Dalarna University's logo and link to the university's website

du.sePublications
Change search
CiteExportLink to record
Permanent link

Direct link
Cite
Citation style
  • apa
  • ieee
  • modern-language-association-8th-edition
  • vancouver
  • chicago-author-date
  • chicago-note-bibliography
  • Other style
More styles
Language
  • de-DE
  • en-GB
  • en-US
  • fi-FI
  • nn-NO
  • nn-NB
  • sv-SE
  • Other locale
More languages
Output format
  • html
  • text
  • asciidoc
  • rtf
Anomaly Detection of Heat Energy Usage in District Heating Substations Using LSTM based Variational Autoencoder Combined with Physical Model
Dalarna University, School of Technology and Business Studies, Microdata Analysis.
Dalarna University, School of Technology and Business Studies, Computer Engineering.ORCID iD: 0000-0002-1429-2345
2020 (English)In: 2020 15th IEEE Conference on Industrial Electronics and Applications (ICIEA), 2020, p. 153-158Conference paper, Published paper (Refereed)
Abstract [en]

District heating systems that distribute heat through pipelines to residential and commercial buildings have been widely used in Northern Europe. The energy efficiency of district heating systems is of great interest to energy stakeholders. However, it is not uncommon that district heating systems fail to achieve the expected performance due to various faults. Identification of such rare observations that are different significantly from the majority of the meter readings data plays a vital role in system diagnose. In this study, a new hybrid approach is proposed for anomaly detection of a district heating substation, which consists of a simplified physical model and a Long Short Term Memory based Variational Autoencoder (LSTM VAE). A dataset of an anonymous substation in Sweden is used as a case study. The performance of two state of art models, LSTM and long short term memory based autoencoder (LSTM AE) are evaluated and compared with the LSTM VAE. Experimental results show that LSTM VAE outperforms the baseline models in terms of Area under receiver operating characteristic (ROC) curve (AUC) and F1 score when an optimal threshold is applied.

Place, publisher, year, edition, pages
2020. p. 153-158
Keywords [en]
Heating systems, Substations, Pipelines, Receivers, Meter reading, Stakeholders, Anomaly detection, Energy system, neural networks, computational intelligence, machine learning
National Category
Electrical Engineering, Electronic Engineering, Information Engineering
Research subject
Research Profiles 2009-2020, Complex Systems – Microdata Analysis
Identifiers
URN: urn:nbn:se:du-35622DOI: 10.1109/ICIEA48937.2020.9248108ISI: 000646627000027Scopus ID: 2-s2.0-85097520827ISBN: 978-1-7281-5169-4 (print)OAI: oai:DiVA.org:du-35622DiVA, id: diva2:1508441
Conference
2020 15th IEEE Conference on Industrial Electronics and Applications (ICIEA), Kristiansand, Norway, 9-13 November 2020.
Available from: 2020-12-10 Created: 2020-12-10 Last updated: 2022-09-19Bibliographically approved
In thesis
1. Applications of artificial neural networks for time series data analysis in energy domain
Open this publication in new window or tab >>Applications of artificial neural networks for time series data analysis in energy domain
2020 (English)Licentiate thesis, comprehensive summary (Other academic)
Abstract [en]

With the development of artificial intelligence techniques and increased installation of smart meters in recent years, time series analysis using historical data in the energy domain becomes applicable. In this thesis, microdata analysis approaches are used, which consist of data acquisition, data processing, data analysis and data modelling, aiming to address two research problems in the energy domain. The first research problem is short-term electricity price forecasting of a deregulated market and the second one is anomaly detection of heat energy usage in district heating substations.

