J.W. Chapman

Stabilizing a multimachine power system via decentralized feedback linearizing excitation control

A new controller for the generator excitation system is described that uses a combination of feedback linearizing and the observation decoupled state space. This creates a controller that can be realistically implemented using only local measurements, and whose performance is consistent with respect to changes in network configuration, loading and power transfer conditions. The control differs in this respect from linear constant-gain controllers such as power system stabilizers, whose characteristics can vary significantly with changes in operating conditions. The design is well-suited to a multimachine setting, in that it is not based on an infinite-bus approximation. Simulations were performed on a 38-bus reduced modelof the Northeast Power Coordinating Council system and benchmarked against simulations in which automatic voltage regulators with power system stabilizers were substituted in place of the nonlinear controls. >

Feedback linearizing excitation control on a full-scale power system model

A type of feedback linearizing excitation control is simulated on a very large (5000-bus) model of the Northeast Power Coordinating Council System. The control is implemented on two groups of machines, one in the New York City/Long Island area and one in the Lake Ontario area of New York State. A series of system faults is simulated, the most serious involving the loss of a very large portion of the transmission capacity from Northwest New York State to the New York City area, under extreme loading conditions. The nonlinear control prevents a breakup of the system which would otherwise result in a widespread blackout, at the expense of a rise in the system voltage at the Lake Ontario end. Simulations are also run with PSS/static exciter combinations on the same groups of generators, as a rough comparison of the performance of conventional vs. feedback linearizing control. >

Interactive object-oriented simulation of interconnected power systems using SIMULINK

An object-oriented power system simulation environment is constructed using the SIMULINK dynamic system modeling software. The environment is well suited to educational purposes, because the user interface is interactive and intuitive with a graphical, object-oriented model representation. For small system studies, a model is constructed in block diagram form with one block for each system component. For large scale simulations, the dynamics of portions of the network can be combined into collective blocks, with parameters managed as data arrays accessed indirectly using string mnemonics. The advanced numerical capabilities built into SIMULINK provide an excellent simulation engine for the nonlinear models. Offline analysis is available through the extensive capabilities of the MATLAB environment.

Stochastic modeling of electric power prices in a multi-market environment

Over the past few years, a number of competitive electric power markets have emerged in the United States. While market structure differs by region, the common denominator has been the high level of price volatility experienced in these markets. As power suppliers, marketers and consumers seek to manage their positions in this volatile environment, understanding the locational spreads in power prices is becoming increasingly important. The paper develops a dynamic model describing the interplay of electricity prices in a multi-market environment. In contrast to most existing price models, it is based on the fundamental interaction of demand and supply processes. We illustrate how delays in the information flow to market participants causes price differentials to occur between markets with unconstrained transmission interfaces. Furthermore we examine how correlation between load processes translates into a correlation between the price processes in a dual-market environment. These results are illustrated in a simulated example. In the context of the new model we examine current state of the art algorithms for valuing locational spread options. The work presented suggests that a single correlation factor may not be sufficient to describe the interplay of prices in a dual-market environment. The model illustrates how the price correlation shifts between two states based on the state of congestion on the system. Possibilities for extensions of current option valuation schemes are discussed.

Some robustness results for feedback linearizing control of generator excitation

Results are presented on the robustness of a feedback linearizing controller (FBLC) for power system generators with respect to uncertainty in parameters that reflect the configuration of the power network. It is shown that the control is locally asymptotically stable over a large range of parameter error. Simulations are presented to illustrate the insensitivity of the control to parameter errors under fault conditions.<<ETX>>

Effects of torsional dynamics on nonlinear generator control

The performance of a feedback-linearizing control for excitation control of a synchronous generator is investigated with respect to unmodeled dynamics of both the turbine generator unit and the transmission network. It is found that certain types of dynamics that were not modeled during the design of the control enter in a manner that does affect the performance of the control, but that preserves the linearity of the closed-loop system. Moreover, the control acts to decouple the dynamics associated with the machine from the dynamics of the transmission grid, thus preventing subsynchronous resonance between the two subsystems when a series capacitor is used to compensate the transmission line. The stability robustness of the feedback-linearizing control is investigated with respect to a structured uncertainty. The uncertainty considered corresponds to the spring modes of the generator shaft and enters in such a way that analysis by Kharitonovs theorem is feasible. It is shown that the control remains stable over a wide range of values of the shaft parameters. A sliding control is designed and compared to the feedback-linearizing control with respect to performance degradation for this type of uncertainty, and it is found that, because of the tight saturation limits on the control signal, the sliding control offers no discernable performance advantage for this type of structured uncertainty.

Control of the inter-area dynamics using FACTS technologies in large electric power systems

In this paper a conceptual formulation of a modular control design for controlling inter-area oscillations in the very large-scale power systems is introduced. The design is based on using the new state-space formulation of power system dynamics recently introduced. This choice of state-space enables one to represent even the nonlinear dynamics in the standard state-space form of ordinary differential equations. Furthermore, it sets the basis for control design in terms of measurements simple to perform, and therefore promises considerable robustness. Next, the physical devices for controlling the inter-area oscillations are assumed to be capable of directly controlling flows of transmission lines on which they are located. Some examples of such devices are phase-shifting transformers and series capacitors. If these devices are controlled by means of fast power electronic switching they can be categorized under the well recognized class of the flexible AC transmission systems (FACTS) devices. This paper is concerned with the conceptual systems issues of the best control design so that certain desired performance of the inter-area dynamics is guaranteed.<<ETX>>

Eigenvalue analysis of the stabilizing effects of feedback linearizing control on subsynchronous resonance

Subsynchronous resonance can occur when a series capacitor is used to compensate for the inductance of a high-voltage power transmission line. If the natural frequency of the network is close to the natural modes of the generator shaft subsystem, a resonance can occur which is potentially damaging. A feedback linearizing excitation control for a synchronous generator is evaluated with respect to its effects on an unstable subsynchronous mode. Although the shaft and transmission dynamics are unmodeled for the purpose of the control design, it has been shown via simulation to stabilize and provide significant damping for the resonant mode. Eigenvalue analysis is performed in this paper to corroborate this finding.

Effects of torsional dynamics on nonlinear generator control. II

The performance of a feedback-linearizing control for excitation control of a synchronous generator is investigated with respect to unmodeled dynamics of both the turbine-generator unit and the transmission network. It is found that certain types of dynamics that were not modeled during the design of the control enter in a manner that does affect the performance of the control, but that preserves the linearity of the closed-loop system. Moreover, the control acts to decouple the dynamics associated with the machine from the dynamics of the transmission grid, thus preventing subsynchronous resonance between the two subsystems when a series capacitor is used to compensate the transmission line.