Research on Predictive Analysis Method of Building Energy Consumption Based on TCN-BiGru-Attention

Building energy consumption prediction has always played a significant role in assessing building energy efficiency, building commissioning, and detecting and diagnosing building system faults. With the progress of society and economic development, building energy consumption is growing rapidly. The...

Full description

Saved in:
Bibliographic Details
Main Authors: Sijia Fu, Rui Zhu, Feiyang Yu
Format: Article
Language:English
Published: MDPI AG 2024-10-01
Series:Applied Sciences
Subjects:
Online Access:https://www.mdpi.com/2076-3417/14/20/9373
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850205280595345408
author Sijia Fu
Rui Zhu
Feiyang Yu
author_facet Sijia Fu
Rui Zhu
Feiyang Yu
author_sort Sijia Fu
collection DOAJ
description Building energy consumption prediction has always played a significant role in assessing building energy efficiency, building commissioning, and detecting and diagnosing building system faults. With the progress of society and economic development, building energy consumption is growing rapidly. Therefore, accurate and effective building energy consumption prediction is the basis of energy conservation. Although there are currently a large number of energy consumption research methods, each method has different applicability and advantages and disadvantages. This study proposes a Time Convolution Network model based on an attention mechanism, which combines the ability of the Time Convolution Network model to capture ultra-long time series information with the ability of the BiGRU model to integrate contextual information, improve model parallelism, and reduce the risk of overfitting. In order to tune the hyperparameters in the structure of this prediction model, such as the learning rate, the size of the convolutional kernel, and the number of recurrent units, this study chooses to use the Golden Jackal Optimization Algorithm for optimization. The study shows that this optimized model has better accuracy than models such as LSTM, SVM, and CNN.
format Article
id doaj-art-44127ca2f37841259e95bdb42a339334
institution OA Journals
issn 2076-3417
language English
publishDate 2024-10-01
publisher MDPI AG
record_format Article
series Applied Sciences
spelling doaj-art-44127ca2f37841259e95bdb42a3393342025-08-20T02:11:08ZengMDPI AGApplied Sciences2076-34172024-10-011420937310.3390/app14209373Research on Predictive Analysis Method of Building Energy Consumption Based on TCN-BiGru-AttentionSijia Fu0Rui Zhu1Feiyang Yu2School of Mathematics and Physics, Shanghai University of Electric Power, Shanghai 200090, ChinaSchool of Mathematics and Physics, Shanghai University of Electric Power, Shanghai 200090, ChinaSchool of Mathematics and Physics, Shanghai University of Electric Power, Shanghai 200090, ChinaBuilding energy consumption prediction has always played a significant role in assessing building energy efficiency, building commissioning, and detecting and diagnosing building system faults. With the progress of society and economic development, building energy consumption is growing rapidly. Therefore, accurate and effective building energy consumption prediction is the basis of energy conservation. Although there are currently a large number of energy consumption research methods, each method has different applicability and advantages and disadvantages. This study proposes a Time Convolution Network model based on an attention mechanism, which combines the ability of the Time Convolution Network model to capture ultra-long time series information with the ability of the BiGRU model to integrate contextual information, improve model parallelism, and reduce the risk of overfitting. In order to tune the hyperparameters in the structure of this prediction model, such as the learning rate, the size of the convolutional kernel, and the number of recurrent units, this study chooses to use the Golden Jackal Optimization Algorithm for optimization. The study shows that this optimized model has better accuracy than models such as LSTM, SVM, and CNN.https://www.mdpi.com/2076-3417/14/20/9373energy consumption predictiontemporal convolutional network (TCN)BiGRUattentiongolden jackal optimization
spellingShingle Sijia Fu
Rui Zhu
Feiyang Yu
Research on Predictive Analysis Method of Building Energy Consumption Based on TCN-BiGru-Attention
Applied Sciences
energy consumption prediction
temporal convolutional network (TCN)
BiGRU
attention
golden jackal optimization
title Research on Predictive Analysis Method of Building Energy Consumption Based on TCN-BiGru-Attention
title_full Research on Predictive Analysis Method of Building Energy Consumption Based on TCN-BiGru-Attention
title_fullStr Research on Predictive Analysis Method of Building Energy Consumption Based on TCN-BiGru-Attention
title_full_unstemmed Research on Predictive Analysis Method of Building Energy Consumption Based on TCN-BiGru-Attention
title_short Research on Predictive Analysis Method of Building Energy Consumption Based on TCN-BiGru-Attention
title_sort research on predictive analysis method of building energy consumption based on tcn bigru attention
topic energy consumption prediction
temporal convolutional network (TCN)
BiGRU
attention
golden jackal optimization
url https://www.mdpi.com/2076-3417/14/20/9373
work_keys_str_mv AT sijiafu researchonpredictiveanalysismethodofbuildingenergyconsumptionbasedontcnbigruattention
AT ruizhu researchonpredictiveanalysismethodofbuildingenergyconsumptionbasedontcnbigruattention
AT feiyangyu researchonpredictiveanalysismethodofbuildingenergyconsumptionbasedontcnbigruattention