Automated Data Filtering Approach for ANN Modeling of Distributed Energy Systems: Exploring the Application of Machine Learning
To realize the distributed generation and to make the partnership between the dispatchable units and variable renewable resources work efficiently, accurate and flexible monitoring needs to be implemented. Due to digital transformation in the energy industry, a large amount of data is and will be ca...
| Main Authors: | , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2020-07-01
|
| Series: | Energies |
| Subjects: | |
| Online Access: | https://www.mdpi.com/1996-1073/13/14/3750 |
| _version_ | 1797561770817093632 |
|---|---|
| author | Homam Nikpey Somehsaraei Susmita Ghosh Sayantan Maity Payel Pramanik Sudipta De Mohsen Assadi |
| author_facet | Homam Nikpey Somehsaraei Susmita Ghosh Sayantan Maity Payel Pramanik Sudipta De Mohsen Assadi |
| author_sort | Homam Nikpey Somehsaraei |
| collection | DOAJ |
| description | To realize the distributed generation and to make the partnership between the dispatchable units and variable renewable resources work efficiently, accurate and flexible monitoring needs to be implemented. Due to digital transformation in the energy industry, a large amount of data is and will be captured every day, but the inability to process them in real time challenges the conventional monitoring and maintenance practices. Access to automated and reliable data-filtering tools seems to be crucial for the monitoring of many distributed generation units, avoiding false warnings and improving the reliability. This study aims to evaluate a machine-learning-based methodology for autodetecting outliers from real data, exploring an interdisciplinary solution to replace the conventional manual approach that was very time-consuming and error-prone. The raw data used in this study was collected from experiments on a 100-kW micro gas turbine test rig in Norway. The proposed method uses Density-Based Spatial Clustering of Applications with Noise (DBSCAN) to detect and filter out the outliers. The filtered datasets are used to develop artificial neural networks (ANNs) as a baseline to predict the normal performance of the system for monitoring applications. Results show that the filtering method presented is reliable and fast, minimizing time and resources for data processing. It was also shown that the proposed method has the potential to enhance the performance of the predictive models and ANN-based monitoring. |
| first_indexed | 2024-03-10T18:19:17Z |
| format | Article |
| id | doaj.art-d09e6c3ceaff4b00b81792f5076fc24d |
| institution | Directory Open Access Journal |
| issn | 1996-1073 |
| language | English |
| last_indexed | 2024-03-10T18:19:17Z |
| publishDate | 2020-07-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Energies |
| spelling | doaj.art-d09e6c3ceaff4b00b81792f5076fc24d2023-11-20T07:29:54ZengMDPI AGEnergies1996-10732020-07-011314375010.3390/en13143750Automated Data Filtering Approach for ANN Modeling of Distributed Energy Systems: Exploring the Application of Machine LearningHomam Nikpey Somehsaraei0Susmita Ghosh1Sayantan Maity2Payel Pramanik3Sudipta De4Mohsen Assadi5Department of Energy and Petroleum Engineering, University of Stavanger, 4036 Stavanger, NorwayDepartment of Computer Science and Engineering, Jadavpur University, Kolkata 700032, IndiaDepartment of Computer Science and Engineering, Jadavpur University, Kolkata 700032, IndiaDepartment of Computer Science and Engineering, Jadavpur University, Kolkata 700032, IndiaDepartment of Computer Science and Engineering, Jadavpur University, Kolkata 700032, IndiaDepartment of Energy and Petroleum Engineering, University of Stavanger, 4036 Stavanger, NorwayTo realize the distributed generation and to make the partnership between the dispatchable units and variable renewable resources work efficiently, accurate and flexible monitoring needs to be implemented. Due to digital transformation in the energy industry, a large amount of data is and will be captured every day, but the inability to process them in real time challenges the conventional monitoring and maintenance practices. Access to automated and reliable data-filtering tools seems to be crucial for the monitoring of many distributed generation units, avoiding false warnings and improving the reliability. This study aims to evaluate a machine-learning-based methodology for autodetecting outliers from real data, exploring an interdisciplinary solution to replace the conventional manual approach that was very time-consuming and error-prone. The raw data used in this study was collected from experiments on a 100-kW micro gas turbine test rig in Norway. The proposed method uses Density-Based Spatial Clustering of Applications with Noise (DBSCAN) to detect and filter out the outliers. The filtered datasets are used to develop artificial neural networks (ANNs) as a baseline to predict the normal performance of the system for monitoring applications. Results show that the filtering method presented is reliable and fast, minimizing time and resources for data processing. It was also shown that the proposed method has the potential to enhance the performance of the predictive models and ANN-based monitoring.https://www.mdpi.com/1996-1073/13/14/3750distributed energy generationautomated data filteringdensity-based clusteringANN-based predictive model |
| spellingShingle | Homam Nikpey Somehsaraei Susmita Ghosh Sayantan Maity Payel Pramanik Sudipta De Mohsen Assadi Automated Data Filtering Approach for ANN Modeling of Distributed Energy Systems: Exploring the Application of Machine Learning Energies distributed energy generation automated data filtering density-based clustering ANN-based predictive model |
| title | Automated Data Filtering Approach for ANN Modeling of Distributed Energy Systems: Exploring the Application of Machine Learning |
| title_full | Automated Data Filtering Approach for ANN Modeling of Distributed Energy Systems: Exploring the Application of Machine Learning |
| title_fullStr | Automated Data Filtering Approach for ANN Modeling of Distributed Energy Systems: Exploring the Application of Machine Learning |
| title_full_unstemmed | Automated Data Filtering Approach for ANN Modeling of Distributed Energy Systems: Exploring the Application of Machine Learning |
| title_short | Automated Data Filtering Approach for ANN Modeling of Distributed Energy Systems: Exploring the Application of Machine Learning |
| title_sort | automated data filtering approach for ann modeling of distributed energy systems exploring the application of machine learning |
| topic | distributed energy generation automated data filtering density-based clustering ANN-based predictive model |
| url | https://www.mdpi.com/1996-1073/13/14/3750 |
| work_keys_str_mv | AT homamnikpeysomehsaraei automateddatafilteringapproachforannmodelingofdistributedenergysystemsexploringtheapplicationofmachinelearning AT susmitaghosh automateddatafilteringapproachforannmodelingofdistributedenergysystemsexploringtheapplicationofmachinelearning AT sayantanmaity automateddatafilteringapproachforannmodelingofdistributedenergysystemsexploringtheapplicationofmachinelearning AT payelpramanik automateddatafilteringapproachforannmodelingofdistributedenergysystemsexploringtheapplicationofmachinelearning AT sudiptade automateddatafilteringapproachforannmodelingofdistributedenergysystemsexploringtheapplicationofmachinelearning AT mohsenassadi automateddatafilteringapproachforannmodelingofdistributedenergysystemsexploringtheapplicationofmachinelearning |