record, field data availability, and failure history. a concise overview of AL selection methodologies
Since these factors change over time, the AL design for documented in the literature.
current production conditions without considering future
production results in high inconstancy rates and fluctuations It is notable that inadequate selection of AL selection
in lifting selection (JPT staff 2014; Lea and Nickens 1999). can lead to frequent replacements within a short timeframe,
The following sections explore old and recent selection resulting in reduced profits and heightened operational costs.
criteria in conventional and unconventional wells It was not until Clegg et al., (1993) unveiled comprehensive
(conventionals and unconventionals) from the literature and reference selection tables and design considerations, which
the various techniques used by the engineers and factors compared seven methods: SRP, PCP, ESP, HPP, HJP, GL and
considered. Plunger, across 31 parameters. These tables represent a
comprehensive selection framework that has been
IJISRT24MAR2108 www.ijisrt.com 2342

consistently utilised by numerous researchers to date, albeit it is crucial to ensure compatibility with existing surface
with minor adjustments and the integration of software tools. production facilities to avoid additional expenses associated
They serve as the cornerstone of many contemporary AL with installing new flowlines and wellhead fittings.
selection methodologies. Bucaram and Patterson (1994) Furthermore, they provided an illustrative example of the
proposed a selection criterion that accounted for factors such selection process for SRPs and outlined various factors to be
as well location, capital expenditures (CAPEX), operating taken into account. It became evident that SRPs were
expenditures (OPEX), production rates, run life, and failure unsuitable for use in gassy and deep wells. The paramount
rates, in addition to fundamental well and reservoir selection criterion revolves around striking a balance between
characteristics including depth, bottom hole pressure (BHP), AL reliability, the desired production rate, and current
gas presence, sand, and solids content. They also highlighted constraints to ensure smooth operation of the pump over an
the significance of considering laterally drilled wells in extended period.
mature fields. When selecting an AL method for a new well,
Table 1: AL Selection Techniques in the Literature

References Selection Criteria No of Screened Analysed Parameters Remarks
AL
Neely et al. 1981 Limitations, IPR 4 (SRP, GL, ESP, Location, flowrate, reservoir and Efficiency is the most
HP) fluid properties, gas, sand, paraffin, important factor to
scale, cost, deviation, skilled consider. Cost for GL.
operators ESP not adoptable for
rate < 150B/D
Brown 1982 Advantages and 6 (SRP, GL, ESP, Flow rate, depth limitations, reservoir Flow rate is the most
disadvantages HPP, HJP, and fluid properties, pressure loss important factor to
tables, elimination Plunger) across the system, location, sand, gas, consider
guidelines, IPR vs paraffin, corrosion, scale, cost, run
TPR life.
Blais 1986 Operating range 7 (SRP, PCP, Depth vs. flow rate -
charts GL, ESP, HPP,
HJP, Plunger)
Clegg et al. 1993 Comprehensive 7 (SRP, PCP, 31 factors among these attributes The crucial
advantages and GL, ESP, HPP, (location, reservoir and fluid consideration is to
disadvantages HJP, Plunger) properties, flow rate, depth sustain the desired rate
tables, design limitations, CAPEX, OPEX, over AL life cycle at
considerations temperature limitations, completion, the minimum OPEX.
efficiency, reliability, flexibility, Field personnel must
system, salvage value, usage, intake, be trained for
noise, surveillance, life cycle, gas, successful operation.
sand/solids, paraffin, corrosion, scale,
deviation)
Bucaram and Guidelines 1 (SRP) Location, CAPEX, OPEX, reliability, Flow rate and
Patterson 1994 flow rate, operating conditions reliability are the most
(casing size, depth, intake capability, critical factors
BHP), gas, sand, paraffin, corrosion,
scale
Espin et al. 1994 Expert systems 10 (Espin) (SRP, Espin [(Quantitative data (well and Espin (some high-
(computer ESP, PCP, HP, reservoir props), qualitative data ranked AL eliminated
Heinze et al. programs for Plunger, CGL, (engineer experience and well because they were not
1989 ranking and IGL, IGL with geographic), production problems economically
eliminating AL Plunger, constant (corrosion, paraffin, sand, gas) and feasible), (SEDLA
Valentin and from least to most slug injection economic evaluation)] software used)
Hofmann 1988 recommended) GL, chamber GL)
Heinze (flow rate, casing size, Valentin (no economic
4 (Heinze) (SRP, deviation, sand, paraffin, scale, data provided for GL),
GL, HP, ESP) corrosion) (OPUS software used)
6 (Valentin) Valentin [well data (location, depth,

(SRP, PCP, ESP, deviation, P/T gradient flow rate),
GL, HPP, HJP) technical data (pump type, viscosity),
economic data]

Heinze et al. Decision tree 7 (SRP, PCP, Economics [CAPEX, OPEX (repair, IPR eliminates 50% of
1995 ESP, GL, HPP, maintenance, replacement, energy, AL. Important
HJP, Plunger) personnel)], depth vs. rate, reservoir considerations are
characteristics, location CAPEX and OPEX
Lea and Nickens Advantages and 5 (SRP, GL, ESP, Location, reservoir and fluid ESPCP is offshore
1999 disadvantages, ESPCP, HP) properties, flow rate, depth recommended for
Blais charts, limitations, AL and equipment run feasible pulling out of
selection by net life, economics (NPV, total cost) failed wire-lined
present value pump.
(NPV) and
feasibility map
Naguib et al. Reservoir 4 (SRP, ESP, GL, Reservoir and fluid properties Economic evaluation
2000 simulation, well HJP) (WC%, GOR, wax content), flow is the primary
performance rate, CAPEX, OPEX, run life, failure elimination factor
analysis, economic rate and workover.
evaluation
Lanier and Ranking matrix 6 (Lanier) (GL, Lanier screening for thermal EOR Lanier (SRP selected
Mahoney 2009 (high, medium, to ESP, SRP, PCP, [reservoir and fluid properties with attached sinker
low recommended) HJP, long-stroke (temperature, viscosity, API˚, WC%, bars to reduce rod
Williams et al. RP) GOR), flow rate, CAPEX, OPEX, buckling)
2008 maintenance, surveillance].
5 (William) William (flow rate is
(ESP, PCP, GL, William (flow rate, depth, BHP, the primary
HJP, SRP) casing size, deviation, gas, elimination factor)
sand/solids)
Mali and Al- Company selection 7 (PCP, SRP, Screening for CHOPS and thermal CAPEX and OPEX
Jasmi 2014 tables ESP, ESPCP, EOR (reservoir and fluid properties, are the critical
MTMPCP, HJP, depth, flow rate, gas, sand, high consideration
GL) temperature, efficiency, CAPEX,
OPEX, run life in vertical and
horizontal wells)
Kaplan and Analysis of axial 4 (SRP, PCP, Screening for heavy oil with CO2 Blending light oil with
Duygu 2014 and radial shear ESP, ESPCP) injection (viscosity, temperature, heavy oil reduces the
stress, and the fluid velocity, rod string length, power required.
torque on the number of strokes, stroke length, rod PCP is more efficient
required power radius, rotation speed, tubing size, and require less
power consumption, pump size) horsepower than SRP
for emulsified heavy
oil
Caicedo et al. Preliminary 6 (SRP, PCP, Screening for high uncertainty Environmental aspects
2015 screening, nodal GL, ESP, HJP, reservoir [location, flow rate, depth, determine the
analysis, reservoir Nitrogen lift) reservoir and fluid properties (PI, selection if the field is
simulation bubble point pressure, pressure, close to urban areas.
temperature, GOR, WC%, viscosity), High uncertainty and
casing and tubing size, high H2S and complex reservoirs
CO2, corrosion, power source, complicate AL
economics, environmental aspects. selection
Kefford and Nodal analysis 7 (SRP, PCP, Consistency, audibility, efficiency, Designed flow rate is
Gaurav 2016 GL, ESP, HPP, technical rigour, vendor the primary selection
HJP, HSP) independence. factor
Ounsakul et al. Machine Learning 4 (SRP, PCP, 17 out of 50 factors in these attributes The selected AL has
2019 ESP, GL) (reservoir and fluid properties, low cost per barrel.
pressures and temperatures, depth,
flow rate, API˚, GOR, gas, sand, cost,
run life)
Hoy et al. 2020 Assessment matrix 6 (GL, SRP, ESP, Screening for polymer EOR 500 ppm polymer
PCP, ESPCP, (reservoir and fluid properties, PI, concentration is a
HJP) depth, deviation, flow rate, polymer constraint for ESP.
concentration, sand, corrosion,

