Model-Predictive Control for Alleviating Transmission Overloads and Voltage Collapse in Large-Scale Electric Power Systems

Emergency control in electric power systems requires rapid identification and implementation of corrective actions. Typically, system operators have performed this service while relying on rules-of-thumb and predetermined control sequences with limited decision support tools. Automatic control schemes offer the potential to improve this process by quickly analyzing large, complex problems to identify the most effective actions. Model-predictive control (MPC) is one such scheme which has a strong record of success in the process industry and has begun receiving attention in power systems applications. Incorporating flexibility into the MPC model using energy storage and temperature-based transmission line limits has shown promising results for relieving transmission overloads on small networks with linear active power models. Separately, MPC has demonstrated its capabilities in correcting transformer-driven voltage collapse behaviors. However, a comprehensive solution combines both aspects into a single controller formulation with knowledge of active and reactive power and voltage magnitude and angle. Additionally, most power system networks are large and result in computationally challenging problem formulations. This work considers these practical limitations and suggests techniques to enable an MPC process capable of operating reliably in the real-world. A new linear controller model is proposed which considers voltage magnitude and angle and both active and reactive power. The new model provides greater accuracy when predicting system behavior and better identifies the actual control needs of the system. The problem size is reduced by limiting the model to only those devices which are significantly affected by the emergency conditions. The new approach is shown to identify controls more rapidly and better suppresses undesirable thermal behavior on overloaded transmission lines while avoiding potential voltage collapse situations.PhDElectrical Engineering: SystemsUniversity of Michigan, Horace H. Rackham School of Graduate Studieshttps://deepblue.lib.umich.edu/bitstream/2027.42/137120/1/jandrewm_1.pd