Measuring accuracy and computational capacity trade-offs in an hourly integrated energy system model

Amirhossein Fattahi; Manuel Sánchez Diéguez; Jos Sijm; Germán Morales España; André Faaij

doi:10.1016/j.adapen.2021.100009

Measuring accuracy and computational capacity trade-offs in an hourly integrated energy system model

Amirhossein Fattahi^*, Manuel Sánchez Diéguez, Jos Sijm, Germán Morales España, André Faaij

^*Corresponding author for this work

Research output: Contribution to journal › Article › Academic › peer-review

Abstract

Improving energy system modeling capabilities can directly affect the quality of applied studies. However, some modeling trade-offs are necessary as the computational capacity and data availability are constrained. In this paper, we demonstrate modeling trade-offs resulting from the modification in the resolution of four modeling capabilities, namely, transitional scope, European electricity interconnection, hourly demand-side flexibility description, and infrastructure representation. We measure the cost of increasing resolution in each capability in terms of computational time and several energy system modeling indicators, notably, system costs, emission prices, and electricity import and export levels. The analyses are performed in a national-level integrated energy system model with a linear programming approach that includes the hourly electricity dispatch with European nodes. We determined that reducing the transitional scope from seven to two periods can reduce the computational time by 75% while underestimating the objective function by only 4.6%. Modelers can assume a single European Union node that dispatches electricity at an aggregated level, which underestimates the objective function by 1% while halving the computational time. Furthermore, the absence of shedding and storage flexibility options can increase the curtailed electricity by 25% and 8%, respectively. Although neglecting flexibility options can drastically decrease the computational time, it can increase the sub-optimality by 31%. We conclude that an increased resolution in modeling flexibility options can significantly improve the results. While reducing the computational time by half, the lack of electricity and gas infrastructure representation can underestimate the objective function by 4% and 6%, respectively.

Original language	English
Article number	100009
Pages (from-to)	1-22
Journal	Advances in Applied Energy
Volume	1
DOIs	https://doi.org/10.1016/j.adapen.2021.100009
Publication status	Published - 23 Feb 2021

Bibliographical note

Funding Information:
The authors wish to acknowledge the support provided by the ESTRAC Integrated Energy System Analysis project financed by the New Energy Coalition (finance code: 656039). The views expressed here are those of the authors alone and do not necessarily reflect the views of the project partners or the policies of the funding partners.

Funding Information:
The authors would like to thank Klara Schure and Robert Koelemeijer from PBL (the Netherlands Environmental Agency) for their efforts in developing the ENSYSI model, which played an important role in the creation of the IESA-Opt model. Furthermore, we want to thank other members of the Energy Transition team at TNO for their help and guidance. The authors wish to acknowledge the support provided by the ESTRAC Integrated Energy System Analysis project financed by the New Energy Coalition (finance code: 656039). The views expressed here are those of the authors alone and do not necessarily reflect the views of the project partners or the policies of the funding partners.

Publisher Copyright:
© 2021 The Author(s)

Funding

The authors wish to acknowledge the support provided by the ESTRAC Integrated Energy System Analysis project financed by the New Energy Coalition (finance code: 656039). The views expressed here are those of the authors alone and do not necessarily reflect the views of the project partners or the policies of the funding partners. The authors would like to thank Klara Schure and Robert Koelemeijer from PBL (the Netherlands Environmental Agency) for their efforts in developing the ENSYSI model, which played an important role in the creation of the IESA-Opt model. Furthermore, we want to thank other members of the Energy Transition team at TNO for their help and guidance. The authors wish to acknowledge the support provided by the ESTRAC Integrated Energy System Analysis project financed by the New Energy Coalition (finance code: 656039). The views expressed here are those of the authors alone and do not necessarily reflect the views of the project partners or the policies of the funding partners.

