Modeling, Analysis and Control Design for Spatially Distributed Systems with Application to Wind Farms

Johns Hopkins University, Baltimore MD

Investigators

Abstract

This work will address challenges arising in the modeling and control of wind farms, enabling power grid operators both to accurately predict capacity under varying operating conditions, as well as to track a desired power output profile. In contrast to controllers that simply maximize output power, this enhanced operational flexibility will allow wind farms to support utilities with valuable services such as frequency regulation and power ramping. The key technical innovation of this project is a reduced-order model that accurately captures dynamic interaction between turbines and the effect of turbine wakes, while remaining tractable enough for real-time computation. The model and controls will be tested using high-fidelity wind farm simulations and actual operator data. The tools to be created will help improve the economic competitiveness of wind power as a core component of the electric power system. The project will provide interdisciplinary research training to graduate students, focusing on the interplay between wind farm flow physics and control theory, and including international research opportunities. The project includes outreach to Baltimore high schools and a STEM summer program for high school students at the Johns Hopkins University. This project will create models, control algorithms and simulation tools for multiscale systems in which mechanical devices must perform coordinated actions within an infinite (or high) dimensional process, with specific application to wind farms. The research goals of the project are as follows: (i) computationally tractable methods to capture the complex interactions between turbines, and to bridge the gap between overly simplified static wake models and slow, costly high-dimensional models; (ii) a model-based control framework to move beyond the traditional control goal of maximum power extraction, and facilitate more complex trajectory tracking; (iii) extension of current high-fidelity simulation tools to be suitable for wind farm model validation and control algorithm testing and refinement; (iv) combined wind farm modeling and control tool development based on a grid simulation system that generates frequency control trajectories and acts as a supervisory control entity to distribute the control objectives over multiple wind farms. The project will provide a large eddy simulation (LES) based model and control validation platform that can facilitate modular changes in wind turbine and wake models along with the control algorithms, so that new models and flow control strategies can be efficiently tested. The new platform will support the coordination of multiple wind farms participating in frequency regulation services, using an interconnected simulated power grid as a supervisory system.