Microgrid Predictive Control With Trajectory Linearization

The patent merges multiple localized PID controllers into a centralized model predictive control system that coordinates control actions across the entire microgrid. The centralized MPC controller integrates information from all generators and loads to compute optimal control inputs that maintain stability during rapid disturbances, resolving the contradiction between localized simplicity and system-wide stability.

The patent implements dynamic control by using a linearized dynamic model of the microgrid that adapts to changing operating conditions. The MPC controller continuously updates control actions based on predicted future states, enabling the system to respond dynamically to rapid disturbances rather than relying on fixed PID parameters that deteriorate under varying conditions.

The patent changes the control parameters from fixed PID gains to dynamic control inputs computed by MPC. The linearized dynamic model provides parameter variations that adapt to different operating conditions and disturbance types, achieving broad-range accuracy without manual retuning. This eliminates the need for complex engineering efforts to tune PID controllers for each specific disturbance scenario.

The patent applies preliminary action by using the linearized dynamic model to predict future system states and compute optimal control inputs before disturbances fully develop. The MPC controller proactively adjusts control actions to prevent instability rather than reactively responding after disturbances occur, achieving high accuracy across various disturbance types without extensive trial-and-error tuning.

The patent replaces complex mechanical tuning and coordination of multiple PID controllers with a computational model-based approach. The linearized dynamic model mathematically represents system behavior, and the MPC controller uses optimization algorithms to compute control actions, substituting mechanical complexity with computational efficiency. This achieves rapid coordinated response across the microgrid without the complexity of tuning and synchronizing multiple independent controllers.

The patent segments the control problem by linearizing the nonlinear microgrid dynamics into manageable linear models around different operating points. This segmentation allows the use of simpler linear control techniques while maintaining accuracy through piecewise linearization, reducing the overall control complexity compared to directly controlling the full nonlinear system.