Operational methods in smart grid with intelligent periphery under uncertainties and severe events

Abstract

To achieve a sustainable and green energy-oriented power energy system, the Energy Internet was described to achieve this goal, in which the grids are envisioned as plug-in, energy shared and distributed controlled. The smart grid intelligence periphery (GRIP) was proposed as a future architecture for the Energy Internet with three basic functionalities, which are system operation with large scale renewables integration, frequency fluctuation alleviation and system failure mitigation. This thesis focuses on the first and third functionalities. In other words, the operational methods addressing uncertainties of renewables and system failures especially from severe events are developed. About the operational method for renewables integration in the GRIP, the risk-limiting dispatch is chosen as the operational method in the GRIP. In order to conquer the limitations in the basic risk-limiting dispatch model, four fundamental requirements for the risk-limiting dispatch are summarized from practical reports and academic research. The four requirements are the transmission network constraint, the dispatch framework composed of multiple delivery periods, alternative risk index which is computational tractable, and the necessity of integrating unit commitment in the context of risk-limiting dispatch. To satisfy such requirements, this thesis extends the basic risk-limiting dispatch to multi-period risk-limiting dispatch and risk-limiting unit commitment. They are in consistency with the current electricity markets, and can mitigate the operating risks through sequentially dispatching units, assisted by updated prediction information for uncertainties of renewables generations. About the operational method for system failure mitigation, this thesis puts emphasis on the system failure mitigation method against severe events, which is called system resilience enhancement. The operational method for system resilience enhancement including three stages which are prior-to-event stage, during-events stage and after-event stage. In the first stage, this thesis designs a resilience assessment criterion to evaluate the system resilience before severe event come, by means of measuring the risk of cascading failure during severe events. A self-organized criticality theory based assessment criterion is proposed so that such assessment can be effectively computed in large complex systems. In the second stage, a model predictive control based sequential operation is proposed to schedule the system resources to enable the system adaptability against unfolding events. Taking advantages of model predictive control, the sequential operation mitigates the damage of severe events by using the updated prediction information of severe events. In the third stage, the proposed optimal breaker sequence operation realizes the restoration strategies after severe events in the node-breaker model. In sum, this thesis contributes to develop the operational methods to accomplish two functionalities in the GRIP, where risk-limiting dispatch aims to tackle the uncertainties due to large scale renewables integration, and system resilience enhancement method spears at mitigate the system failures due to severe events. The achievements of this thesis carry a major step forward on the realistic implementation of the GRIP.