Research Objectives and Approaches: The objective of this research is to develop a fundamental understanding of the properties of electro-mechanical waves that propagate as frequency and voltage disturbances over the electrical grid. These waves reveal much about the short-term dynamics and stability of the electrical grid and will become increasingly important in the future as time-intermittent renewable generation resources are incorporated into the grid. The approach is to leverage the new, wide-area, high-resolution phasor measurement unit monitoring and apply methods and approaches from the fields of nonlinear systems, applied mathematics, statistical physics, and signal analysis to study the dynamics of voltage, power flow, and frequency disturbances on large-scale transmission grids.
Intellectual Merit: Studying electro-mechanical wave phenomena in large power systems is a promising new approach that draws on numerous disciplines and will rapidly advance the frontiers of the physics-based understanding and data-driven inference of electrical grid states and fast dynamics. This new understanding will enable the design of smart grid controls to prevent economically-costly cascading grid failures.
Broader Impacts: New time-intermittent renewable resources create a need for effective ways of discovering how an event in one location is manifested in the rest of the grid. Electro-mechanical waves provide one such relationship and will be key to early detection of problematic grid behavior and design and activation of advanced control systems to restore reliable operation. Our project will develop a better fundamental understanding of electrical grid behavior, which will ultimately translate into improved operation, better control, and lowered cost.
The research part of the project focused on the development of novel mathematical tools for analysis and control of future power systems. New technologies coming to power systems, such as Phasor Measurement Units, Renewable wind and solar generation, large scale stand-alone storage will require a major revamp of decision and control layers employed for planning and operation of the power grid. Within the scope of the project team of PI Turitsyn has addressed a number of practical questions related to operation of future power systems with some of the most advanced mathematical tools originating from the fields of theoretical physics and applied mathematics. All research projects were carried out with graduate and undergraduate students and postdoctoral scholars who were provided an opportunity to prepare themselves for a career in and exciting research field of extreme importance to national interests. Three main areas where significant progress has been made are summarized below: Loss of system stability accompanied most of the major blackouts that happened in the world in the last decades. Fast and intelligent response of system operators in critical situation is the only way to avoid propagation of cascades in dangerous situations typically initiated by loss of some system components. One of the challenges that operators face nowadays is the difficulty of tracking the state of the system and estimating the risks of system collapse in real-time. Novel Phasor Measurement Unit (PMU) technology provides an opportunity to address this challenge by real-time sensing of voltage and currents all over the grid. However, tracking of the modes in the grid is complicated by ambient stochastic fluctuations that are naturally observed due to variations in system loads. These fluctuations get amplified as the system approaches the instability threshold. We have developed a rigorous mathematical theory describing the amplification of fluctuations close to stability margin and used this theory to develop the security indicators reflecting the risks of system collapse as well as practical algorithms for identification of the most dangerous direction in the phase space that could be applied in real-time operations. Our algorithms were validated by simulations of realistic dynamic models of power grids and reported in several prestigious power system conferences. Introduction of renewable generation to power grid will help US to reduce its dependence on foreign fossil fuels and alleviate the damage to climate by CO2 emissions. At the same time it will also make secure operations of the power grid a much more difficult task. Wind and solar resources experience random uncontrollable and unpredictable fluctuations of power output that need to be somehow balanced by available controllable resources. Stand-alone electricity storage is one possible way of solving the intermittency problem. However, a variety of technological limitations of existing storage technologies makes it difficult to assess their effectiveness in practical deployments. We have developed a mathematical framework based on modern convex optimization and stochastic analysis for assessment of the storage technologies and the overall costs of reducing the amount of fluctuations. The framework allows to naturally compare different storage resources find optimal sizing of storage systems, and optimal control strategies. Furthermore, we have explored the "value of forecast" question by comparing the effectiveness of mitigation strategies in the presence of imperfect forecast of wind and solar energy outputs. Within the scope of this project we have provided a summer research opportunity for minority undergraduate student who decided to pursue his career in academia. Most of the modern power systems are designed to operate under the N-1 security constraint which requires that failure of any single component should not result in violation of any technological constraints in the system. This requirement implies that the actual risks in the system are associated with N-2 contingencies corresponding to simultaneous failure of two components at once. Identification of these scenarios is extremely challenging computation wise and most of the operators rely on poorly justified heuristics or engineering judgement in their risk assessment and contingency planning procedures. We have addressed this problem by deriving a simple but rigorous algorithm for screening of N-2 contingencies that could be naturally implemented in real time even for large-scale grids. On medium size models the algorithm was operating several thousand times faster in comparison to more traditional alternatives. In a followup work this algorithm was used for assessment of common contingencies and identification of key vulnerabilities of modern power systems. The development of algorithm attracted significant attention of mass medium and increased awareness of the important challenges that the community faces these days.
