Probabilistic Forecasting of Day-Ahead Electricity Prices and their Volatility with LSTMs

Julius Trebbien; Sebastian Pütz; Benjamin Schäfer; Heidi S. Nygård; Leonardo Rydin Gorjão; Dirk Witthaut

doi:10.1109/ISGTEUROPE56780.2023.10407112

Probabilistic Forecasting of Day-Ahead Electricity Prices and their Volatility with LSTMs

Julius Trebbien
, Sebastian Pütz
, Benjamin Schäfer
, Heidi S. Nygård
, Leonardo Rydin Gorjão
, Dirk Witthaut

Research output: Chapter in Book/Report/Conference proceeding › Conference Article in proceeding › Academic › peer-review

Abstract

Accurate forecasts of electricity prices are crucial for the management of electric power systems and the development of smart applications. European electricity prices have risen substantially and became highly volatile after the Russian invasion of Ukraine, challenging established forecasting methods. Here, we present a Long Short-Term Memory (LSTM) model for the German-Luxembourg day-ahead electricity prices addressing these challenges. The recurrent structure of the LSTM allows the model to adapt to trends, while the joint prediction of both mean and standard deviation enables a probabilistic prediction. Using a physics-inspired approach–superstatistics–to derive an explanation for the statistics of prices, we show that the LSTM model faithfully reproduces both prices and their volatility.

Original language	English
Title of host publication	2023 IEEE PES Innovative Smart Grid Technologies Europe (ISGT EUROPE)
Place of Publication	Grenoble, France, 2023
Publisher	IEEE Canada
Number of pages	5
ISBN (Electronic)	979-8-3503-9678-2
ISBN (Print)	979-8-3503-9679-9
DOIs	https://doi.org/10.1109/ISGTEUROPE56780.2023.10407112
Publication status	Published - 30 Jan 2024
Externally published	Yes
Event	IEEE PES ISGT EUROPE 2023 - Université Grenoble Alpes, Grenoble, France Duration: 23 Oct 2023 → 26 Oct 2023 https://attend.ieee.org/isgt-europe-2023/

Conference

Conference	IEEE PES ISGT EUROPE 2023
Abbreviated title	ISGT Europe 2023
Country/Territory	France
City	Grenoble
Period	23/10/23 → 26/10/23
Internet address	https://attend.ieee.org/isgt-europe-2023/

Keywords

Adaptation models
Superstatistics
Predictive models
Probabilistic logic
Smart grids
probabilistic forecasting
Electricity prices
Day-ahead electricity prices
German-Luxembourg electricity prices
LSTM
Volatility
Heavy tailed distributions

Access to Document

10.1109/ISGTEUROPE56780.2023.10407112

Cite this

@inproceedings{14bca41f36524d558b67ca6b4c46052b,

title = "Probabilistic Forecasting of Day-Ahead Electricity Prices and their Volatility with LSTMs",

abstract = "Accurate forecasts of electricity prices are crucial for the management of electric power systems and the development of smart applications. European electricity prices have risen substantially and became highly volatile after the Russian invasion of Ukraine, challenging established forecasting methods. Here, we present a Long Short-Term Memory (LSTM) model for the German-Luxembourg day-ahead electricity prices addressing these challenges. The recurrent structure of the LSTM allows the model to adapt to trends, while the joint prediction of both mean and standard deviation enables a probabilistic prediction. Using a physics-inspired approach–superstatistics–to derive an explanation for the statistics of prices, we show that the LSTM model faithfully reproduces both prices and their volatility.",

keywords = "Adaptation models, Superstatistics, Predictive models, Probabilistic logic, Smart grids, probabilistic forecasting, Electricity prices, Day-ahead electricity prices, German-Luxembourg electricity prices, LSTM, Volatility, Heavy tailed distributions",

author = "Julius Trebbien and Sebastian P{\"u}tz and Benjamin Sch{\"a}fer and Nyg{\aa}rd, \{Heidi S.\} and \{Rydin Gorj{\~a}o\}, Leonardo and Dirk Witthaut",

