Oil Production Optimisation using Piecewise Linear Approximations (MILP): Computational Performance Comparison vs. MINLP Formulation

Emmanuel I. Epelle; Dimitrios I. Gerogiorgis

doi:10.1016/B978-0-12-823377-1.50209-3

Oil Production Optimisation using Piecewise Linear Approximations (MILP): Computational Performance Comparison vs. MINLP Formulation

Emmanuel I. Epelle, Dimitrios I. Gerogiorgis

School of Engineering

Research output: Chapter in Book/Report/Conference proceeding › Chapter (peer-reviewed) › peer-review

Abstract

Typical daily operations of oil and gas production systems are characterised by numerous decisions that must be carefully made if field profitability is to be sustained. These systems are usually nonlinear, nonconvex and involve binary decision variables; hence, the application of mathematical optimisation often results in an MINLP formulation. Piecewise linearisation techniques based on Special Ordered Sets of type 2 (SOS2) constraints have been used to approximate the nonlinear functions of the optimisation problem for complexity reduction. However, a computational analysis of these MILP-based formulations in comparison to their MINLP equivalents in oil production systems is scarce in literature. In this study, the benefits of MILP reformulation are applied to a synthetic but realistic case. In comparing both formulations, we evaluate solution sensitivity to the number of breakpoints, solution time, accuracy, and ease of automation.

Original language	English
Title of host publication	Computer Aided Chemical Engineering
Subtitle of host publication	30th European Symposium on Computer Aided Process Engineering
Editors	Sauro Pierucci, Flavio Manenti, Giulia Luisa Bozzano, Davide Manca
Publisher	Elsevier
Pages	1249-1254
Number of pages	6
Volume	48
ISBN (Print)	1570-7946
DOIs	https://doi.org/10.1016/B978-0-12-823377-1.50209-3
Publication status	Published - 19 Oct 2020
Event	30th European Symposium on Computer Aided Process Engineering - Online, Milan, Italy Duration: 31 Aug 2020 → 2 Sept 2020 https://www.aidic.it/escape30/enter.php

Publication series

Name	Computer Aided Chemical Engineering
Publisher	Elsevier
Volume	48
ISSN (Print)	1570-7946

Conference

Conference	30th European Symposium on Computer Aided Process Engineering
Abbreviated title	ESCAPE 30
Country/Territory	Italy
City	Milan
Period	31/08/20 → 2/09/20
Internet address	https://www.aidic.it/escape30/enter.php

Keywords / Materials (for Non-textual outputs)

Mixed-integer optimisation
piecewise linear approximation
oil production

Access to Document

10.1016/B978-0-12-823377-1.50209-3

358EpelleAccepted author manuscript, 1.13 MB

http://www.sciencedirect.com/science/article/pii/B9780128233771502093

Cite this

Epelle, E. I., & Gerogiorgis, D. I. (2020). Oil Production Optimisation using Piecewise Linear Approximations (MILP): Computational Performance Comparison vs. MINLP Formulation. In S. Pierucci, F. Manenti, G. L. Bozzano, & D. Manca (Eds.), Computer Aided Chemical Engineering: 30th European Symposium on Computer Aided Process Engineering (Vol. 48, pp. 1249-1254). (Computer Aided Chemical Engineering; Vol. 48). Elsevier. https://doi.org/10.1016/B978-0-12-823377-1.50209-3

Epelle, Emmanuel I. ; Gerogiorgis, Dimitrios I. / Oil Production Optimisation using Piecewise Linear Approximations (MILP): Computational Performance Comparison vs. MINLP Formulation. Computer Aided Chemical Engineering: 30th European Symposium on Computer Aided Process Engineering. editor / Sauro Pierucci ; Flavio Manenti ; Giulia Luisa Bozzano ; Davide Manca. Vol. 48 Elsevier, 2020. pp. 1249-1254 (Computer Aided Chemical Engineering).

@inbook{644749ea14ec4a4d92938e36620604c9,

title = "Oil Production Optimisation using Piecewise Linear Approximations (MILP): Computational Performance Comparison vs. MINLP Formulation",

abstract = "Typical daily operations of oil and gas production systems are characterised by numerous decisions that must be carefully made if field profitability is to be sustained. These systems are usually nonlinear, nonconvex and involve binary decision variables; hence, the application of mathematical optimisation often results in an MINLP formulation. Piecewise linearisation techniques based on Special Ordered Sets of type 2 (SOS2) constraints have been used to approximate the nonlinear functions of the optimisation problem for complexity reduction. However, a computational analysis of these MILP-based formulations in comparison to their MINLP equivalents in oil production systems is scarce in literature. In this study, the benefits of MILP reformulation are applied to a synthetic but realistic case. In comparing both formulations, we evaluate solution sensitivity to the number of breakpoints, solution time, accuracy, and ease of automation.",

keywords = "Mixed-integer optimisation, piecewise linear approximation, oil production",