As a result of electricity market deregulation, third party suppliers can enter the market and consumers are free to choose electricity suppliers, which leads to a more transparent and competitive market. Accurate short-term electricity price forecasting is crucial to the market participants in terms of maximizing profits, risk management and other short-term market operations. Literature review is performed aiming to identify the suitable methods. It is concluded that long short-term memory (LSTM) based methods are superior to other methods for time series analysis. Since the gating mechanisms of long short-term memory alleviate the problem of gradient vanishing. Another conclusion form the literature is that hybrid approach that consists of two or more artificial intelligence algorithms complimenting each other is more effective to solve complex real world problem. Based on the conclusions derived, a hybrid approach based on bidirectional LSTM (BDLSTM) and Catboost is proposed for short-term electricity price forecasting of NordPool. Performance of support vector regression (SVR), ARIMA, ensemble tree, multi-layer perception (MLP), gated recurrent unit (GRU), BDLSTM and LSTM are evaluated. Experiment results show that BDLSTM outperforms the other models in terms of Mean percentage error (MAPE), root mean square error (RMSE) and mean absolute error (MAE).

Statistics show that market shares of district heating have increased steadily in the past five decades. District heating shares approximately 55% of the heat supply market in Sweden. Therefore, energy efficiency of district heating systems is of great interest to energy stakeholders. Anomalies are rare observations deviated significantly from the majority of the data, and such suspicious observations are important indicators of potential faults. To reduce the financial loss and improve energy efficiency, detecting anomalies from meter readings is essential. Another type of neural network architecture, LSTM variational autoencoder (LSTMVAE) combined with a heat signature model is proposed for anomaly detection using the dataset from an anonymous substation in Sweden. Results show that the proposed method outperforms other two baseline models LSTM, LSTM autoencoder (LSTMAE) in terms of F1 score and AUC.

In this thesis, various approaches based on neural networks are explored to solve different time series data analysis in the energy domain, aiming for supporting decision makings of market participants to maximize profits, enhancing risk managements and improving energy efficiency. Although, two problems domains are covered, methods reviewed and applied in the thesis can be tailored for other energy time series analysis problems as well.

Place, publisher, year, edition, pages
Borlänge: Dalarna University, 2020
Series
Dalarna Licentiate Theses ; 14
Keywords
Deregulated energy market, electricity prices, district heating, energy efficiency, neural networks
National Category
Energy Engineering Energy Systems Computer Sciences
Research subject
Research Profiles 2009-2020, Complex Systems – Microdata Analysis
Identifiers
urn:nbn:se:du-35598 (URN)978-91-88679-08-6 (ISBN)
Presentation
2021-01-22, digital seminar, 10:00 (English)
Opponent
Supervisors
Available from: 2020-12-11 Created: 2020-12-10 Last updated: 2023-08-17Bibliographically approved
2. Machine learning for building energy system analysis
Open this publication in new window or tab >>Machine learning for building energy system analysis
2022 (English)Doctoral thesis, comprehensive summary (Other academic)
Abstract [en]

