Applications of synchronized phasor measurements for state estimation, voltage stability, and damping control

Abstract

For future work, we have plans to extend the PSE observable region to the entire New York State, and apply the method to other control areas, such as ISO-New England or areas within the US Western Interconnection. The voltage stability application is being further developed in a project funded by the Consortium for Electric Reliability Technology Solutions (CERTS) for the US Western Interconnection. Future work on the adaptive damping controller includes development of a discrete-time system realization and application to larger systems with multiple modes of oscillation and multiple distributed damping controllers.

Phasor measurement units (PMUs) are being installed in power grids all over the world, providing system operators with many new benefits. Compared to conventional Supervisory Control and Data Acquisition (SCADA) measurements, the main advantages of PMU data are higher sampling rates, phase angle measurement, and synchronization by Global Positioning System (GPS) clock signal. The high-fidelity measurements provided by PMU technology provide increased system visibility to grid operators, allowing them to observe and track the dynamic propagation of disturbances throughout the system. As system operators look to leverage their investment in PMUs, they seek applications that make use of these synchronized phasor measurements (also known as synchrophasors) to improve energy reliability, efficiency, and economy. In this thesis, we present three practical applications of synchronized phasor measurements: phasor state estimation for improving PMU data quality, voltage stability analysis for power transfer-limited interfaces, and a networked adaptive controller for damping interarea oscillations.

The first step to incorporating synchrophasors in power system operation is ensuring the measurements are of good quality, which is usually accomplished using a state estimator. In previous work at RPI, a method for phasor state estimation (PSE) was proposed to estimate and eliminate phase angle biases and jumps associated with erroneous measurements. The method was applied to real PMU data a very limited portion of a transmission network.

In this work, we extend the PSE method to include current phasor scaling error correction and estimation of transmission line parameters and transformer tap ratios. The error correction and line parameter estimation are provided simultaneously with the system state, using an augmented state vector in a nonlinear Gauss-Newton iteration. We applied the method to real PMU data from several disturbance events on the Central New York transmission system. Using data from 6 PMUs deployed on 6 substations, the PSE provides observability for a total of 13 substations, including power transfer-limited interfaces. Finally, we analyze the residuals after the least-square estimation to demonstrate the data quality improvement of our method.

The second proposed application in this thesis uses the results of the PSE method to calculate the voltage stability margin of a power transfer interface. The main concept is to use the PMU data to construct a reduced model of the external (unobservable) system. Then, a loadflow method can be used to determine the maximum power transfer of constrained transmission path and PV curves can be computed. Previous research at RPI used PMU data from a single substation to compute PV curves for a major load area. In this work, we leverage multiple PMUs, a larger observable region, and improved data quality from the PSE to improve the accuracy of our models. We construct Thévenin equivalent models for external power injections and use a quasi-steady-state model for a Flexible AC Power System (FACTS) controller that injects reactive power dynamically. We compute the PV curves for multiple buses along the transfer path and compare them to the PMU data to validate the reduced models.

Finally, we present a networked, adaptive controller for damping interarea oscillations. Previous work has demonstrated the benefits of using remote signals for damping control, and one approach is to synthesize these remote signals using high-quality local measurements. The advent of high-fidelity synchronized phasor data allows us to directly measure these remote signals and transmit them over communication links to the damping controller. However, communication delays and PMU phasor computation time introduce a time-varying latency in the input signal that requires additional considerations in the controller design.

Input signal delay introduces phase lag to the system, so we supplement the classical control architecture with a set of phase lead compensators to account for time-varying communication latency. As latency varies with the level of network congestion, we adaptively select among a set of 5 phase lead compensators to account for delays ranging from 0 to 250 ms. Switching instability is prevented by ensuring that the adaptive switching is slower than an average dwell time condition. We adapt a Lyapunov function-based stability proof from switching control theory to prove the stability of our adaptive controller under this condition. Good performance of the adaptive controller is demonstrated in a simulation of a two-area, four-machine system using remote phasor measurements with variable latency.