A Semi-Analytical, Heterogeneous Multiscale Method for Simulation of Inverter-Dense Power Grids

In the pursuit of low-carbon, clean electricity generation, renewable resources such as wind and solar are gaining prominence in a power grid that traditionally relies on fossil fuels to generate electricity. However, the integration of renewables through power electronic inverters introduces uncertainties to grid planning and operations due to their intermittent and variable nature. While battery-based energy storage systems can mitigate the impacts from renewables, they are also integrated through power electronic inverters. All these factors are shaping today’s power grid into an inverter-dense grid, posing a considerable challenge in the computer simulation of its tremendously complex, multi-timescale dynamics, ranging from microseconds to hours. This project aims to establish a unified multi-timescale modeling and solution framework addressing this challenge, and will bring transformative changes to the field of power system simulation. The intellectual merits of the project include assimilating heterogeneous multiscale methods and semi-analytical solution methods for highly accurate and efficient simulations. The broader impacts of the project include an in-depth understanding of the dynamics of inverter-dense grids, the establishment of a high-performance multi-timescale simulator paradigm, the fostering of interdisciplinary education and the facilitation of power grid decarbonization. Power system simulation is crucial for assessing grid stability and reliability under disturbances. The multi-timescale behaviors of an inverter-dense grid may involve electromagnetic transient, electromechanical and quasi-steady-state dynamics. Accurate simulation of such a system requires solving numerous nonlinear and stiff differential and algebraic equations at a small time step over an extended period. The heterogeneous multiscale methods developed by the project will allow case-specific model reduction and solver parameterization to enable automatic zoom-in and -out of the simulation resolution between macroscopic and microscopic timescales on the fly of simulation. Variable-step and -order mechanisms will be employed in each timescale through semi-analytical solvers. The validation studies will encompass sub-synchronous oscillation, transient stability, and black-start scenarios in realistic grid models with a penetration of inverter-based resources ranging from 50% to 100%. The results from the project have the potential to advance knowledge in efficient simulation methodologies, which are also applicable to other multi-timescale networked systems. 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.