maturity, robustness (efficiency),

reliability)
Zein El Din Weatherford 8 (SRP, PCP, Max flow rate, max depth, max PCP can handle
Shoukry et al. selection tables GL, ESP, HPP, temperature, API˚, gas, solids, viscosity up to 100000
2020 HJP, Plunger, corrosion, maintenance, power cp. Have lower
Foam Lift) source, location, efficiency. CAPEX and OPEX.
Simple to operate
Crnogorac et al. Fuzzy logic 5 (SRP, PCP, 14 factors (power source, automation, AL of a new well is
2020 GL, ESP, HJP) maintenance, flow rate, depth, the one that best
temperature, fluid density and matched to AL
viscosity, deviation, corrosion, solids, database
paraffin, GOR, WC%)
Adam et al. 2022 Decision-making 5 (Adam) (SRP, Adam 15 factors among these Adam (flow rate and
approaches PCP, GL, ESP, attributes (reservoir and fluid depth are the essential
Alemi et al. (TOPSIS model, HJP) properties, flow rate, depth, flow line consideration)
2010, 2011 ELECTRE model) pressure and temperature, sand,
mathematical 5 (Fatahi and corrosion, contaminants, recovery, Fatahi compared their
Fatahi et al. models for ranking Alemi) (SRP, location, power source) results with
2011, 2012 and elimination PCP, GL, ESP, Schlumberger
HJP) Fatahi and Alemi 25 factors among standard AL selection
these attributes (reservoir and fluid tables
properties, flow rate, depth,
completion, casing size, deviation,
sand, corrosion, contaminants,
recovery, stability, location, power
source, service, stimulation)
Mahdi et al. 2023 Machine learning 4 (SRP, PCP, 9 production parameters Gas and cumulative
GL, ESP) + production are the
natural flow critical factors
Unconventionals
Khan et al. 2014 Selection in GL, PCP, ESP, Drilling conditions, depth, casing Economic evaluation
unconventional by ESPCP, SRP, size, reservoir and fluid properties is the critical selection
Oyewole 2016 analysing field Jets, Plunger, (porosity, permeability, saturation, factor in
conditions using Foam, GOR, GLR, pressures, temperatures), unconventionals
Valbuena et al. simulation, IPR, Compression, flow rate, depletion period, gas, sand,
2016 nodal analysis, Velocity Strings solids, surface facilities, pump size,
NPV CAPEX, OPEX, well-integrity, run
Liu and Zerpa life.
2016
Kefford and
Gaurav 2016
Escobar Patron et
al. 2018
Chow et al. 2020
Lane and
Chokshi 2016
Temizel et al.
2020
A. Computer Programming and Nodal Analysis Applications techniques. This program utilised field data categorised into
The advent of computer programming and simulation three main groups: (1) quantitative data encompassing well
tools for AL selection emerged during the late 1980s and and reservoir properties, (2) qualitative data including
early 1990s. Espin et al., (1994) pioneered the development engineer expertise and well geographical considerations, and
of a coding program designed to assist engineers in selecting (3) production challenges such as corrosion, paraffin
the appropriate AL method from a pool of 10 lifting deposition, sand influx, and gas interference, alongside

economic evaluations. The program ranked lifting methods Matondang et al. (2011) introduced an alternative
on a scale ranging from 1 (least recommended) to 5 (most method for AL selection combining GL mandrels and ESP.
recommended). Despite some lifting methods receiving high Initially, this approach aimed to alleviate gas-related issues
scores, they were disregarded due to their lack of economic affecting the pump by venting gas through the casing and
feasibility, and instead, lower-ranked methods were adopted. subsequently reducing water cut (WC%). The
The same methodology was employed by (Heinze et al. 1989; implementation proved successful, with gas released through
Valentin and Hofmann 1988). Other ranking computer the mandrels merging with the gas expelled from the ESP gas
programs and mathematical models were also applied by separator, resulting in a production increase from 350 to 500
(Alemi et al. 2010, 2011; Fatahi et al. 2011, 2012). Lanier and B/D and a reduction in water production. This technique
Mahoney (2009) employed a ranking matrix to assess six facilitated the adoption of ESP in high GOR wells, contingent
lifting methods: GL, ESP, SRP, PCP, HJP, and long-stroke upon the suitability of well completion for this hybrid
pump. They scrutinized technical and operational constraints application. In a separate study, Zulkapli et al. (2014)
associated with AL, along with CAPEX and OPEX, within assessed ESP production in the Bokor offshore field in
the context of a thermally recovered heavy oil reservoir Malaysia following the replacement of a dual string GL
situated in Oman, aiming to enhance production rates. system, prompted by escalating water production and
Notably, the study revealed that elevated temperatures posed inadequate gas supply. They employed nodal analysis using
significant challenges for both GL and ESP methods. GL was the commercial simulator PIPESIM to simulate performance.
found unsatisfactory due to the costly gas supply and low Despite the theoretical feasibility of GL, their findings
GOR. Meanwhile, the high operational and capital expenses revealed deficiencies in gas supply from a nearby field and
rendered HJP and ESP unfeasible options. Although Metal- complications with compressor functionality, impairing the
to-Metal PCP (MTMPCP) boasted economical operational efficiency of GL. Additionally, discrepancies in real-time
costs, it was disqualified due to design rate limitations, measurements impeded the optimization process, leading to
susceptibility to sand production, and a history of failures, misinterpretation—a widespread issue globally.
much like the long-stroke pump. Furthermore, Jet was Consequently, ESP was chosen due to the low GOR of the
excluded due to its substantial power requirement for lifting wells and absence of historical sand production. Rather than
fluids with a density below 14 API, coupled with limited discontinuing GL, Alshmakhy et al. (2020, 2019) pursued a
historical data on pump performance in the field. Ultimately, novel approach to enhance its efficiency. They introduced
all attempts to introduce new AL alternatives proved digital optimization techniques for both single and dual-string
unsuccessful, prompting the continued utilization of the GL systems within an onshore field in the UAE, aiming to
primary SRP, albeit with modifications such as attached mitigate common challenges such as casing pressure
sinker bars to mitigate rod buckling. Similarly, Williams et al. instability, temperature fluctuations, and injection rate
(2008) employed a matrix screening approach to optimise control issues. The implementation involved deploying a
five lifting methods: ESP, PCP, GL, HJP, and SRP in a Digital Intelligent Artificial Lift (DIAL) system, featuring up
Colombian oil field, addressing prevalent challenges such as to six injection orifices and an electric cable connected to the
depth, gas presence, and solid content, each of which mandrels, enabling remote control of GL orifice operations
influences the efficacy of the method. The selection process from the surface. Additionally, the system facilitated real-
was refined by utilizing flow rates spanning from 0 to 750 time monitoring of pressure and temperature, with an
B/D as a discriminating factor. Their findings indicated GL anticipated 20% increase in oil production. The potential
as suitable across all flow rate ranges, whereas PCP was efficacy of this technology appears promising, particularly if
deemed effective for productions up to 300 B/D. SRP and deployed offshore, where the cost associated with workovers
ESP were identified as optimal choices for flow rates ranging is substantially higher. In a similar vein, Caicedo et al. (2015)
between 300 and 750 B/D, while GL and ESP were preferred conducted nodal analysis to ascertain the most suitable AL
for flow rates exceeding 750 B/D. Another investigation method for a high-uncertainty, large reserve field in Abu
conducted by Naguib et al. (2000) focused on four AL Dhabi, presuming no natural flow. Given the presence of H₂S
methods; SRP, ESP, GL, and HJP in an Egyptian oil field. and the proximity to residential areas, safety considerations
The study entailed reservoir simulation and well performance predominated the selection process to avert potential leaks
analysis to ascertain the optimal AL method. SRP and HJP that could jeopardize human safety. Following analysis
were discounted due to their incompatibility with high accounting for various GOR, Water Cut (WC%), and
reservoir volumes and wax content. Ultimately, GL was reservoir pressure values, it was determined that AL would
selected due to the convenience of gas supply availability be necessary if the reservoir pressure dipped below 2500 psi,
from a nearby company, while ESP was chosen to regulate WC% exceeded 90%, and GOR remained below 3000
flow rates, particularly concerning the presence of high scf/STB. SRP and PCP were eliminated due to the risk of
associated gas, albeit with the intention of installing a stuffing box leakage, while GL was not feasible due to
downhole gas separator. Subsequently, a further evaluation inadequate gas supply. Finally, ESP was selected, with
was conducted between the two remaining candidates. GL particular attention to ensuring that GOR did not surpass 1500
exhibited lower CAPEX and OPEX compared to ESP, scf/STB. Kefford and Gaurav (2016) undertook an
alongside a superior recovery factor. However, ESP assessment of multiple lifting methods, employing adjusted
demonstrated advantages in scenarios involving high correlations and iterative calculations. Their study
production rates, increasing water cut, and inadequate gas encompassed an analysis of specific reservoir attributes and
supply. operational factors across three fields, including
unconventional reservoirs, with the aim of estimating

