Research Areas

Matrix and Tensor Computation
In power system planning and operation, it is important to solve an optimization problem in an efficient way. Due to the nonlinearity and nonconvexity, we often find a numerical solution iteratively by solving Ax = b for x. Factorization is widely used for doing so, which makes it possible to utilize the sparsity in the power system optimization. We focus on matrix update - LU, QR, and EVD, and plan to work on tensor computation.

Algorithm for Solving AC Optimal Power Flow
AC OPF finds an optimal solution that satisfies all the economic, policy, operational, and engineering constraints imposed on the power systems. This problem is the building block to formulate the operation and the integration of smart grid and renewable technologies. While important, the problem is a large-scale and highly non-linear optimization that is difficult to solve, and the solution is likely a local solution if found. The GDOP seeks an efficient way to find a Newton-Raphson step with a low rank update.

Grid Integration of Renewable Energy Technologies
In the effort for reducing greenhouse gas emission and for energy independence, renewables have attracted attention over a decade or so time. However, mainly due to the geographically limited energy resources, intermittent nature, and insufficient support from the grid systems, their integration is behind the target values such as renewable portfolio standard. The GDOP addresses the challenges to provide a solution to meet the goal in an optimal way through a systematic approach. We introduce the stochastic nature of renewables and construction decision in the AC OPF framework.

Integration of Smart Grid Technologies
Smart grid is an umbrella term to modernize our grid systems – transmission and distribution networks. It is hoped that smart grid brings a technically feasible and socially agreeable solution to reduce pollution and to improve energy efficiency. The GDOP works on demand-side management, network switching, and storage devices for achieving those goals. Due to the multidisciplinary nature of the research, the GDOP simplifies the nonlinear feature of power flow to accommodate increased computational complexity. Our research activity includes problem formulation, algorithm development and implementation, and parallel computation.

System Monitoring and Control
Recently, around the world, significant efforts are made to monitor and control large-scale power systems. Thanks to the development of phase measurement units (PMU), it is possible to monitor the systems precisely in real-time (60 measurements in a second). India and China have aggressive plans to install PMUs at every substation. DOE also plans to invest on multi-billions to install PMUs. However, the tools to utilize the large-scale data set are not yet developed. The GDOP seeks a system model that allows us to monitor and control to capture events in real-time.

Power System Operation
Power system operators are facing new challenges because of the increased uncertainty due to the integration of renewable generation and smart grid technologies. While difficult, it is important to maintain the system reliability in an acceptable level. The GDOP incorporates these challenges in security-constrained unit-commitment (SCUC) and security-constrained economic dispatch (SCED) problems. In the formulations, we frequently update new forecast such as the load and the resource forecast as available. The GDOP also implements the environmental and social impact of power system operation.

Sensitivity Analysis and Market Power
Market power is a serious issue because it can reduce market efficiency significantly. Conventional indices may not capture the existence of market power in real-time. The GDOP observes that market power is closely related to the location with respect to limiting constraints by series of sensitivity analyses. The situation becomes more complicated than ever when combined with intermittent energy resources and “wind-first” policy. We provide technical solutions to stabilize the electricity market even in the presence of uncertainty.

Visual Computation
Visual computation provides a tool for a decision maker to check how their control action affect power systems in a visual way. Without sufficient backgrounds on the power systems, the decision maker can determine a proper control action to achieve an objective that the decision maker may have in mind.

Test of Ideas, Tools, and Policies on the Electricity Markets
There are many “disasters” that was originally considered brilliant idea and implemented without any testing such as “Pay-as-Bid” markets in California during energy crisis. Simple tests would prevent such disasters because multiple experiments performed by different groups have revealed the potential problems. The GDOP combines classroom teaching to real world experience in the form of virtual auction. Students participate in electricity trading, and test various ideas, tools, and policy. This is “win-win” situation: while students gain real world knowledge, researchers test their idea. Currently, we are seeking support from industry: for tailoring their problems for test; for providing insights of the problems; and for supporting the experiments to incentivize students financially.

The GDP focuses on demand-side management, storage device operation, electric vehicle charging schedule.