author = "Epelle, {Emmanuel I.} and Gerogiorgis, {Dimitrios I.}",

year = "2020",

month = oct,

day = "19",

doi = "10.1016/B978-0-12-823377-1.50209-3",

language = "English",

isbn = "1570-7946",

volume = "48",

series = "Computer Aided Chemical Engineering",

publisher = "Elsevier",

pages = "1249--1254",

editor = "Sauro Pierucci and Flavio Manenti and Bozzano, {Giulia Luisa} and Davide Manca",

booktitle = "Computer Aided Chemical Engineering",

address = "Netherlands",

note = "30th European Symposium on Computer Aided Process Engineering, ESCAPE 30 ; Conference date: 31-08-2020 Through 02-09-2020",

url = "https://www.aidic.it/escape30/enter.php",

}

Epelle, EI & Gerogiorgis, DI 2020, Oil Production Optimisation using Piecewise Linear Approximations (MILP): Computational Performance Comparison vs. MINLP Formulation. in S Pierucci, F Manenti, GL Bozzano & D Manca (eds), Computer Aided Chemical Engineering: 30th European Symposium on Computer Aided Process Engineering. vol. 48, Computer Aided Chemical Engineering, vol. 48, Elsevier, pp. 1249-1254, 30th European Symposium on Computer Aided Process Engineering, Milan, Italy, 31/08/20. https://doi.org/10.1016/B978-0-12-823377-1.50209-3

Oil Production Optimisation using Piecewise Linear Approximations (MILP): Computational Performance Comparison vs. MINLP Formulation. / Epelle, Emmanuel I.; Gerogiorgis, Dimitrios I.
Computer Aided Chemical Engineering: 30th European Symposium on Computer Aided Process Engineering. ed. / Sauro Pierucci; Flavio Manenti; Giulia Luisa Bozzano; Davide Manca. Vol. 48 Elsevier, 2020. p. 1249-1254 (Computer Aided Chemical Engineering; Vol. 48).

Research output: Chapter in Book/Report/Conference proceeding › Chapter (peer-reviewed) › peer-review

TY - CHAP

T1 - Oil Production Optimisation using Piecewise Linear Approximations (MILP): Computational Performance Comparison vs. MINLP Formulation

AU - Epelle, Emmanuel I.

AU - Gerogiorgis, Dimitrios I.

PY - 2020/10/19

Y1 - 2020/10/19

N2 - Typical daily operations of oil and gas production systems are characterised by numerous decisions that must be carefully made if field profitability is to be sustained. These systems are usually nonlinear, nonconvex and involve binary decision variables; hence, the application of mathematical optimisation often results in an MINLP formulation. Piecewise linearisation techniques based on Special Ordered Sets of type 2 (SOS2) constraints have been used to approximate the nonlinear functions of the optimisation problem for complexity reduction. However, a computational analysis of these MILP-based formulations in comparison to their MINLP equivalents in oil production systems is scarce in literature. In this study, the benefits of MILP reformulation are applied to a synthetic but realistic case. In comparing both formulations, we evaluate solution sensitivity to the number of breakpoints, solution time, accuracy, and ease of automation.

AB - Typical daily operations of oil and gas production systems are characterised by numerous decisions that must be carefully made if field profitability is to be sustained. These systems are usually nonlinear, nonconvex and involve binary decision variables; hence, the application of mathematical optimisation often results in an MINLP formulation. Piecewise linearisation techniques based on Special Ordered Sets of type 2 (SOS2) constraints have been used to approximate the nonlinear functions of the optimisation problem for complexity reduction. However, a computational analysis of these MILP-based formulations in comparison to their MINLP equivalents in oil production systems is scarce in literature. In this study, the benefits of MILP reformulation are applied to a synthetic but realistic case. In comparing both formulations, we evaluate solution sensitivity to the number of breakpoints, solution time, accuracy, and ease of automation.

KW - Mixed-integer optimisation

KW - piecewise linear approximation

KW - oil production

U2 - 10.1016/B978-0-12-823377-1.50209-3

DO - 10.1016/B978-0-12-823377-1.50209-3

M3 - Chapter (peer-reviewed)

SN - 1570-7946

VL - 48

T3 - Computer Aided Chemical Engineering

SP - 1249

EP - 1254

BT - Computer Aided Chemical Engineering

A2 - Pierucci, Sauro

A2 - Manenti, Flavio

A2 - Bozzano, Giulia Luisa

A2 - Manca, Davide

PB - Elsevier

T2 - 30th European Symposium on Computer Aided Process Engineering

Y2 - 31 August 2020 through 2 September 2020

ER -

Epelle EI , Gerogiorgis DI. Oil Production Optimisation using Piecewise Linear Approximations (MILP): Computational Performance Comparison vs. MINLP Formulation. In Pierucci S, Manenti F, Bozzano GL, Manca D, editors, Computer Aided Chemical Engineering: 30th European Symposium on Computer Aided Process Engineering. Vol. 48. Elsevier. 2020. p. 1249-1254. (Computer Aided Chemical Engineering). doi: 10.1016/B978-0-12-823377-1.50209-3

Oil Production Optimisation using Piecewise Linear Approximations (MILP): Computational Performance Comparison vs. MINLP Formulation

Abstract

Publication series

Conference

Keywords / Materials (for Non-textual outputs)

Access to Document

Fingerprint

Cite this