production rates and assessing AL capacity to manage with multiple stages to convert kinetic energy into potential
associated gas. The objective was to broaden the selection energy, the PVP exhibited an efficiency approximately 50%
methodology and introduce a novel criterion instead of the lower than ESP, reaching up to 33%. However, it
conventional Blais method. Nodal analysis was utilised to demonstrated superior performance in handling additional
compute well performance, determining the AL methods head compared to ESP. Performance tests of the PVP
capable of achieving both maximum and targeted production indicated that a single stage operating at 60Hz could generate
rates, while considering factors such as gas handling, a head of 75.5, equivalent to 32.7 psi. Following installation
wellhead pressure, and power requirements. Alferov et al. for a pilot test onshore, the pump achieved a positive
(2015) and Khabibullin and Krasnov (2015) explored the efficiency of 6-8% after four months of operation, determined
impact of varying parameters: reservoir pressure, BHP, by the difference between consumed and delivered power.
WC%, PI, GOR, and flowline pressure on the CAPEX and
OPEX of AL methods in Russian fields. Alferov et al. (2015) In a distinct approach, Kaplan and Duygu (2014)
argued against the practicality of relying on outdated investigated a selection strategy in a Turkish heavy oil field
selection technical tables, asserting that such tables are undergoing CO₂ injection for enhanced oil recovery (EOR).
primarily derived from the operational history of AL under Their analysis focused on axial and radial shear stress, as well
diverse field conditions. Their case study examined the field as torque requirements for two AL methods: Beam Pumping
implementation of Simultaneous Water Alternating Gas Unit (BPU) and PCP. Due to elevated temperatures, ESP was
(WAG) in a low-permeability, heterogeneous reservoir deemed unsuitable. Although BPU had been effectively
characterized by paraffin, salt, and corrosion. SRP, PCP and producing oil, issues such as emulsion formation and high
HJP were eliminated due to inadequate equipment viscosity led to rod failures and restricted oil production
availability. The most viable and cost-effective AL options volumes. Comparative analysis revealed that the power
identified for the field development plan (FDP) were ESP and required to handle radial shear stress and torque for PCP was
GL, owing to their respective capabilities in managing lower than that required to manage axial shear stress for BPU.
fluctuations in WC% and GOR, respectively. Khabibullin and This reduction in power could be attained by adjusting the
Krasnov (2015) AL selection map for a new field revealed RPM and employing a larger pump. Consequently, PCP was
comparable results for ESP and GL at BHP of 100 atm BHP, selected to replace the SRP in the field. Mali and Al-Jasmi
highlighting a preference for 40 atm for optimal applicability. (2014) implemented a selection screening process for cold
heavy oil production with sand (CHOPS) and CSS thermal
B. Other AL Selection Experiments recovery techniques in a Kuwaiti oil field. The FDP targeted
Fraga et al. (2020) introduced a new pump system, a maximum production of 300 B/D for cold oil and 1000 B/D
termed the progressive vortex pump (PVP), which combines for hot oil, under conditions of 12 API density, GOR, and well
PCP and ESP mechanisms. Developed by Petrobras, the PVP depths up to 3000 ft. Seven AL candidates were assessed,
aims to optimize production and address the challenges posed including SRP, ESP, Electrical Submersible PCP (ESPCP),
by high temperatures in cyclic steam stimulation (CSS) and GL, HJP, PCP, and MTMPCP. Selection criteria are detailed
steam flooding (SF) operations, while also accommodating in Table 2. Ultimately, PCP and MTMPCP were chosen for
varying flow rates. Comprising a rotor, stator, and diffuser their lower CAPEX and OPEX.
Table 2: AL Comparison for Heavy Oil Production (Mali and Al-Jasmi 2014)
Parameters SRP ESP PCP Jet ESPCP GL
Capital Cost Low High Low High Moderate High
Operating Cost Low Moderate Low High Moderate Moderate
Run life in vertical wells Average Average Average High Average High
Run life in horizontal Low Average Low High Average High
wells
Ability to handle sand Average Low Average Good Average Average
content
Efficiency Average Low Average Low Average Average
Suitability for thermal Applicable Applicable Applicable Applicable Not Applicable
production Applicable
Operational Flexibility Average Good Good Low Average Good
Ability to handle gas Average Good Good Good Good Good
content
Production Handling Good Average Good Average Average Good
Capacity
Hoy et al. (2020) conducted an evaluation of existing ESP, and PCP, aiming to determine their reliability in
lifting methods with the aim of selecting an appropriate achieving the desired flow rate. Their findings indicated that
system for a polymer EOR application within an Austrian oil ESP and SRP emerged as the optimal lifting methods.
field. Their investigation focused on assessing the impact of However, while ESP demonstrated capability in managing
viscosity changes and head column variations on GL, SRP, fluid head, it proved inadequate in handling a polymer

concentration of 500 ppm. Conversely, SRP exhibited some for achieving optimal AL selection (Table 3) with the main
friction-related issues. In a complementary study, Zein El Din goal of extending run life and maximising revenue.
Shoukry et al. (2020) delineated a set of parameters crucial
Table 3: AL Selection Parameters (Zein El Din Shoukry et al. 2020)

AL Gas Lift Foam Plunger Rod Lift PCP ESP HJP HPP
Lift
Max 18,000 ft 22,000 ft 19,000 ft 16,000 ft <9,000 ft 15,000 ft 20,000 ft 17,000 ft
Depth
Max 75,000 B/D 500 B/D 200 B/D 6,000 B/D 5,000 B/D 60,000 B/D 35000 B/D 8,000 B/D
Volume
Max 450°F 400°F 550°F 550°F 302°F 482°F 550°F 550°F
Temp
Corrosion Good to Excellent Excellen Good to Good Good Excellent Good
Handling excellent t excellent
Gas Excellent Excellent Excellen Fair to Good Fair Good Fair
Handling t good
Solids Good Good Fair Fair to Excellent Sand<40ppm Good Fair
Handling good
Fluid >15° >8° >15° >8° 8°<API<45 Viscosity<400 ≥6° >8°
Gravity ° cp
(°API)
Servicing Wireline or Capillary Wellhea Workover Wireline or Wireline or Hydraulic Hydraulic
workover unit d or workover workover rig or wireline or wireline
rig catcher pulling rig rig
or
wireline
Prime Compresso Well Well Gas or Gas or Electric Gas or Gas or
Mover r natural natural electric electric electric electric
energy energy
Offshore Excellent Good N/A Limited Good Excellent Excellent Good
System 10% to N/A N/A 45% to 55% to 35% to 60% 10% to 45% to
Efficiency 30% 60% 75% 30% 55%
Crnogorac et al. (2020) presented a study aimed at III. AL SELECTION IN UNCONVENTIONALS

selecting the optimal AL method through the application of
fuzzy logic and mathematical models. The model's In unconventionals application, the most used ALs are
effectiveness is contingent upon a predefined dataset ESP, GL, SRP, Jets, and Plunger lift (Table 4). The
encompassing five lifting methods and may not be universally installation of AL occurs either subsequent to the decline in
applicable when different input parameters or alternative AL natural well flow or at the onset of production (Chow et al.
techniques are utilised. The selection of AL for a new well is 2020). The typical operational lifespan of ESP systems ranges
based on aligning the characteristics of the prospective AL between 6 to 9 months. An essential consideration for SRP
with those stored in the AL database. (Adam et al., 2022) pertains to maintaining side loads within the range of 200
introduced a novel selection methodology tailored for lbf/25 ft; deviations beyond this limit necessitate exploring
Sudanese oil fields. Their decision-making model adapted the alternative AL methods. GL systems are suitable for use
approach proposed by (Alemi et al., 2010), incorporating within deviations of up to 75°. Jets, distinguished by their
TOPSIS (Technique for Order Preference by Similarity to the absence of moving components, exhibit an adeptness in
Ideal Solution) and integrating Analytic Hierarchy Process handling solids. Conversely, Plunger Lift systems are
(AHP) for parameter weighting, with the aim of facilitating deployed to manage lower production rates, typically around
informed decision-making. By considering the desired flow 200 STB/D (Kolawole et al. 2019; Pankaj et al. 2018).
rate and other pertinent parameters, the model ranked the
most suitable AL method. However, the study suggests that Table 4: AL used in Unconventionals (Kolawole et al. 2019)
incorporating economic evaluations into the decision-making AL Percentage of Application
process could enhance the robustness of the results. GL 40%
ESP 36%
SRP 13%
Jets 4%
Plunger 7%