Keywords

Energy system modeling
Flexibility
Hourly temporal resolution
Modeling capabilities
Modeling trade-offs

Access to Document

10.1016/j.adapen.2021.100009Licence: CC BY

1-s2.0-S2666792421000020-mainFinal published version, 3.96 MBLicence: CC BY

Cite this

@article{9a1217f79c1646168eb2910f9daf92b5,

title = "Measuring accuracy and computational capacity trade-offs in an hourly integrated energy system model",

abstract = "Improving energy system modeling capabilities can directly affect the quality of applied studies. However, some modeling trade-offs are necessary as the computational capacity and data availability are constrained. In this paper, we demonstrate modeling trade-offs resulting from the modification in the resolution of four modeling capabilities, namely, transitional scope, European electricity interconnection, hourly demand-side flexibility description, and infrastructure representation. We measure the cost of increasing resolution in each capability in terms of computational time and several energy system modeling indicators, notably, system costs, emission prices, and electricity import and export levels. The analyses are performed in a national-level integrated energy system model with a linear programming approach that includes the hourly electricity dispatch with European nodes. We determined that reducing the transitional scope from seven to two periods can reduce the computational time by 75\% while underestimating the objective function by only 4.6\%. Modelers can assume a single European Union node that dispatches electricity at an aggregated level, which underestimates the objective function by 1\% while halving the computational time. Furthermore, the absence of shedding and storage flexibility options can increase the curtailed electricity by 25\% and 8\%, respectively. Although neglecting flexibility options can drastically decrease the computational time, it can increase the sub-optimality by 31\%. We conclude that an increased resolution in modeling flexibility options can significantly improve the results. While reducing the computational time by half, the lack of electricity and gas infrastructure representation can underestimate the objective function by 4\% and 6\%, respectively.",

keywords = "Energy system modeling, Flexibility, Hourly temporal resolution, Modeling capabilities, Modeling trade-offs",

author = "Amirhossein Fattahi and \{S{\'a}nchez Di{\'e}guez\}, Manuel and Jos Sijm and \{Morales Espa{\~n}a\}, Germ{\'a}n and Andr{\'e} Faaij",

note = "Funding Information: The authors wish to acknowledge the support provided by the ESTRAC Integrated Energy System Analysis project financed by the New Energy Coalition (finance code: 656039). The views expressed here are those of the authors alone and do not necessarily reflect the views of the project partners or the policies of the funding partners. Funding Information: The authors would like to thank Klara Schure and Robert Koelemeijer from PBL (the Netherlands Environmental Agency) for their efforts in developing the ENSYSI model, which played an important role in the creation of the IESA-Opt model. Furthermore, we want to thank other members of the Energy Transition team at TNO for their help and guidance. The authors wish to acknowledge the support provided by the ESTRAC Integrated Energy System Analysis project financed by the New Energy Coalition (finance code: 656039). The views expressed here are those of the authors alone and do not necessarily reflect the views of the project partners or the policies of the funding partners. Publisher Copyright: {\textcopyright} 2021 The Author(s)",

year = "2021",

month = feb,

day = "23",

doi = "10.1016/j.adapen.2021.100009",

language = "English",

volume = "1",

pages = "1--22",

journal = "Advances in Applied Energy",

issn = "2666-7924",

publisher = "Elsevier Ltd",

}

TY - JOUR

T1 - Measuring accuracy and computational capacity trade-offs in an hourly integrated energy system model

AU - Fattahi, Amirhossein

AU - Sánchez Diéguez, Manuel

AU - Sijm, Jos

AU - Morales España, Germán

AU - Faaij, André

N1 - Funding Information: The authors wish to acknowledge the support provided by the ESTRAC Integrated Energy System Analysis project financed by the New Energy Coalition (finance code: 656039). The views expressed here are those of the authors alone and do not necessarily reflect the views of the project partners or the policies of the funding partners. Funding Information: The authors would like to thank Klara Schure and Robert Koelemeijer from PBL (the Netherlands Environmental Agency) for their efforts in developing the ENSYSI model, which played an important role in the creation of the IESA-Opt model. Furthermore, we want to thank other members of the Energy Transition team at TNO for their help and guidance. The authors wish to acknowledge the support provided by the ESTRAC Integrated Energy System Analysis project financed by the New Energy Coalition (finance code: 656039). The views expressed here are those of the authors alone and do not necessarily reflect the views of the project partners or the policies of the funding partners. Publisher Copyright: © 2021 The Author(s)

