Operation and Management of Distributed Energy Resources with Artificial Intelligence
The high proliferation of DERs creates opportunities and challenges to power system operations and management. Energy management and control of DERs are complex due to the availability of energy resources in high numbers with low capacities. DERs are deployed in power distribution systems as grid-connected DERs and standalone power systems, commonly termed distributed energy systems (DESs). This dissertation emphasizes the application of data-driven computational intelligence methods to counter the challenges posed by DES integration.
The first main contribution of this dissertation is a methodology to quantify demand flexibility in temperature-controlled loads. Flexibility quantification is achieved through indoor temperature predictions.
The second contribution is the distributed demand response (DR) management framework developed based on demand flexibility, considering fairness, accessibility, and reliability of DR. Distributed control of the demand is achieved by optimizing DR and distributing DR requests based on demand flexibility. Typical results show the advantage of demand flexibility quantification and DR dispatch optimization.
The third main contribution is the functional situational awareness (FSA) framework for critical energy systems, defining criticality and degradation for components, subsystems and the system. The FSA framework is applied to a hybrid electric vehicle power system (VPS). It utilizes measurements to derive the functionality of components. Subsystem and VPS FSA are inferred as a fusion of component and subsystem FSA, respectively.
Finally, a study is conducted to use FSA inferences in a dynamic energy management system (DEMS-FSA) to improve the reliability and longevity of the VPS. Energy dispatches are adapted to system modes: normal, critical and degraded. Simulations are conducted for an HEV consisting of five subsystems, and the results suggest the efficacy of FSA inferences and DEMS-FSA.