The rapid decline in production rates observed in through NPV analysis. This methodology was applied in a
unconventionals over a short span of time, often spanning just field case study. Ultimately, the study underscored the
a few years or even months, poses a significant challenge paramount importance of economic evaluation in guiding the
necessitating the replacement of AL and an increase in AL selection process. Oyewole (2016 presented a field case
OPEX. In the context of unconventional reservoirs, the size study aimed at selecting an appropriate AL system capable of
of the casing plays a pivotal role in the design and selection managing the rapid decline in production. The selection
of AL systems. A larger casing diameter results in higher gas process was structured into four distinct categories: (1)
production through the annulus to the surface, thereby technical considerations, encompassing production rates and
influencing the performance of AL (Parshall 2013). Recent associated gas production to ascertain depletion periods; (2)
advancements in AL technology aim to enhance performance reservoir/fluid properties and drilling conditions; (3) surface
in unconventionals. Notable developments include the facilities; and (4) economic evaluation. Regardless of the
integration of permanent magnet motors and optimized stage recommended AL methods, economic factors emerged as the
designs for ESP, the implementation of controlled valves in primary selection criterion. In a separate study, Liu and Zerpa
GL, tailpipe designs tailored to manage slugging, and the (2016) conducted a cost analysis of AL methods (Table 5) to
introduction of the Geared Centrifugal Pump (GCP), which identify a suitable approach for a hydrate reservoir in Alaska
operates akin to an ESP albeit with surface-driven rods characterized by low pressure, low Gas-Liquid Ratio (GLR),
powered by hydraulic and electric sources, offering enhanced low reservoir and surface temperatures, and sand production
suitability for gas-rich environments compared to issues. The authors determined PCP to be a viable option;
conventional ESPs (Parshall 2013; Stephenson 2020). however, its inability to manage sand production led to
Valbuena et al. (2016) introduced a methodological premature failure. Moreover, the high CAPEX associated
framework for selecting suitable AL systems in horizontal gas with ESP and the low GLR posed challenges for GL
wells, incorporating technical and economic considerations. suitability. In addition to CAPEX, (Khan et al. 2014)
The technical screening process evaluated the constraints of incorporated various factors, including workover costs,
various lifting methods in terms of production rates, depth OPEX, oil prices, oil treatment and transportation expenses,
versus rate, reservoir/fluid properties, and gas handling, along with maximum NPV, into their selection strategy for
drawing upon standard selection tables and charts. four AL methods; GL, ESP, ESPCP, SRP for Shale play
Subsequently, the feasibility of lifting methods was assessed horizontal wells. Their analysis also considered natural flow
through Net Present Value (NPV) calculations. In addition to conditions and the optimal interval for transitioning to
conventional selection criteria employed across the oil and alternative lifting methods. The results indicated that using
gas industry (OGI), the authors categorized selection factors ESP followed by SRP after a two-year period yielded greater
into three distinct groups: weighting factors, representing the profitability compared to employing a single or multiple
significance of each factor in the selection process and rated lifting methods. While using a single lifting method resulted
on a scale of 1 to 10; suitability factors, determined through in reduced efficiency, employing three methods significantly
mathematical equations; and economic factors, quantified increased the CAPEX of production.
Table 5: Summary of AL Method Feasibility for Hydrate Reservoir (Liu and Zerpa 2016)
AL ESP PCP SRP HP GL Plunger Compress Foam Vel
String
Shallow Well Well Well Well Well V. well Well Well V. well
depth suited suited suited suited suited suited suited suited suited
Offshore Maybe Maybe Poorly Poorly V. well Well Maybe Maybe V. well
suited suited suited suited suited
Permafrost Well Well Well Well Well V. well Well Well Well
suited suited suited suited suited suited suited suited suited
Low Poorly Maybe Well Maybe Maybe V. well V. well V. well Maybe
production suited suited suited suited suited
Low GLR Well Well Well Well Poorly Poorly V. poorly V. Poorly Poorly
suited suited suited suited suited suited suited suited suited
Low BHP Maybe Maybe V.well Maybe Poorly Well Well Well Maybe
suited suited suited suited suited
Viscous Poorly Maybe Poorly Poorly Poorly V.Poorly Poorly V.Poorly V.Poorly
production suited suited suited suited suited suited suited suited
Sandy Poorly Maybe Poorly Poorly Maybe V.poorly Poorly Well Well
production suited suited suited suited suited suited suited
Secondary Poorly Maybe Poorly Poorly Maybe V.poorly Poorly Well Well
hydrate suited suited suited suited suited suited suited
Ice Poorly Maybe Poorly Poorly Maybe V.poorly Poorly Well Well
suited suited suited suited suited suited suited

Slow Well Well Well Well Maybe Poorly Maybe Maybe Poorly
pressure suited suited suited suited suited suited
building up
Low Well V. well Well Well Well Well Well Well Well
reservoir suited suited suited suited suited suited suited suited suited
temperature
CAPEX 115,000 35,000 45,000 45,000 25,000 10,000 20,000 7,500 10,000
Pankaj et al. (2018) conducted an analysis of reservoir plots were employed to validate these steps, followed by
characteristics, including porosity, permeability, saturation, consideration of company specifications pertaining to
geological structures, and GORs, alongside varying temperature, pressure, and production rates. Furthermore,
production rates ranging from 500 to 2500 STB/D, employing additional analysis encompassed various aspects such as well
simulators to identify suitable AL for deep horizontal shale integrity and a comparison between qualitative and
wells. Their findings identified GL and Jets as the most quantitative well parameters that may limit the application of
suitable options, although Jets proved unsuitable for low-rate AL techniques, culminating in the ranking of each lifting
operations. Similarly, Escobar Patron et al. (2018) utilised method based on its applicability. Lane and Chokshi (2016)
simulators to ascertain the optimal AL methods for and Temizel et al. (2020) summarised the use of AL in
addressing production decline challenges in unconventionals unconventionals into four production stages (Fig. 1):
within the US. The software analysed input parameters such
as well depth, reservoir and fluid properties, and solids  High rates Jet is used for cleaning operations to remove
content to screen out AL methods based on known hydraulic fracturing fluids and this continues until
constraints. Subsequently, nodal analysis and NPV production declines.
calculations were employed for forecasting purposes,  GL, Plunger, and Foam Lift are to handle gas slug flow.
identifying AL systems capable of achieving the desired  GL, Jet, and ESP are used in early production, and an
production rate at minimal expense. Simulation scenarios amalgamation of GL and Jet/Foam could be applied
spanning 1, 3, and 6 years were considered, revealing that the depending on completion.
well could naturally flow for 3 months in all scenarios, after  SRP is used in the later production period after the decline
which ESPs and Jets were identified as suitable for higher occurs.
flow rates, followed by SRPs as production declined. Chow
et al. (2020) devised a selection tool for determining the most  Additional Selection Considerations Recommended by the
viable lifting method for offshore unconventionals. Analyses Authors are:
of well and fluid properties were conducted, and well  Integrated planning for well completion, considering
performance was validated independently for the selected several future AL.
method. Pump feasibility was evaluated in three stages:  AL life cycle estimation to reduce workover cost.
firstly, assessing the pump's capacity to handle Gas Volume  Continuous well parameters surveillance for production
Fraction (GVF); secondly, ensuring compatibility between optimisation.
pump and casing sizes; and finally, aligning reservoir
deliverability with pump size. Subsequent calculations and
Fig. 1: AL Life Stages in Unconventionals (Lane and Chokshi 2016)

