Discovering Power System Dynamics through Time-Frequency Representation of Ambient Data

Simple item page
dc.contributor.authorDixit, Vishal Sateeshen
dc.contributor.committeechairCenteno, Virgilio A.en
dc.contributor.committeechairLiu, Chen-Chingen
dc.contributor.committeememberMishra, Chetanen
dc.contributor.committeememberKekatos, Vasileiosen
dc.contributor.departmentElectrical Engineeringen
dc.date.accessioned2024-06-13T08:02:31Zen
dc.date.available2024-06-13T08:02:31Zen
dc.date.issued2024-06-12en
dc.description.abstractWith the proliferation of renewable energy and its integration into the modern power grid, we face some new issues. Aside from the increased switching rate, which results in faster dynamic behavior, realistic models for these Inverter-Based Resources (IBRs) are not widely available. This complicates researching the behavior of this quickly changing system, and without proper models, simulations may not be totally reliable. To address this, it is recommended that measurement data be used, which includes the entire grid and all of its unique characteristics. Signal processing techniques have been employed exclusively to construct spectrograms, which are time-frequency representations of a signal's spectral information. These spectrograms show ridges that represent the system's changing modes. It can be extremely beneficial to track these modes and generate labeled data representing the evolution of modes as the system evolves. This labeled data can aid in the development of correlation and causation hypotheses linking specific abnormal behaviors to proximity to instability. This can also assist analyze these IBRs and identify flaws in their modeling. This thesis describes a step-by-step process for creating spectrograms, reducing them for better visualization, and then estimating mode evolution with a ridge-tracking algorithm based on penalized jump criteria. The results show that the tracker works effectively with both synthetic and real PMU data.en
dc.description.abstractgeneralBy 2050, the number of electric cars on the road will increase almost tenfold, and renewable energy will make up almost 50% of the global power mix. While this is great news for the environment, it also poses new challenges to the power sector in ensuring the reliable delivery of clean energy. To address this, we need to collect real-time information about the system. A spectrogram is a visual representation of the power grid's dynamic behavior, providing essential information about frequency and power. Despite its extensive use in biomedical data, this tool is not used much in the power system industry. Spectrograms can be used as a forensic tool or preventive measure to detect system instability. Our project aims to track the system's dynamic behavior using ambient data, which is shown to be rich in information. The proposed algorithm suggests a detailed step-by-step methodology to use this tool for system identification and monitoring. The work's novelty lies in the tracking algorithm developed to identify and track the spectral components in a time-frequency representation.en
dc.description.degreeMaster of Scienceen
dc.format.mediumETDen
dc.identifier.othervt_gsexam:40940en
dc.identifier.urihttps://hdl.handle.net/10919/119423en
dc.language.isoenen
dc.publisherVirginia Techen
dc.rightsIn Copyrighten
dc.rights.urihttp://rightsstatements.org/vocab/InC/1.0/en
dc.subjectSpectrogramen
dc.subjectAmbient dataen
dc.subjectTime-frequencyen
dc.subjectspectral analysisen
dc.subjectsynchrophasorsen
dc.titleDiscovering Power System Dynamics through Time-Frequency Representation of Ambient Dataen
dc.typeThesisen
thesis.degree.disciplineElectrical Engineeringen
thesis.degree.grantorVirginia Polytechnic Institute and State Universityen
thesis.degree.levelmastersen
thesis.degree.nameMaster of Scienceen

Files

Original bundle
Now showing 1 - 1 of 1
Name:
Dixit_VS_T_2024.pdf
Size:
13.5 MB
Format:
Adobe Portable Document Format
Download

Collections

Masters Theses