Efficient algorithms to expedite transient stability analysis of power systems

• Main Author: Zadehkhost, Sajjad
• Language: English
• Published: University of British Columbia 2015
• Online Access: http://hdl.handle.net/2429/52816

With rapid increase in complexity of modern power systems, there is a strong need for better computational tools to ensure the reliable operation of electrical grids. These tools need to be accurate, computationally efficient, and capable of using advanced measurement devices. In this context, transient stability assessment (TSA) is an important study that determines system’s dynamic security margins following a major disturbance. The TSA consists of a set of differential-algebraic equations (DAEs), which are typically solved using time-domain simulation (TDS) approach. While being very accurate, the TDS requires significant computational resources when applied to practical power systems. This problem becomes more significant in transient stability monitoring (TSM), wherein the computational performance of the TDS is typically the bottleneck. This research is to investigate available challenges in the TSM applications and develop new algorithms to help realizing a practical monitoring tool for transient stability studies. The thesis focuses on three research thrusts: i) dynamic reduction of power system to reduce problem size; ii) advanced computation approaches to expedite the TDS method; iii) integration of PMU measurements into the TSM. Initially, a new adaptive aggregation algorithm for dynamic reduction is proposed, wherein parameters of generators and structure of transmission network are considered to aggregate coherent generators and create a reduced-order system. Also, a new criterion is defined to monitor validity of the constructed reduced system. It is shown that the proposed technique is more accurate than traditional aggregation methods. To expedite the TDS approach, this thesis presents two new integration techniques, which are called Multi-Decomposition Approach (MDA) and Successive Linearization and Integration Technique (SLIT). In these methods, the nonlinear DAEs are decomposed into a series of linear subsystems, which participate in approximating actual solution. It is demonstrated that sequential and parallel versions of the MDA and SLIT are faster than state-of-the-art integration techniques. Finally, a dynamic state estimator based on Extended Kalman Filter is developed to convert the PMU measurements into a set of state variables suitable for transient stability studies. Computer studies show that the proposed framework provides accurate results in highly disturbed power systems with fairly low PMU sampling rates. === Applied Science, Faculty of === Electrical and Computer Engineering, Department of === Graduate