IV. ML APPLICATION IN OIL AND GAS
Over the past decade, the utilisation of ML in the OGI has experienced a steady increase across various domains. Fig. 2
illustrates the recent applications of ML in the OGI, as evidenced by data obtained from Google Scholar (Pandey et al. 2020).
Fig. 2: Keywords Search on Google Scholar (Pandey et al. 2020)
In the OGI, ML has been applied to achieve significant accuracy and adaptability of ML algorithms in
advancements in big-data analysis, often compared with accommodating data anomalies within petroleum
traditional correlations and commercial software solutions. engineering. It is acknowledged that data collected in this
Theoretical and empirical correlations are at times deemed field is susceptible to errors, stemming from factors such as
impractical and limited to specific properties and datasets sensor inaccuracies or human oversight during data
(Khan et al. 2019). Elichev et al. (2019) contended that ML collection. The emphasis lies in training the ML models to
algorithms may lack accuracy when confronted with effectively handle such anomalies and derive approximate or
inadequate or uncertain data, highlighting the necessity to semi-corrected data through problem-solving approaches.
mitigate noise stemming from errors to ensure robust outputs. Over 500 papers addressing ML applications in the oil and
Conversely, several studies (Andrianova et al. 2018; Daigle gas industry have been published on OnePetro (Hajizadeh,
and Griffith 2018; Shoeibi Omrani et al. 2019) contradicted 2019). Table 7 provides examples of ML applications within
(Elichev et al. 2019) prospective by demonstrating the the oil and gas sector.
Table 7: ML Application Examples in Oil and Gas

Work Application
Andrianova et al. 2018 PVT analysis
Anifowose et al. 2017 Reservoir characterisation uncertainty
Ahmadi and Chen 2019; Elichev et al. 2019; Alakbari et al. 2017; Reservoir and fluid properties
Onwuchekwa 2018; Ramirez et al. 2017
Al-Alwani et al. 2019; Al Selaiti et al. 2020; Boguslawski et al. 2019; Bowie Well performance, production optimisation
2018; Cao et al. 2016; Han et al. 2019; Herve et al. 2020; Khan et al. 2019; and forecast
Luo et al. 2018; Pennel et al. 2018; Saghir et al. 2020
Noshi and Schubert 2018; Pollock et al. 2018 Drilling and directional drilling optimisation
Pankaj et al. 2018; Prosper and West 2018 Completion design in unconventionals
Chiroma et al. 2016 Oil prices estimation
Hajizadeh 2019 Strategic planning and development projects
V. ML APPLICATION IN AL SELECTION deliverability and forecast production, thereby identifying an

appropriate lifting method capable of achieving the desired
As previously discussed, the selection of AL methods flow rate. Traditional approaches have often involved the use
predominantly relies on the expertise of engineers and the of commercial simulators to design lifting methods, a process
historical performance of mature wells. Engineers typically that can be repetitive and tedious over time (Kefford and
conduct well performance and nodal analyses to assess well Gaurav 2016). In recent years, several researchers have

turned to ML techniques to analyse well performance and through data and compare various parameters. In contrast, the
reservoir/fluid properties. However, the application of ML as application of ML techniques offers a promising solution to
a selection technique for AL methods remains relatively the complexities of AL selection. Mahdi et al. (2023)
underexplored. Ounsakul et al. (2019) applied supervised ML demonstrates how ML models can significantly streamline
to determine the optimal lifting method from ESP, PCP, GL, the selection process by analysing large datasets and
and SRP. Their model utilises field data upon which identifying patterns that may not be immediately apparent to
algorithms are trained to analyse for AL selection. The human analysts. By leveraging ML algorithms, engineers can
objective was to enhance the selection criteria by minimising make more informed decisions about the most suitable AL
human errors. Three distinct algorithms, namely Naive method for a given well, potentially leading to improvements
Bayes, decision tree, and neural network, were employed to in production and revenue. Overall, while traditional
assess 30,000 samples encompassing reservoir, fluid, and selection techniques have their advantages, they are often
economic factors. Their findings demonstrated the capability limited by their rigidity and reliance on manual analysis. In
of ML to identify optimal pumps and reduce the lifecycle contrast, ML-based approaches offer the potential to
costs of producing wells compared to human decision- revolutionize AL selection by automating data analysis and
making processes. The author of this article recently providing more accurate predictions.
conducted AL selection research in a Sudanese oil field using
ML techniques (Mahdi et al. 2023). The analysis involved VII. CONCLUSION
several production parameters of four lifting methods along
with natural flow deployed across 24 wells over a period of The same AL have been used for decades with some
16 years, comprising a dataset of over 450,000 samples. Key modernisations. Most AL selection in the literature played
factors influencing AL selection were identified as gas around the prevalent selection criteria focusing on studying
content, wellhead pressure, and cumulative fluid production. and analysing reservoir parameters, fluid properties, well
Production performance and economic analyses were productivity, surface facilities, power requirements,
conducted to compare the actual AL performance in the field environmental aspects, corrosion, solids, paraffin handling,
with the predictions generated by ML. The results revealed gas handling, well completion and design, and economic
that the ML-predicted AL exhibited superior production factors, including workover and maintenance. The
performance compared to the actual implementations. Syed approaches applied in the literature might look the same,
et al. (2020) conducted AL system optimisation using ML however, every field has its peculiar circumstances and
techniques to facilitate the selection and monitoring of AL reservoir/fluid properties. Few applications of ML and AI in
systems within shale gas fields. Diverging from conventional AL selection are found in the literature, and the area is still
ML approaches, their study incorporated the consideration of fertile for more research. The current state of the art in AL is
the optimal timing for replacing the current AL systems, not fully optimised to meet the demands of the industry,
aiming to prevent pump failures and enhance profitability. especially in unconventional reservoirs. To address this gap,
This aspect of AL replacement during ongoing operations future research in AL should focus on developing innovative
may encounter resistance from oil companies, given the technologies, improving the understanding of the failure
potential reluctance to replace operative AL systems. mechanisms, reinforcing ML applications, and enhancing the
Additionally, the researchers investigated monitoring and design and operation practices of the existing systems.
maintenance practices, which are deemed essential within the Despite some innovations in gas handling and pump designs,
OGI to ensure operational efficiency and equipment integrity. problems still exist, and long-lasting AL seems to be the
In a related study, Ranjan et al., (2015) employed an Artificial inevitable challenge in the petroleum industry.
Neural Network (ANN) to optimise GL operations in an
offshore field situated in India. They developed a simplistic  Funding: This research received no external funding.
model consisting of 10 neurons representing reservoir and  Conflicts of Interest: The authors declare no conflict of
well parameters, with a single hidden layer employed as input interest.
to determine the optimal gas injection rate required to achieve
maximum oil production. The ML model served to validate
nodal analysis outcomes and provided engineers with a time-
saving alternative to laborious calculations.
VI. DISCUSSION
The traditional methods of AL selection, relying on

selection tables and flow rate and depth limitations, possess
several drawbacks. These methods often lack flexibility, as
they rely on predefined tables that may not accurately reflect
the unique characteristics of each well. Additionally, they
may have limitations in terms of the range of flow rates and
depths they can effectively handle. While these methods have
the advantage of simplicity and ease of use, they can also be
time-consuming, requiring engineers to manually screen

ABBREVIATION
AL Artificial lift MTMPCP Metal to metal PCP

SRP Sucker rod pump ML Machine learning
GL Gas lift CAPEX Capital expenditure
PCP Progressive cavity pump OPEX Operational expenditures
ESP Electrical submersible pump GOR Gas oil ratio
HJP Hydraulic jet pump IPR Inflow performance relationship
HPP Hydraulic piston pump WC% Water cut
Electrical submersible progressive
ESPCP PI Productivity index
cavity pump
Conventionals Conventional wells Unconventionals Unconventional wells
REFERENCES [8]. Alferov, A. V., Lutfurakhmanov, A. G., Litvinenko,