year = "2024",

month = jan,

day = "30",

doi = "10.1109/ISGTEUROPE56780.2023.10407112",

language = "English",

isbn = "979-8-3503-9679-9",

booktitle = "2023 IEEE PES Innovative Smart Grid Technologies Europe (ISGT EUROPE)",

publisher = "IEEE Canada",

address = "Canada",

note = "IEEE PES ISGT EUROPE 2023, ISGT Europe 2023 ; Conference date: 23-10-2023 Through 26-10-2023",

url = "https://attend.ieee.org/isgt-europe-2023/",

}

Trebbien, J, Pütz, S, Schäfer, B, Nygård, HS, Rydin Gorjão, L & Witthaut, D 2024, Probabilistic Forecasting of Day-Ahead Electricity Prices and their Volatility with LSTMs. in 2023 IEEE PES Innovative Smart Grid Technologies Europe (ISGT EUROPE). IEEE Canada, Grenoble, France, 2023, IEEE PES ISGT EUROPE 2023, Grenoble, France, 23/10/23. https://doi.org/10.1109/ISGTEUROPE56780.2023.10407112

Probabilistic Forecasting of Day-Ahead Electricity Prices and their Volatility with LSTMs. / Trebbien, Julius; Pütz, Sebastian; Schäfer, Benjamin et al.
2023 IEEE PES Innovative Smart Grid Technologies Europe (ISGT EUROPE). Grenoble, France, 2023: IEEE Canada, 2024.

Research output: Chapter in Book/Report/Conference proceeding › Conference Article in proceeding › Academic › peer-review

TY - GEN

T1 - Probabilistic Forecasting of Day-Ahead Electricity Prices and their Volatility with LSTMs

AU - Trebbien, Julius

AU - Pütz, Sebastian

AU - Schäfer, Benjamin

AU - Nygård, Heidi S.

AU - Rydin Gorjão, Leonardo

AU - Witthaut, Dirk

PY - 2024/1/30

Y1 - 2024/1/30

N2 - Accurate forecasts of electricity prices are crucial for the management of electric power systems and the development of smart applications. European electricity prices have risen substantially and became highly volatile after the Russian invasion of Ukraine, challenging established forecasting methods. Here, we present a Long Short-Term Memory (LSTM) model for the German-Luxembourg day-ahead electricity prices addressing these challenges. The recurrent structure of the LSTM allows the model to adapt to trends, while the joint prediction of both mean and standard deviation enables a probabilistic prediction. Using a physics-inspired approach–superstatistics–to derive an explanation for the statistics of prices, we show that the LSTM model faithfully reproduces both prices and their volatility.

AB - Accurate forecasts of electricity prices are crucial for the management of electric power systems and the development of smart applications. European electricity prices have risen substantially and became highly volatile after the Russian invasion of Ukraine, challenging established forecasting methods. Here, we present a Long Short-Term Memory (LSTM) model for the German-Luxembourg day-ahead electricity prices addressing these challenges. The recurrent structure of the LSTM allows the model to adapt to trends, while the joint prediction of both mean and standard deviation enables a probabilistic prediction. Using a physics-inspired approach–superstatistics–to derive an explanation for the statistics of prices, we show that the LSTM model faithfully reproduces both prices and their volatility.

KW - Adaptation models

KW - Superstatistics

KW - Predictive models

KW - Probabilistic logic

KW - Smart grids

KW - probabilistic forecasting

KW - Electricity prices

KW - Day-ahead electricity prices

KW - German-Luxembourg electricity prices

KW - LSTM

KW - Volatility

KW - Heavy tailed distributions

U2 - 10.1109/ISGTEUROPE56780.2023.10407112

DO - 10.1109/ISGTEUROPE56780.2023.10407112

M3 - Conference Article in proceeding

SN - 979-8-3503-9679-9

BT - 2023 IEEE PES Innovative Smart Grid Technologies Europe (ISGT EUROPE)

PB - IEEE Canada

CY - Grenoble, France, 2023

T2 - IEEE PES ISGT EUROPE 2023

Y2 - 23 October 2023 through 26 October 2023

ER -

Probabilistic Forecasting of Day-Ahead Electricity Prices and their Volatility with LSTMs

Abstract

Conference

Keywords

Access to Document

Fingerprint

Cite this