Positive-Unlabelled Learning based Novelty Detection for Industrial Chillers
Chiller systems are used in many different applications in both the industrial and the commercial sector. They are considered major energy consumers and thus contribute a non-negligible factor to environmental pollution as well as to the overall operating cost. In addition, chillers, especially in...
| Main Authors: | , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
TIB Open Publishing
2021-06-01
|
| Series: | TH Wildau Engineering and Natural Sciences Proceedings |
| Subjects: | |
| Online Access: | https://www.tib-op.org/ojs/index.php/th-wildau-ensp/article/view/32 |
| _version_ | 1811287714895495168 |
|---|---|
| author | Ron van de Sand Sandra Corasaniti Jörg Reiff-Stephan |
| author_facet | Ron van de Sand Sandra Corasaniti Jörg Reiff-Stephan |
| author_sort | Ron van de Sand |
| collection | DOAJ |
| description |
Chiller systems are used in many different applications in both the industrial and the commercial sector. They are considered major energy consumers and thus contribute a non-negligible factor to environmental pollution as well as to the overall operating cost. In addition, chillers, especially in industrial applications, are often associated with high reliability requirements, as unplanned system downtimes are usually costly. As many studies over the past decades have shown, the presence of faults can lead to significant performance degradation and thus higher energy consumption of these systems. Thus, data-driven fault detection plays an ever-increasing role in terms of energy efficient control strategies. However, labelled data to train associated algorithms are often only available to a limited extent, which consequently inhibits the broad application of such technologies. Therefore, this paper presents an approach that exploits only a small amount of labelled and large amounts of unlabelled data in the training phase in order to detect fault related anomalies. For this, the model utilizes the residual space of the data transformed through principal component analyses in conjunction with a biased support vector machine, which can be ascribed to the concept of semi-supervised learning, or more specifically, positive-unlabelled learning.
|
| first_indexed | 2024-04-13T03:23:07Z |
| format | Article |
| id | doaj.art-ca15bb4a5d9e4264922c96b496f5f9cd |
| institution | Directory Open Access Journal |
| issn | 2748-8829 |
| language | English |
| last_indexed | 2024-04-13T03:23:07Z |
| publishDate | 2021-06-01 |
| publisher | TIB Open Publishing |
| record_format | Article |
| series | TH Wildau Engineering and Natural Sciences Proceedings |
| spelling | doaj.art-ca15bb4a5d9e4264922c96b496f5f9cd2022-12-22T03:04:43ZengTIB Open PublishingTH Wildau Engineering and Natural Sciences Proceedings2748-88292021-06-01110.52825/thwildauensp.v1i.32Positive-Unlabelled Learning based Novelty Detection for Industrial ChillersRon van de Sand0Sandra Corasaniti1Jörg Reiff-Stephan2Technical University of Applied Sciences Wildau University of Rome Tor Vergata Technical University of Applied Sciences Wildau Chiller systems are used in many different applications in both the industrial and the commercial sector. They are considered major energy consumers and thus contribute a non-negligible factor to environmental pollution as well as to the overall operating cost. In addition, chillers, especially in industrial applications, are often associated with high reliability requirements, as unplanned system downtimes are usually costly. As many studies over the past decades have shown, the presence of faults can lead to significant performance degradation and thus higher energy consumption of these systems. Thus, data-driven fault detection plays an ever-increasing role in terms of energy efficient control strategies. However, labelled data to train associated algorithms are often only available to a limited extent, which consequently inhibits the broad application of such technologies. Therefore, this paper presents an approach that exploits only a small amount of labelled and large amounts of unlabelled data in the training phase in order to detect fault related anomalies. For this, the model utilizes the residual space of the data transformed through principal component analyses in conjunction with a biased support vector machine, which can be ascribed to the concept of semi-supervised learning, or more specifically, positive-unlabelled learning. https://www.tib-op.org/ojs/index.php/th-wildau-ensp/article/view/32Chiller CBM Machine LearningEnergy Efficiency |
| spellingShingle | Ron van de Sand Sandra Corasaniti Jörg Reiff-Stephan Positive-Unlabelled Learning based Novelty Detection for Industrial Chillers TH Wildau Engineering and Natural Sciences Proceedings Chiller CBM Machine Learning Energy Efficiency |
| title | Positive-Unlabelled Learning based Novelty Detection for Industrial Chillers |
| title_full | Positive-Unlabelled Learning based Novelty Detection for Industrial Chillers |
| title_fullStr | Positive-Unlabelled Learning based Novelty Detection for Industrial Chillers |
| title_full_unstemmed | Positive-Unlabelled Learning based Novelty Detection for Industrial Chillers |
| title_short | Positive-Unlabelled Learning based Novelty Detection for Industrial Chillers |
| title_sort | positive unlabelled learning based novelty detection for industrial chillers |
| topic | Chiller CBM Machine Learning Energy Efficiency |
| url | https://www.tib-op.org/ojs/index.php/th-wildau-ensp/article/view/32 |
| work_keys_str_mv | AT ronvandesand positiveunlabelledlearningbasednoveltydetectionforindustrialchillers AT sandracorasaniti positiveunlabelledlearningbasednoveltydetectionforindustrialchillers AT jorgreiffstephan positiveunlabelledlearningbasednoveltydetectionforindustrialchillers |