K. V., & Zdolnik, S. E. (2015, October). Artificial lift
[1]. Adam, A. M., Mohamed Ali, A. A., Elsadig, A. A., & strategy selection within field development planning.
Ahmed, A. A. (2022, March). An Intelligent Selection In SPE Russian Petroleum Technology Conference?
Model for Optimum Artificial Lift Method Using (pp. SPE-176670). SPE.
Multiple Criteria Decision-Making Approach. In http://dx.doi.org/10.2118/176670-MS
Offshore Technology Conference Asia (p. [9]. Alshmakhy, A., Al Daghar, K., Punnapala, S.,
D021S007R008). OTC. AlShehhi, S., Ben Amara, A., Makin, G., & Faux, S.
http://dx.doi.org/10.4043/31482-MS (2019, September). First Digital Intelligent Artificial
[2]. Ahmadi, M. A., & Chen, Z. (2019). Machine learning Lift Production Optimization Technology in UAE
models to predict bottom hole pressure in multi‐phase Dual-String Gas Lift Well-Business Case and
flow in vertical oil production wells. The Canadian Implementation Plan. In SPE Annual Technical
Journal of Chemical Engineering, 97(11), 2928-2940. Conference and Exhibition? (p. D021S030R005).
http://dx.doi.org/10.1002/cjce.23526 SPE. http://dx.doi.org/10.2118/196146-MS
[3]. Alakbari, F. S., Elkatatny, S., & Baarimah, S. O. [10]. Alshmakhy, A., Punnapala, S., AlShehhi, S., Ben
(2016, November). Prediction of bubble point Amara, A., Makin, G., & Faux, S. (2020, January).
pressure using artificial intelligence AI techniques. In First Digital Intelligent Artificial Lift Production
SPE middle east artificial lift conference and Optimization Technology in UAE Dual-String Gas
exhibition (p. D021S005R004). SPE. Lift Well–Completion and Installation
http://dx.doi.org/10.2118/184208-MS Considerations. In International Petroleum
[4]. Al-Alwani, M. A., Britt, L., Dunn-Norman, S., Technology Conference (p. D021S043R003). IPTC.
Alkinani, H. H., Al-Hameedi, A. T., & Al-Attar, A. http://dx.doi.org/10.2523/IPTC-19620-MS
(2019, June). Production performance estimation from [11]. Andrianova, A., Simonov, M., Perets, D., Margarit,
stimulation and completion parameters using machine A., Serebryakova, D., Bogdanov, Y., ... & Bukharev,
learning approach in the Marcellus Shale. In ARMA A. (2018, October). Application of machine learning
US Rock Mechanics/Geomechanics Symposium (pp. for oilfield data quality improvement. In SPE Russian
ARMA-2019). ARMA. Petroleum Technology Conference? (p.
[5]. Carlos, M., Luigi, S., Dorzhi, B., Yara, A., Erismar, D023S029R005). SPE.
R., Richard, M., & Quijada, D. (2020). Robust Data http://dx.doi.org/10.2118/191601-18RPTC-MS
Driven Well Performance Optimization Assisted by [12]. Anifowose, F. A., Labadin, J., & Abdulraheem, A.
Machine Learning Techniques for Natural Flowing (2017). Ensemble machine learning: An untapped
and Gas-Lift Wells in Abu Dhabi. modeling paradigm for petroleum reservoir
[6]. Alemi, M., Jalalifar, H., Kamali, G., & Kalbasi, M. characterization. Journal of Petroleum Science and
(2010). A prediction to the best artificial lift method Engineering, 151, 480-487.
selection on the basis of TOPSIS model. Journal of http://dx.doi.org/10.1016/j.petrol.2017.01.024
Petroleum and Gas Engineering, 1(1), 009-015. [13]. Blais, R. (1986). Artificial Lift Methods. poster,
[7]. Alemi, M., Jalalifar, H., Kamali, G. R., Kalbasi, M., & PennWell Publ. Co., Tulsa.
Research, P. E. D. E. C. (2011). A mathematical [14]. Boguslawski, B., Boujonnier, M., Bissuel-Beauvais,
estimation for artificial lift systems selection based on L., & Saghir, F. (2018, November). Edge Analytics at
ELECTRE model. Journal of Petroleum Science and the Wellhead: Designing Robust Machine Learning
Engineering, 78(1), 193-200. Models for Artificial Lift Failure Detection. In Abu
http://dx.doi.org/10.1016/j.petrol.2011.05.014. Dhabi International Petroleum Exhibition and
Conference (p. D031S078R004). SPE.
http://dx.doi.org/10.2118/192886-MS.

[15]. Bowie, B. (2018). Machine learning applied to [26]. Escobar Patron, K., Zhang, K., Xu, T., Lu, H., & Cui,
optimize Duvernay well performance, SPE 189823. In S. (2018, September). Case study of artificial lift
SPE Canada Unconventional Resources Conference, strategy selection and optimization for unconventional
Calgary, Alberta (pp. 1-24). oil wells in the Williston Basin. In SPE Liquids-Rich
http://dx.doi.org/10.2118/189823-MS Basins Conference-North America (p.
[16]. Brown, K. E. (1982). Overview of artificial lift D011S001R005). SPE.
systems. Journal of Petroleum Technology, 34(10), http://dx.doi.org/10.2118/191793-MS
2384-2396. http://dx.doi.org/10.2118/9979-PA [27]. Espin, D. A., Gasbarri, S., & Chacin, J. E. Expert
[17]. Bucaram, S. M. (1994). Managing artificial lift. system for selection of optimum Artificial Lift
Journal of Petroleum Technology, 46(04), 335-340. method. SPE Latin America/Caribbean Petroleum
http://dx.doi.org/10.2118/26212-PA Engineering Conference, Buenos Aires: OnePetro;
[18]. Caicedo, S., Montoya, C., Abboud, J., & Tiar, S. 1994. http://dx.doi.org/10.2118/26967-MS
(2015, November). A Systematic Integrated Approach [28]. Fatahi, E., Jalalifar, H., Pourafshari, P., & Moradi, B.
to Evaluate Artificial Lift Requirements While (2012). Selection of the best artificial lift method for
Dealing With High Uncertainty. In Abu Dhabi one of the Iranian oil field using multiple attribute
International Petroleum Exhibition and Conference (p. decision making methods. International Journal of
D031S043R005). SPE. Engineering and Technology, 2(2), 188-193.
http://dx.doi.org/10.2118/177436-MS [29]. Fatahi, E., Jalalifar, H., Pourafshari, P., & Rostami, A.
[19]. Cao, Q., Banerjee, R., Gupta, S., Li, J., Zhou, W., & J. (2011). Selection of the best artificial lift method in
Jeyachandra, B. (2016, June). Data driven production one of the iranian oil field by the employment of
forecasting using machine learning. In SPE Argentina ELECTRE model. British Journal of Applied Science
Exploration and Production of unconventional & Technology, 1(4), 172-180.
resources symposium (p. D021S006R001). SPE. http://dx.doi.org/10.9734/BJAST/2011/585
http://dx.doi.org/10.2118/180984-MS [30]. Fraga, R. S., Castellões, O. G., Assmann, B. W.,
[20]. Chiroma, H., Abdul-Kareem, S., Shukri Mohd Noor, Estevam, V., de Moura, G. T., Schröer, I. N., & do
A., Abubakar, A. I., Sohrabi Safa, N., Shuib, L., ... & Amaral, L. G. (2020). Progressive Vortex Pump: A
Herawan, T. (2016). A review on artificial intelligence New Artificial Lift Pumped Method. SPE Production
methodologies for the forecasting of crude oil price. & Operations, 35(02), 454-463.
Intelligent Automation & Soft Computing, 22(3), 449- http://dx.doi.org/10.2118/200497-PA
462. [31]. Hajizadeh, Y. (2019). Machine learning in oil and gas;
http://dx.doi.org/10.1080/10798587.2015.1092338 a SWOT analysis approach. Journal of Petroleum
[21]. Chow, J., Gamboa, J., Garcia, G. A., Price, T., & Hall, Science and Engineering, 176, 661-663.
C. (2020, November). Verifying feasibility of artificial http://dx.doi.org/10.1016/j.petrol.2019.01.113
lift methods in rapid selection tool. In SPE Artificial [32]. Han, D., Kwon, S., Son, H., & Lee, J. (2020,
Lift Conference and Exhibition-Americas. OnePetro. February). Production forecasting for shale gas well in
http://dx.doi.org/10.2118/201129-MS transient flow using machine learning and decline
[22]. Clegg, J. D., Bucaram, S. M., & Hein Jr, N. W. (1993). curve analysis. In Asia Pacific Unconventional
Recommendations and Comparisons for Selecting Resources Technology Conference, Brisbane,
Artificial-Lift Methods (includes associated papers Australia, 18-19 November 2019 (pp. 1510-1527).
28645 and 29092). Journal of Petroleum Technology, Unconventional Resources Technology Conference.
45(12), 1128-1167. http://dx.doi.org/10.2118/24834- [33]. Heinze, L. R., Thornsberry, K., & Witt, L. D. (1989,
PA March). AL: an expert system for selecting the optimal
[23]. Crnogorac, M., Tanasijević, M., Danilović, D., pumping method. In SPE Oklahoma City Oil and Gas
Karović Maričić, V., & Leković, B. (2020). Selection Symposium/Production and Operations Symposium
of Artificial Lift Methods: A Brief Review and New (pp. SPE-18872). SPE.
Model Based on Fuzzy Logic. Energies, 13(7), 1758. http://dx.doi.org/10.2118/18872-MS
http://dx.doi.org/10.3390/en13071758 [34]. Heinze, L. R., Winkler, H. W., & Lea, J. F. (1995,
[24]. Daigle, H., & Griffith, N. (2018, September). April). Decision Tree for selection of Artificial Lift
Optimizing nanoparticle-stabilized emulsion behavior method. In SPE Oklahoma City Oil and Gas
in porous media through electrostatic interactions. In Symposium/Production and Operations Symposium
SPE Annual Technical Conference and Exhibition? (p. (pp. SPE-29510). SPE.
D021S015R003). SPE. http://dx.doi.org/10.2118/29510-MS
http://dx.doi.org/10.2118/191596-MS [35]. Herve, P., Prado, G., & Rosner, M. (2020, May). How
[25]. Elichev, V., Bilogan, A., Litvinenko, K., Khabibullin, Machine Learning is Improving Production on
R., Alferov, A., & Vodopyan, A. (2019, October). Offshore Platforms. In Offshore Technology
Understanding well events with machine learning. In Conference (p. D022S002R008). OTC.
SPE Russian Petroleum Technology Conference? (p. http://dx.doi.org/10.4043/30782-MS
D013S005R003). SPE.
http://dx.doi.org/10.2118/196861-MS

