Haes Alhelou, Hassan

Frequency Control in Power Systems Considering Unknown Inputs and Measurements Fault Using Dynamic State Estimation

AbstractFrequency control is among the most important tasks in modern powersystems due to the increase of their active power imbalances between gen eration and demand uncertainties and complexities Such new energysystems need new control approaches that can handle practical challengessuch as unknown inputs failures in measurement systems high uncer tainties and cyber attacks This dissertation proposes a novel fully de centralized load frequency control which is unknown as secondary con trol approach based on dynamic state estimation DSE The proposedapproach employs an unknown input observer UIO for each power sys tem area for tracking the dynamic states in real time operation Inorder to achieve the fully decentralized control approach for secondaryfrequency control loop preserving the hierarchical concept the demand uctuation and tie line power deviations are modeled as unknown in puts to the UIO of each area An optimal state feedback control methodbased on the observed states is used next to control the frequency of eacharea separately This approach reduces the complexity of the designedobserver achieves fully decentralized control improves the robustnessof the controllers and makes the estimation process highly e cient ac curate and easy to implement Based on the suggested UIO a novelunknown input functional observer based optimal load frequency controlapproach is also proposed for real world complex nonlinear power sys tems In the proposed control approach the control signal applied toeach power plant is directly estimated via the well designed functionalobserver The proposed functional dynamic estimator is designed to han dle parametric and nonparametric uncertainties control loop and sensorfaults unknown inputs and cyber attacks The functional estimator foreach power plant is decoupled from the other plants resulting in a morefeasible implementation reducing the complexity of the estimator andimproving the reliability of the proposed control system Furthermore the proposed UIO is further developed mathematically to have the abil ity of online detection and isolation of practical issues associated withmeasurement and communication parts such as faulty sensors error inmeasurement and cyber attacks This contribution helps power systemoperator to eliminate the impacts of the mentioned practical and techni cal challenges on the performance of the control systems which increasesthe overall reliability and security of frequency control in modern powersystems Moreover the proposed dynamic observer is further developedto online estimate the magnitude and locations of unknown inputs Pro viding such knowledge of unknown inputs e g disturbances to operatorcan highly help in improving the control and protection performance in