Dissertation
DATA-DRIVEN ALGORITHMS FOR POWER SYSTEM LOAD MODELPARAMETER ESTIMATION AND VOLTAGE STABILITY ASSESSMENT
Washington State University
Doctor of Philosophy (PhD), Washington State University
01/2021
DOI:
https://doi.org/10.7273/000005444
Handle:
https://hdl.handle.net/2376/119014
Abstract
Voltage-dependent behavior of power system load greatly influences the power system's operational characteristics, and thus appropriate load modeling is vital for accurate power system studies. The availability of fast and accurate PMU measurements at substations encourages researchers to develop aggregate load models that can be utilized in transmission voltage stability assessment. The power system's voltage stability is immensely important for secure and optimal operation and is directly impacted by load behavior. The power system's long-term voltage stability status is affected by the steady-state behavior of the power system load. It is crucial to consider the composite (dynamic + static) behavior of load for determining short-term voltage stability status. The development of accurate voltage stability assessment tools for different timing paradigms is paramount for the secure and reliable operation of the power grid. \parThe first part of this thesis considers the development of robust load parameter estimation algorithms. This thesis develops several algorithms for accurate steady-state load parameter estimation. Firstly the regression-based load parameter estimation approaches normally considered in the literature are enhanced by leveraging more information about the numerical characteristics of PMU data. Secondly, a framework is developed to use powerful machine-learning tools for steady-state load parameter estimation. A composite load parameter estimation algorithm is developed to consider electronic load behavior in aggregate load parameter estimation. The developed load parameter estimation algorithms contribute to state of the art by improving load parameter tracking accuracy.
The second part of this thesis considers the development of novel voltage stability assessment tools considering the evolving behavior of power systems. A hybrid voltage stability assessment tool is developed that appropriately considers generators' reactive power reserve status while computing the voltage stability index. Secondly, a continuation power flow tool is developed to consider the impact of the distribution system infrastructure on voltage stability status in detail. Such kind of a tool is needed to accommodate the active behavior of distribution for voltage stability assessment. Thirdly a short-term voltage stability assessment framework is developed using a 1D-convolutional neural network for early voltage collapse detection and severity quantification of short-term voltage stability events.
Metrics
10 File views/ downloads
48 Record Views
Details
- Title
- DATA-DRIVEN ALGORITHMS FOR POWER SYSTEM LOAD MODELPARAMETER ESTIMATION AND VOLTAGE STABILITY ASSESSMENT
- Creators
- Syed Muhammad Hur Rizvi
- Contributors
- Anurag K. Srivastava (Advisor)Anjan Bose (Committee Member)Vaithianath Mani Venkatasubramanian (Committee Member)
- Awarding Institution
- Washington State University
- Academic Unit
- School of Electrical Engineering and Computer Science
- Theses and Dissertations
- Doctor of Philosophy (PhD), Washington State University
- Publisher
- Washington State University
- Number of pages
- 260
- Identifiers
- 99900592359901842
- Language
- English
- Resource Type
- Dissertation