[36]. Hoy, M., Knauhs, P., Langbauer, C., Pratscher, H. P., [46]. Lanier, G. H., & Mahoney, M. (2009). Pushing the
Cimitoglu, T., Marschall, C., ... & Hurch, S. (2020, Limit: High-Rate-Artificial-Lift Evaluation for a Sour,
November). Artificial Lift Selection and Testing for an Heavy-Oil, Thermal EOR Project in Oman. SPE
EOR Redevelopment Project–Lessons Learned from Production & Operations, 24(04), 579-589.
Field Pilots, Laboratory and Pump Test Facilities. In http://dx.doi.org/10.2118/115849-PA
SPE Artificial Lift Conference and Exhibition- [47]. Lea, J. F., & Nickens, H. V. (1999, March). Selection
Americas. OnePetro. of artificial lift. In SPE Oklahoma City Oil and Gas
http://dx.doi.org/10.2118/201128-MS Symposium/Production and Operations Symposium
[37]. Hoy, M., Kometer, B., Bürßner, P., Puscalau, G., & (pp. SPE-52157). SPE.
Eder, S. (2018, August). SRP equipment http://dx.doi.org/10.2118/52157-MS
customization creating value by increasing run life in [48]. Liu, Z., & Zerpa, L. E. (2016, May). Preliminary study
a low oil price environment. In SPE Artificial Lift of liquid loading problems for gas hydrate wells and
Conference and Exhibition-Americas? (p. selection of artificial lift methods. In SPE Western
D012S001R001). SPE. Regional Meeting (pp. SPE-180391). SPE.
http://dx.doi.org/10.2118/190958-MS http://dx.doi.org/10.2118/180391-MS
[38]. JPT staff, _. (2014). Techbits: Artificial Lift Selection [49]. Luo, G., Tian, Y., Bychina, M., & Ehlig-Economides,
Discussed at Workshop. Journal of Petroleum C. (2018, September). Production optimization using
Technology, 66(03), 38-40. machine learning in Bakken shale. In Unconventional
http://dx.doi.org/10.2118/0314-0038-JPT Resources Technology Conference, Houston, Texas,
[39]. Kaplan, V., & Duygu, E. (2014, May). Selection and 23-25 July 2018 (pp. 2174-2197). Society of
Optimization of Artificial Lift System in Heavy Oil Exploration Geophysicists, American Association of
Fields. In SPE Latin America and Caribbean Petroleum Geologists, Society of Petroleum
Petroleum Engineering Conference (p. Engineers. http://dx.doi.org/10.15530/urtec-2018-
D031S025R003). SPE. 2902505
http://dx.doi.org/10.2118/169288-MS [50]. Mahdi, M. A. A., Amish, M., & Oluyemi, G. (2023).
[40]. Kefford, P. A., & Gaurav, M. (2016, September). Well An Artificial Lift Selection Approach Using Machine
performance calculations for artificial lift screening. Learning: A Case Study in Sudan. Energies, 16(6),
In SPE Annual Technical Conference and Exhibition? 2853. http://dx.doi.org/10.3390/en16062853
(p. D031S048R001). SPE. [51]. Mali, P., & Al-Jasmi, A. (2014, June). Evaluation of
http://dx.doi.org/10.2118/181344-MS artificial lift modes for heavy oil reservoirs. In SPE
[41]. Khabibullin, R. A., & Krasnov, V. A. (2015, October). Canada Heavy Oil Conference (p. D021S011R006).
An approach for artificial lift applicability maps SPE. http://dx.doi.org/10.2118/170040-MS
construction. In SPE Russian Petroleum Technology [52]. Matondang, A. N., Ibnu, A., & Subiantoro, E. (2011,
Conference? (pp. SPE-176673). SPE. June). Application of hybrid artificial lift to produce
http://dx.doi.org/10.2118/176673-MS multizone with high GOR, contrast PI and contrast
[42]. Khan, M. R., Alnuaim, S., Tariq, Z., & Abdulraheem, water cut. In SPE Brasil Offshore (pp. SPE-143745).
A. (2019, March). Machine learning application for oil SPE. http://dx.doi.org/10.2118/143745-MS
rate prediction in artificial gas lift wells. In SPE [53]. Naguib, M. A., Bayoumi, A., El-Emam, N., & El
middle east oil and gas show and conference (p. Battrawy, A. (2000, October). Guideline of artificial
D032S085R002). SPE. lift selection for mature field. In SPE Asia Pacific Oil
http://dx.doi.org/10.2118/194713-MS and Gas Conference and Exhibition (pp. SPE-64428).
[43]. Khan, N., Ganzer, L., Elichev, V., & Ali, N. (2014, SPE. http://dx.doi.org/10.2118/64428-MS
May). An integrated life-time artificial lift selection [54]. Neely, B., Gipson, F., Clegg, J., Capps, B., & Wilson,
approach for tight/shale oil production. In SPE P. (1981, October). Selection of artificial lift method.
Hydrocarbon Economics and Evaluation Symposium In SPE Annual Technical Conference and Exhibition?
(p. D011S006R002). SPE. (pp. SPE-10337). SPE.
http://dx.doi.org/10.2118/169833-MS http://dx.doi.org/10.2118/10337-MS
[44]. Kolawole, O., Gamadi, T. D., & Bullard, D. (2020). [55]. Noshi, C. I., & Schubert, J. J. (2018, October). The
Artificial lift system applications in tight formations: role of machine learning in drilling operations; a
The state of knowledge. SPE Production & review. In SPE Eastern Regional Meeting (p.
Operations, 35(02), 422-434. D043S005R006). SPE.
http://dx.doi.org/10.2118/196592-PA http://dx.doi.org/10.2118/191823-18ERM-MS
[45]. Lane, W., & Chokshi, R. (2014, August). [56]. Onwuchekwa, C. (2018, August). Application of
Considerations for optimizing artificial lift in machine learning ideas to reservoir fluid properties
unconventionals. In SPE/AAPG/SEG Unconventional estimation. In SPE Nigeria Annual International
Resources Technology Conference (pp. URTEC- Conference and Exhibition (pp. SPE-193461). SPE.
1921823). URTeC. http://dx.doi.org/10.15530/urtec- http://dx.doi.org/10.2118/193461-MS
2014-1921823.

