Collaborative Research: Dynamic Grid Optimization under High Renewable Penetration: Multistage Algorithms and Stability Augmentation

Growth of renewable energy leads to new challenges for electric power grid planning and operation. Many renewable energy resources, such as solar and wind, heavily depend on the weather conditions that are inherently uncertain. Such uncertainty is usually revealed progressively over time. Consequently, the grid planning and operation decisions need be adjusted accordingly across multiple stages to achieve optimal efficiency. The multistage decision structure calls for study on multistage grid optimization algorithms that can accommodate the discrete decisions, such as battery charging versus discharging decisions, and scale well with the number of renewable resources, which can go up to tens of thousands. Moreover, several major tripping and disturbance incidences in the past decade have underscored the heightened stability concerns of a power grid with high renewable penetration. In contrast to conventional thermal generators that have large rotating masses to stabilize themselves, renewable resources are typically power electronics-interfaced resources, which lead to lower system inertia, faster grid dynamics, more frequent disturbances, and greater control difficulty. Hence, it is increasingly essential to integrate stability considerations into grid optimization algorithms to enhance reliable power system operation. This research will include open-source implementations of the algorithms developed, which can provide a computational infrastructure and benchmark for assessing long-term energy integration plans, or for evaluating the daily operational efficiency and reliability of power grids. To address these critical challenges of uncertainty and stability, this project aims to develop novel dynamic grid optimization algorithms and modeling tools to effectively accommodate high penetration of renewable energy and ensure reliable grid operation. The first part of this project is focused on a class of algorithms, called stochastic dual dynamic programming, for multistage stochastic optimization models. The investigators will fundamentally advance these algorithms to handle both continuous and discrete grid decisions effectively, and to enable better statistical guarantees by exploiting the structure of grid optimization with renewable uncertainty. The second part of this project plans to directly integrate stability considerations into the objective function and constraints of grid optimization, establishing a framework of stability-augmented grid optimization. Such framework enhances conventional grid planning and operation decisions to be stability-informed and optimizes both the economic efficiency and dynamic stability performance. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.