Hybrid Optimization based on Deep Learning Approach for Short-Term Load Forecast of Electricity Demand in Buildings

Due to the growing popularity of microgrids in buildings, the foreseeable electricity demand for a building draws the attention of many researchers. The precise short-term demand forecast efficiently directs building managers and operators for interactions with electrical distribution systems, daily...

Full description

Saved in:
Bibliographic Details
Main Authors: Charan Sekhar Makula, Ratna Dahiya
Format: Article
Language:English
Published: OICC Press 2024-06-01
Series:Majlesi Journal of Electrical Engineering
Subjects:
Online Access:https://oiccpress.com/mjee/article/view/7986
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850085462033563648
author Charan Sekhar Makula
Ratna Dahiya
author_facet Charan Sekhar Makula
Ratna Dahiya
author_sort Charan Sekhar Makula
collection DOAJ
description Due to the growing popularity of microgrids in buildings, the foreseeable electricity demand for a building draws the attention of many researchers. The precise short-term demand forecast efficiently directs building managers and operators for interactions with electrical distribution systems, daily operational decisions, and energy conservation. This research proposes a hybrid optimization-based deep learning (DL) approach to increase the accuracy of short-term forecasts. The present work employs the Bilateral Long Short-Term Memory (BiLSTM) network-based DL technique because the BiLSTM technique has an exceptional ability to manage nonlinear interactions in data and learn the temporal dependencies. The performance of the BiLSTM technique is improved by using the optimally determined hyperparameters via a hybrid optimization algorithm that combines particle swarm optimization (PSO) and grey wolf optimization (GWO). The exploration ability of GWO and exploitation ability of PSO are effectively combined in the hybrid optimization GWO-PSO. The performance of the recommended approach is assessed using a case study of an educational building. The performance of the proposed model is compared to existing nonoptimal BiLSTM and single optimization-based BiLSTM for short-term load forecast.
format Article
id doaj-art-b5b83012dd4f418e92dd1cc3b663f44b
institution DOAJ
issn 2345-377X
2345-3796
language English
publishDate 2024-06-01
publisher OICC Press
record_format Article
series Majlesi Journal of Electrical Engineering
spelling doaj-art-b5b83012dd4f418e92dd1cc3b663f44b2025-08-20T02:43:43ZengOICC PressMajlesi Journal of Electrical Engineering2345-377X2345-37962024-06-0118210.57647/j.mjee.2024.1802.32Hybrid Optimization based on Deep Learning Approach for Short-Term Load Forecast of Electricity Demand in BuildingsCharan Sekhar MakulaRatna DahiyaDue to the growing popularity of microgrids in buildings, the foreseeable electricity demand for a building draws the attention of many researchers. The precise short-term demand forecast efficiently directs building managers and operators for interactions with electrical distribution systems, daily operational decisions, and energy conservation. This research proposes a hybrid optimization-based deep learning (DL) approach to increase the accuracy of short-term forecasts. The present work employs the Bilateral Long Short-Term Memory (BiLSTM) network-based DL technique because the BiLSTM technique has an exceptional ability to manage nonlinear interactions in data and learn the temporal dependencies. The performance of the BiLSTM technique is improved by using the optimally determined hyperparameters via a hybrid optimization algorithm that combines particle swarm optimization (PSO) and grey wolf optimization (GWO). The exploration ability of GWO and exploitation ability of PSO are effectively combined in the hybrid optimization GWO-PSO. The performance of the recommended approach is assessed using a case study of an educational building. The performance of the proposed model is compared to existing nonoptimal BiLSTM and single optimization-based BiLSTM for short-term load forecast.https://oiccpress.com/mjee/article/view/7986Bilateral Long Short Term MemoryBuildingsGWOPSOHyperparameter Optimization.Short Term Load forecast
spellingShingle Charan Sekhar Makula
Ratna Dahiya
Hybrid Optimization based on Deep Learning Approach for Short-Term Load Forecast of Electricity Demand in Buildings
Majlesi Journal of Electrical Engineering
Bilateral Long Short Term Memory
Buildings
GWOPSO
Hyperparameter Optimization.
Short Term Load forecast
title Hybrid Optimization based on Deep Learning Approach for Short-Term Load Forecast of Electricity Demand in Buildings
title_full Hybrid Optimization based on Deep Learning Approach for Short-Term Load Forecast of Electricity Demand in Buildings
title_fullStr Hybrid Optimization based on Deep Learning Approach for Short-Term Load Forecast of Electricity Demand in Buildings
title_full_unstemmed Hybrid Optimization based on Deep Learning Approach for Short-Term Load Forecast of Electricity Demand in Buildings
title_short Hybrid Optimization based on Deep Learning Approach for Short-Term Load Forecast of Electricity Demand in Buildings
title_sort hybrid optimization based on deep learning approach for short term load forecast of electricity demand in buildings
topic Bilateral Long Short Term Memory
Buildings
GWOPSO
Hyperparameter Optimization.
Short Term Load forecast
url https://oiccpress.com/mjee/article/view/7986
work_keys_str_mv AT charansekharmakula hybridoptimizationbasedondeeplearningapproachforshorttermloadforecastofelectricitydemandinbuildings
AT ratnadahiya hybridoptimizationbasedondeeplearningapproachforshorttermloadforecastofelectricitydemandinbuildings