[57]. Ounsakul, T., Sirirattanachatchawan, T., [68]. Saghir, F., Gilabert, H., & Mancuso, B. M. (2020,
Pattarachupong, W., Yokrat, Y., & Ekkawong, P. October). Application of Augmented Intelligence and
(2019, March). Artificial lift selection using machine Edge Analytics In Upstream Production Operations:
learning. In International petroleum technology An Innovative Approach for Optimizing Artificial Lift
conference (p. D021S042R003). IPTC. Systems Performance. In SPE Annual Technical
http://dx.doi.org/10.2523/IPTC-19423-MS Conference and Exhibition? (p. D031S022R002).
[58]. Oyewole, P. (2016, October). Artificial lift selection SPE. http://dx.doi.org/10.2118/201516-MS
strategy to maximize unconventional oil and gas assets [69]. Shoeibi Omrani, P., Dobrovolschi, I., Belfroid, S.,
value. In SPE Artificial Lift Conference and Kronberger, P., & Munoz, E. (2018, November).
Exhibition-Americas? (p. D031S009R002). SPE. Improving the accuracy of virtual flow metering and
http://dx.doi.org/10.2118/181233-MS back-allocation through machine learning. In Abu
[59]. Pandey, Y. N., Rastogi, A., Kainkaryam, S., Dhabi International Petroleum Exhibition and
Bhattacharya, S., Saputelli, L., Pandey, Y. N., ... & Conference (p. D021S035R004). SPE.
Saputelli, L. (2020). Toward oil and gas 4.0. Machine http://dx.doi.org/10.2118/192819-MS
Learning in the Oil and Gas Industry: Including [70]. Stephenson, G. (2020). Technology Focus: Artificial
Geosciences, Reservoir Engineering, and Production Lift (March 2020). Journal of Petroleum Technology,
Engineering with Python, 1-40. 72(03), 47-47. http://dx.doi.org/10.2118/0320-0047-
http://dx.doi.org/10.1007/978-1-4842-6094-4_1 JPT
[60]. Pankaj, P., Geetan, S., MacDonald, R., Shukla, P., [71]. Syed, F. I., Alshamsi, M., Dahaghi, A. K., &
Sharma, A., Menasria, S., ... & Judd, T. (2018, April). Neghabhan, S. (2022). Artificial lift system
Application of data science and machine learning for optimization using machine learning applications.
well completion optimization. In Offshore Petroleum, 8(2), 219-226.
Technology Conference (p. D041S051R005). OTC. http://dx.doi.org/10.1016/j.petlm.2020.08.003
http://dx.doi.org/10.4043/28632-MS [72]. Temizel, C., Canbaz, C. H., Betancourt, D., Ozesen,
[61]. Pankaj, P., Patron, K. E., & Lu, H. (2018, August). A., Acar, C., Krishna, S., & Saputelli, L. (2020,
Wellbore Modeling and Reservoir Characterization October). A comprehensive review and optimization
for the Application of Artificial Lift in Deep of artificial lift methods in unconventionals. In SPE
Horizontal Wells in the Unconventional Reservoirs. In Annual Technical Conference and Exhibition? (p.
SPE Artificial Lift Conference and Exhibition- D041S053R008). SPE.
Americas? (p. D022S004R001). SPE. http://dx.doi.org/10.2118/201692-MS
http://dx.doi.org/10.2118/190922-MS [73]. Valbuena, J., Pereyra, E., & Sarica, C. (2016,
[62]. Parshall, J. (2013). Challenges, Opportunities Abound October). Defining the artificial lift system selection
for Artificial Lift. Journal of Petroleum Technology, guidelines for horizontal wells. In SPE Artificial Lift
65(03), 70-75. http://dx.doi.org/10.2118/0313-0070- Conference and Exhibition-Americas? (p.
JPT D031S009R003). SPE.
[63]. Pennel, M., Hsiung, J., & Putcha, V. B. (2018, April). http://dx.doi.org/10.2118/181229-MS
Detecting failures and optimizing performance in [74]. Valentin, E. P., & Hoffmann, F. C. (1988, October).
artificial lift using machine learning models. In SPE OPUS: An Expert Advisor for Artificial Lift. In SPE
Western Regional Meeting (p. D041S011R008). SPE. Annual Technical Conference and Exhibition? (pp.
http://dx.doi.org/10.2118/190090-MS SPE-18184). SPE. http://dx.doi.org/10.2118/18184-
[64]. Pollock, J., Stoecker-Sylvia, Z., Veedu, V., Panchal, MS
N., & Elshahawi, H. (2018, April). Machine learning [75]. Williams, S., Rozo, R., Aya, F. P., & Salazar
for improved directional drilling. In Offshore Hernández, J. I. (2008, September). Artificial Lift
Technology Conference (p. D031S031R001). OTC. Optimization in the Orito Field. In SPE Annual
http://dx.doi.org/10.4043/28633-MS Technical Conference and Exhibition? (pp. SPE-
[65]. Prosper, C., & West, D. (2018, October). Case study 116659). SPE. http://dx.doi.org/10.2118/116659-MS
applied machine learning to optimise PCP completion [76]. Zein El Din Shoukry, A., Soltys, T. W., Bettenson, J.,
design in a CBM field. In SPE Asia Pacific Oil and & Ariza, G. (2020, January). First Successful
Gas Conference and Exhibition (p. D021S016R002). Installation of Progressing Cavity Pump System in an
SPE. http://dx.doi.org/10.2118/192002-MS Oil Well at the Kingdom of Saudi Arabia. In
[66]. Ramirez, A. M., Valle, G. A., Romero, F., & Jaimes, International Petroleum Technology Conference (p.
M. (2017, May). Prediction of PVT properties in crude D023S173R001). IPTC.
oil using machine learning techniques MLT. In SPE http://dx.doi.org/10.2523/IPTC-19774-Abstract
Latin America and Caribbean Petroleum Engineering [77]. Zulkapli, M. H., Salim, M. M., Zaini, M. Z., Rivero
Conference (p. D021S009R002). SPE. Colmenares, M. E., Curteis, C., & Sepulveda, W.
http://dx.doi.org/10.2118/185536-MS (2014, December). The Evolution of Artificial Lift
[67]. Ranjan, A., Verma, S., & Singh, Y. (2015, March). Completions in an Offshore Brownfield in Malaysia.
Gas lift optimization using artificial neural network. In In International Petroleum Technology Conference
SPE Middle East Oil and Gas Show and Conference (pp. IPTC-18146). IPTC.
(pp. SPE-172610). SPE. http://dx.doi.org/10.2523/IPTC-18146-MS
http://dx.doi.org/10.2118/172610-MS
View publication stats

Artificial Lift Selection Methods in Conventional

Uploaded by

Copyright:

Available Formats

Artificial Lift Selection Methods in Conventional

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Artificial Lift Selection Methods in Conventional

Uploaded by

Copyright:

Available Formats

See discussions, stats, and author profiles for this publication at: https://www.researchgate.

Artiﬁcial Lift Selection Methods in Conventional and Unconventional Wells: A

Mohaned Mahdi Mohamed Amish

SEE PROFILE SEE PROFILE

The user has requested enhancement of the downloaded file.

Artificial Lift Selection Methods in Conventional and

Abstract:- Artificial lift (AL) selection is an important II. AL SELECTION IN CONVENTIONALS

IJISRT24MAR2108 www.ijisrt.com 2342

Table 1: AL Selection Techniques in the Literature

6 (Valentin) Valentin [well data (location, depth,

IJISRT24MAR2108 www.ijisrt.com 2343

IJISRT24MAR2108 www.ijisrt.com 2344

maturity, robustness (efficiency),

Chow et al. 2020

IJISRT24MAR2108 www.ijisrt.com 2345

IJISRT24MAR2108 www.ijisrt.com 2346

IJISRT24MAR2108 www.ijisrt.com 2347

Table 3: AL Selection Parameters (Zein El Din Shoukry et al. 2020)

Crnogorac et al. (2020) presented a study aimed at III. AL SELECTION IN UNCONVENTIONALS

IJISRT24MAR2108 www.ijisrt.com 2348

IJISRT24MAR2108 www.ijisrt.com 2349

Fig. 1: AL Life Stages in Unconventionals (Lane and Chokshi 2016)

IJISRT24MAR2108 www.ijisrt.com 2350

IV. ML APPLICATION IN OIL AND GAS

Fig. 2: Keywords Search on Google Scholar (Pandey et al. 2020)

Table 7: ML Application Examples in Oil and Gas

V. ML APPLICATION IN AL SELECTION deliverability and forecast production, thereby identifying an

IJISRT24MAR2108 www.ijisrt.com 2351

The traditional methods of AL selection, relying on

IJISRT24MAR2108 www.ijisrt.com 2352

AL Artificial lift MTMPCP Metal to metal PCP

REFERENCES [8]. Alferov, A. V., Lutfurakhmanov, A. G., Litvinenko,

IJISRT24MAR2108 www.ijisrt.com 2353

IJISRT24MAR2108 www.ijisrt.com 2354

IJISRT24MAR2108 www.ijisrt.com 2355

IJISRT24MAR2108 www.ijisrt.com 2356

View publication stats

You might also like