Applying Neural Networks to Large-Scale Distribution System Analysis: an Empirical Computational Perspective

Sjoerd Doumen; R. Bernards; Johan Morren; N. G. Paterakis

doi:10.1109/SEST48500.2020.9203347

Applying Neural Networks to Large-Scale Distribution System Analysis: an Empirical Computational Perspective

Sjoerd Doumen, R. Bernards, Johan Morren, N. G. Paterakis

Onderzoeksoutput: Hoofdstuk in Boek/Rapport/Congresprocedure › Conferentiebijdrage › Academic › peer review

Samenvatting

Research has shown that Neural Networks (NNs)
are capable of accurate and quick low voltage (LV) grid analysis.
Therefore, NNs could be a viable method for the middle longterm
scenario tool (MLT), a tool created and used by Enexis,
one of the major distribution system operator (DSO) in the
Netherlands, to analyze future scenarios of LV grids. However,
the tool analyzes a substantial amount of LV grids, and each
would require a NN. This amount of NNs necessitates a single
network architecture and training method for all NNs, which can be achieved by knowing hyperparameters beforehand, since determining hyperparameters is computationally costly. This paper estimates how long it would take to train a substantial amount of NNs, determines if hyperparameters are shareable between NNs of similar-sized LV grids and if hyperparameters are predictable based on LV grid sizes. The results of hyperparameter sharing show comparable performance between NNs, however, differences start to occur for larger LV grids. Predicting hyperparameters based on LV grid size gives an unsatisfactory
performance.

Originele taal-2	Engels
Titel	2020 International Conference on Smart Energy Systems and Technologies (SEST)
Plaats van productie	Istanbul
Aantal pagina's	6
ISBN van elektronische versie	9781728147017
DOI's	https://doi.org/10.1109/SEST48500.2020.9203347
Status	Gepubliceerd - 7 sep 2020

Toegang tot document

10.1109/SEST48500.2020.9203347

Link to publication in Scopus

Vingerafdruk Duik in de onderzoeksthema's van 'Applying Neural Networks to Large-Scale Distribution System Analysis: an Empirical Computational Perspective'. Samen vormen ze een unieke vingerafdruk.

Citeer dit

@inproceedings{14da21bd2120482ea4e4ba534b245137,

title = "Applying Neural Networks to Large-Scale Distribution System Analysis: an Empirical Computational Perspective",

abstract = "Research has shown that Neural Networks (NNs)are capable of accurate and quick low voltage (LV) grid analysis.Therefore, NNs could be a viable method for the middle longtermscenario tool (MLT), a tool created and used by Enexis,one of the major distribution system operator (DSO) in theNetherlands, to analyze future scenarios of LV grids. However,the tool analyzes a substantial amount of LV grids, and eachwould require a NN. This amount of NNs necessitates a singlenetwork architecture and training method for all NNs, which can be achieved by knowing hyperparameters beforehand, since determining hyperparameters is computationally costly. This paper estimates how long it would take to train a substantial amount of NNs, determines if hyperparameters are shareable between NNs of similar-sized LV grids and if hyperparameters are predictable based on LV grid sizes. The results of hyperparameter sharing show comparable performance between NNs, however, differences start to occur for larger LV grids. Predicting hyperparameters based on LV grid size gives an unsatisfactoryperformance.",

keywords = "Grid Analysis, Load Flow, Low Voltage Grids, Neural Networks",

author = "Sjoerd Doumen and R. Bernards and Johan Morren and Paterakis, {N. G.}",

year = "2020",

month = sep,

day = "7",

doi = "10.1109/SEST48500.2020.9203347",

language = "English",

booktitle = "2020 International Conference on Smart Energy Systems and Technologies (SEST)",

}

Applying Neural Networks to Large-Scale Distribution System Analysis: an Empirical Computational Perspective. / Doumen, Sjoerd; Bernards, R.; Morren, Johan ; Paterakis, N. G.

2020 International Conference on Smart Energy Systems and Technologies (SEST). Istanbul, 2020. 9203347.

Onderzoeksoutput: Hoofdstuk in Boek/Rapport/Congresprocedure › Conferentiebijdrage › Academic › peer review

TY - GEN

T1 - Applying Neural Networks to Large-Scale Distribution System Analysis: an Empirical Computational Perspective

AU - Doumen, Sjoerd

AU - Bernards, R.

AU - Morren, Johan

AU - Paterakis, N. G.

PY - 2020/9/7

Y1 - 2020/9/7

N2 - Research has shown that Neural Networks (NNs)are capable of accurate and quick low voltage (LV) grid analysis.Therefore, NNs could be a viable method for the middle longtermscenario tool (MLT), a tool created and used by Enexis,one of the major distribution system operator (DSO) in theNetherlands, to analyze future scenarios of LV grids. However,the tool analyzes a substantial amount of LV grids, and eachwould require a NN. This amount of NNs necessitates a singlenetwork architecture and training method for all NNs, which can be achieved by knowing hyperparameters beforehand, since determining hyperparameters is computationally costly. This paper estimates how long it would take to train a substantial amount of NNs, determines if hyperparameters are shareable between NNs of similar-sized LV grids and if hyperparameters are predictable based on LV grid sizes. The results of hyperparameter sharing show comparable performance between NNs, however, differences start to occur for larger LV grids. Predicting hyperparameters based on LV grid size gives an unsatisfactoryperformance.

AB - Research has shown that Neural Networks (NNs)are capable of accurate and quick low voltage (LV) grid analysis.Therefore, NNs could be a viable method for the middle longtermscenario tool (MLT), a tool created and used by Enexis,one of the major distribution system operator (DSO) in theNetherlands, to analyze future scenarios of LV grids. However,the tool analyzes a substantial amount of LV grids, and eachwould require a NN. This amount of NNs necessitates a singlenetwork architecture and training method for all NNs, which can be achieved by knowing hyperparameters beforehand, since determining hyperparameters is computationally costly. This paper estimates how long it would take to train a substantial amount of NNs, determines if hyperparameters are shareable between NNs of similar-sized LV grids and if hyperparameters are predictable based on LV grid sizes. The results of hyperparameter sharing show comparable performance between NNs, however, differences start to occur for larger LV grids. Predicting hyperparameters based on LV grid size gives an unsatisfactoryperformance.

KW - Grid Analysis

KW - Load Flow

KW - Low Voltage Grids

KW - Neural Networks

UR - http://www.scopus.com/inward/record.url?scp=85093649635&partnerID=8YFLogxK

U2 - 10.1109/SEST48500.2020.9203347

DO - 10.1109/SEST48500.2020.9203347

M3 - Conference contribution

BT - 2020 International Conference on Smart Energy Systems and Technologies (SEST)

CY - Istanbul

ER -