Deep learning based on multi-decomposition for short-term load forecasting

Seon Hyeog Kim, Gyul Lee, Gu Young Kwon, Do In Kim, Yong June Shin

Research output: Contribution to journalArticle

3 Citations (Scopus)

Abstract

Load forecasting is a key issue for efficient real-time energy management in smart grids. To control the load using demand side management accurately, load forecasting should be predicted in the short term. With the advent of advanced measuring infrastructure, it is possible to measure energy consumption at sampling rates up to every 5 min and analyze the load profile of small-scale energy groups, such as individual buildings. This paper presents applications of deep learning using feature decomposition for improving the accuracy of load forecasting. The load profile is decomposed into a weekly load profile and then decomposed into intrinsic mode functions by variational mode decomposition to capture periodic features. Then, a long short-term memory network model is trained by three-dimensional input data with three-step regularization. Finally, the prediction results of all intrinsic mode functions are combined with advanced measuring infrastructure measured in the previous steps to determine an aggregated output for load forecasting. The results are validated by applications to real-world data from smart buildings, and the performance of the proposed approach is assessed by comparing the predicted results with those of conventional methods, nonlinear autoregressive networks with exogenous inputs, and long short-term memory network-based feature decomposition.

Original languageEnglish
Article number3433
JournalEnergies
Volume11
Issue number12
DOIs
Publication statusPublished - 2018 Dec 1

Fingerprint

Short-term Load Forecasting
Load Forecasting
Decomposition
Intrinsic Mode Function
Decompose
Infrastructure
Smart Grid
Energy Management
Memory Term
Memory Model
Nonlinear networks
Intelligent buildings
Network Model
Energy Consumption
Energy management
Regularization
Real-time
Energy utilization
Three-dimensional
Learning

All Science Journal Classification (ASJC) codes

  • Renewable Energy, Sustainability and the Environment
  • Energy Engineering and Power Technology
  • Energy (miscellaneous)
  • Control and Optimization
  • Electrical and Electronic Engineering

Cite this

Kim, Seon Hyeog ; Lee, Gyul ; Kwon, Gu Young ; Kim, Do In ; Shin, Yong June. / Deep learning based on multi-decomposition for short-term load forecasting. In: Energies. 2018 ; Vol. 11, No. 12.
@article{01aa81bf175d42e7b34868a7a76f93a0,
title = "Deep learning based on multi-decomposition for short-term load forecasting",
abstract = "Load forecasting is a key issue for efficient real-time energy management in smart grids. To control the load using demand side management accurately, load forecasting should be predicted in the short term. With the advent of advanced measuring infrastructure, it is possible to measure energy consumption at sampling rates up to every 5 min and analyze the load profile of small-scale energy groups, such as individual buildings. This paper presents applications of deep learning using feature decomposition for improving the accuracy of load forecasting. The load profile is decomposed into a weekly load profile and then decomposed into intrinsic mode functions by variational mode decomposition to capture periodic features. Then, a long short-term memory network model is trained by three-dimensional input data with three-step regularization. Finally, the prediction results of all intrinsic mode functions are combined with advanced measuring infrastructure measured in the previous steps to determine an aggregated output for load forecasting. The results are validated by applications to real-world data from smart buildings, and the performance of the proposed approach is assessed by comparing the predicted results with those of conventional methods, nonlinear autoregressive networks with exogenous inputs, and long short-term memory network-based feature decomposition.",
author = "Kim, {Seon Hyeog} and Gyul Lee and Kwon, {Gu Young} and Kim, {Do In} and Shin, {Yong June}",
year = "2018",
month = "12",
day = "1",
doi = "10.3390/en11123433",
language = "English",
volume = "11",
journal = "Energies",
issn = "1996-1073",
publisher = "Multidisciplinary Digital Publishing Institute (MDPI)",
number = "12",

}

Deep learning based on multi-decomposition for short-term load forecasting. / Kim, Seon Hyeog; Lee, Gyul; Kwon, Gu Young; Kim, Do In; Shin, Yong June.

In: Energies, Vol. 11, No. 12, 3433, 01.12.2018.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Deep learning based on multi-decomposition for short-term load forecasting

AU - Kim, Seon Hyeog

AU - Lee, Gyul

AU - Kwon, Gu Young

AU - Kim, Do In

AU - Shin, Yong June

PY - 2018/12/1

Y1 - 2018/12/1

N2 - Load forecasting is a key issue for efficient real-time energy management in smart grids. To control the load using demand side management accurately, load forecasting should be predicted in the short term. With the advent of advanced measuring infrastructure, it is possible to measure energy consumption at sampling rates up to every 5 min and analyze the load profile of small-scale energy groups, such as individual buildings. This paper presents applications of deep learning using feature decomposition for improving the accuracy of load forecasting. The load profile is decomposed into a weekly load profile and then decomposed into intrinsic mode functions by variational mode decomposition to capture periodic features. Then, a long short-term memory network model is trained by three-dimensional input data with three-step regularization. Finally, the prediction results of all intrinsic mode functions are combined with advanced measuring infrastructure measured in the previous steps to determine an aggregated output for load forecasting. The results are validated by applications to real-world data from smart buildings, and the performance of the proposed approach is assessed by comparing the predicted results with those of conventional methods, nonlinear autoregressive networks with exogenous inputs, and long short-term memory network-based feature decomposition.

AB - Load forecasting is a key issue for efficient real-time energy management in smart grids. To control the load using demand side management accurately, load forecasting should be predicted in the short term. With the advent of advanced measuring infrastructure, it is possible to measure energy consumption at sampling rates up to every 5 min and analyze the load profile of small-scale energy groups, such as individual buildings. This paper presents applications of deep learning using feature decomposition for improving the accuracy of load forecasting. The load profile is decomposed into a weekly load profile and then decomposed into intrinsic mode functions by variational mode decomposition to capture periodic features. Then, a long short-term memory network model is trained by three-dimensional input data with three-step regularization. Finally, the prediction results of all intrinsic mode functions are combined with advanced measuring infrastructure measured in the previous steps to determine an aggregated output for load forecasting. The results are validated by applications to real-world data from smart buildings, and the performance of the proposed approach is assessed by comparing the predicted results with those of conventional methods, nonlinear autoregressive networks with exogenous inputs, and long short-term memory network-based feature decomposition.

UR - http://www.scopus.com/inward/record.url?scp=85059255500&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85059255500&partnerID=8YFLogxK

U2 - 10.3390/en11123433

DO - 10.3390/en11123433

M3 - Article

VL - 11

JO - Energies

JF - Energies

SN - 1996-1073

IS - 12

M1 - 3433

ER -