Thesis
An ensemble based algorithm for synchrophasor data anomaly detection
Washington State University
Master of Science (MS), Washington State University
2017
Handle:
https://hdl.handle.net/2376/100312
Abstract
Phasor Measurement Units (PMUs) provide the high resolution real-time synchronized phasor data for power system operation and control. PMUs feeds data to local Phasor Data Concentrator (PDC) and controllers located at substations before data delivery to control centers. PMU data quality is critical in wide-area applications. A number of methods are developed to detect anomalies in time series data, tailored for different scenarios. However, in practical PMU data analysis, applying a standalone method (with fixed parameters) does not always achieve high accuracy. In this thesis, an unsupervised ensemble learning method is applied to develop fast, scalable methods for PMU bad data detection. The ensemble based detection invokes a set of base detectors to generate anomaly scores of the PMU data, and makes decisions by aggregating the scores from each detector. This thesis develops 1) a learning algorithm to train the ensemble learner, and 2) an online algorithm for inferencing the anomaly scores with the ensemble learner. Using both simulated and real-world PMU data, this work validates the performance of the developed approach to detect diversified errors with high accuracy. Developed algorithm and outperforms its counterparts using single standalone method. In addition, impact of bad data detection algorithms on the wide area applications have been also provided.
Metrics
16 File views/ downloads
22 Record Views
Details
- Title
- An ensemble based algorithm for synchrophasor data anomaly detection
- Creators
- Mengze Zhou
- Contributors
- Anurag K. Srivastava (Degree Supervisor)
- Awarding Institution
- Washington State University
- Academic Unit
- Electrical Engineering and Computer Science, School of
- Theses and Dissertations
- Master of Science (MS), Washington State University
- Publisher
- Washington State University; [Pullman, Washington] :
- Identifiers
- 99900524883601842
- Language
- English
- Resource Type
- Thesis