PY - 2021/2/23

Y1 - 2021/2/23

N2 - Improving energy system modeling capabilities can directly affect the quality of applied studies. However, some modeling trade-offs are necessary as the computational capacity and data availability are constrained. In this paper, we demonstrate modeling trade-offs resulting from the modification in the resolution of four modeling capabilities, namely, transitional scope, European electricity interconnection, hourly demand-side flexibility description, and infrastructure representation. We measure the cost of increasing resolution in each capability in terms of computational time and several energy system modeling indicators, notably, system costs, emission prices, and electricity import and export levels. The analyses are performed in a national-level integrated energy system model with a linear programming approach that includes the hourly electricity dispatch with European nodes. We determined that reducing the transitional scope from seven to two periods can reduce the computational time by 75% while underestimating the objective function by only 4.6%. Modelers can assume a single European Union node that dispatches electricity at an aggregated level, which underestimates the objective function by 1% while halving the computational time. Furthermore, the absence of shedding and storage flexibility options can increase the curtailed electricity by 25% and 8%, respectively. Although neglecting flexibility options can drastically decrease the computational time, it can increase the sub-optimality by 31%. We conclude that an increased resolution in modeling flexibility options can significantly improve the results. While reducing the computational time by half, the lack of electricity and gas infrastructure representation can underestimate the objective function by 4% and 6%, respectively.

AB - Improving energy system modeling capabilities can directly affect the quality of applied studies. However, some modeling trade-offs are necessary as the computational capacity and data availability are constrained. In this paper, we demonstrate modeling trade-offs resulting from the modification in the resolution of four modeling capabilities, namely, transitional scope, European electricity interconnection, hourly demand-side flexibility description, and infrastructure representation. We measure the cost of increasing resolution in each capability in terms of computational time and several energy system modeling indicators, notably, system costs, emission prices, and electricity import and export levels. The analyses are performed in a national-level integrated energy system model with a linear programming approach that includes the hourly electricity dispatch with European nodes. We determined that reducing the transitional scope from seven to two periods can reduce the computational time by 75% while underestimating the objective function by only 4.6%. Modelers can assume a single European Union node that dispatches electricity at an aggregated level, which underestimates the objective function by 1% while halving the computational time. Furthermore, the absence of shedding and storage flexibility options can increase the curtailed electricity by 25% and 8%, respectively. Although neglecting flexibility options can drastically decrease the computational time, it can increase the sub-optimality by 31%. We conclude that an increased resolution in modeling flexibility options can significantly improve the results. While reducing the computational time by half, the lack of electricity and gas infrastructure representation can underestimate the objective function by 4% and 6%, respectively.

KW - Energy system modeling

KW - Flexibility

KW - Hourly temporal resolution

KW - Modeling capabilities

KW - Modeling trade-offs

UR - https://www.scopus.com/pages/publications/85118556267

U2 - 10.1016/j.adapen.2021.100009

DO - 10.1016/j.adapen.2021.100009

M3 - Article

AN - SCOPUS:85118556267

SN - 2666-7924

VL - 1

SP - 1

EP - 22

JO - Advances in Applied Energy

JF - Advances in Applied Energy

M1 - 100009

ER -

Measuring accuracy and computational capacity trade-offs in an hourly integrated energy system model

Abstract

Bibliographical note

Funding

Keywords

Access to Document

Other files and links

Fingerprint

Cite this