Buildings account for approximately 40% of the global energy, and Heating, Ventilation, and Air Conditioning (HVAC) contributes to a large proportion of building energy consumption. Two main negative characteristics that contribute to performance degradation and energy waste in an HVAC system are inappropriate control strategies and faults. These two negative characteristics are the focus areas of this thesis. As a type of HVAC system, district heating (DH) plays an important role in sustainable thermal energy production and dominates the heat market in Sweden. However, analysis of district heating has not been well studied in the literature, therefore, district heating substations are selected as the target systems for analysis in this thesis, where night setback control setting, and abrupt fault/anomaly detection are the two problem domains analyzed. One of the main reasons for these problems is the lack of effective methods for detecting such negative characteristics. The knowledge gaps found in these two problem domains are that studies focusing on night setback identification are not found in the literature, and while deep learning (DL) methods show great potential in terms of fault detection and diagnosis (FDD), they are not systematically evaluated, and the application of deep learning methods for HVAC fault detection and diagnosis is scarce. Therefore, this thesis aims at addressing these research gaps. Different machine learning (ML) methods are investigated for night setback identification in the first study using a relatively small dataset. The main contributions of this thesis to microdata analysis are that it proposes a new perspective of sequential data analysis by converting the original time series data into corresponding images. Thus the original research problem is shifted from time series analysis to image analysis, which enables the use of transfer learning to improve model performance. In addition, the proposed approach can be customized and generalized to analyze other building energy systems as well as solving other research problems in the energy domain. The second part of the thesis contributes to the understanding of how data is analysed by performing a meta-analysis. It not only analyses the effectiveness of deep learning methods for performing HVAC FDD quantitively, but also investigates outlier studies and publication bias. These analysis results give future researchers in this field an unbiased perspective and help them avoid overestimating effectiveness, either due to outlier studies or publication bias. The result shows that bidirectional long short-term memory (BDLSTM) with attention mechanism outperforms other benchmarking methods included in this study. However, the labeling process used in this study is difficult, since by inspecting the original time series data, the dataset is balanced, whereas in real life, there are much fewer substations with night setback settings enabled. Most importantly, this approach is difficult to generalize to a large number of substations. To address these drawbacks, the second study reformulated the research problem from time series to image analysis by converting the original time series data into corresponding heat map images. Then, a transfer learning approach is proposed to classify these images. Results show that MobilenetV2 outperforms other benchmark models in terms of f1 score, while Squeezenets are relatively lighted weighted in terms of training time and model sizes and works well This approach provides a feasible solution that can be used for a relatively large, real-world, and imbalanced dataset with overall high accuracy. It is also concluded that night setback patterns are more easily observed from heatmap images, compared to the original time series. Therefore, the proposed approach reduces the complexity of data labeling, and, even if the model fails, it is flexible and easy to switch to manual judgment. In terms of fault detection and diagnosis, the third study reviews and analyzes deep learning methods for HVAC fault detection and diagnosis in a systematic way. It is concluded that long short-term memory (LSTM) and 1D convolutional neural network (CNN) are effective methods for HVAC FDD by design without extra data transformation, while 2D CNNs gain their popularity in recent years due to the increased diversity of collected data. Meta-analysis results show that deep learning methods are effective for HVAC fault detection and diagnosis. However, publication bias exists, which suggests small studies without significant results are not published, hence, the effectiveness of the included studies can be overestimated. Based on the conclusion from the systematic review, an LSTM-based model is proposed for abrupt faults/anomaly detection. Results show that performance varies based on the threshold values. A lower threshold value results in a better recall, while a higher threshold value results in better precision, and there is a trade-off between recall and precision. Therefore, the choice of the threshold depends on the cost of investigating false alarms versus that of missed out faults.

Place, publisher, year, edition, pages
Borlänge: Dalarna University, 2022
Series
Dalarna Doctoral Dissertations ; 22
Keywords
district heating, machine learning, deep learning, HVAC, neural networks
National Category
Computer Engineering
Identifiers
urn:nbn:se:du-42714 (URN)978-91-88679-38-3 (ISBN)
Public defence
2022-11-25, room 322, Borlänge, 13:00 (English)
Opponent
Supervisors
Available from: 2022-10-24 Created: 2022-09-19 Last updated: 2023-03-17Bibliographically approved

Open Access in DiVA

No full text in DiVA

Other links

Publisher's full textScopus

Authority records

Zhang, FanFleyeh, Hasan

Search in DiVA

By author/editor
Zhang, FanFleyeh, Hasan
By organisation
Microdata AnalysisComputer Engineering
Electrical Engineering, Electronic Engineering, Information Engineering

Search outside of DiVA

GoogleGoogle Scholar

doi
isbn
urn-nbn

Altmetric score

doi
isbn
urn-nbn
Total: 519 hits
CiteExportLink to record
Permanent link

Direct link
Cite
Citation style
  • apa
  • ieee
  • modern-language-association-8th-edition
  • vancouver
  • chicago-author-date
  • chicago-note-bibliography
  • Other style
More styles
Language
  • de-DE
  • en-GB
  • en-US
  • fi-FI
  • nn-NO
  • nn-NB
  • sv-SE
  • Other locale
More languages
Output format
  • html
  • text
  • asciidoc
  • rtf