Constructing sustainable power grids : environment-friendly and resilient operation

Abstract

With increased pressures and requirements introduced by the environment, economy and society, it becomes vitally important to construct sustainable power grids. While many issues have to be resolved to attain power grid sustainability, this thesis focuses on three critical topics, i.e. integration of renewable energy sources (RESs), emission control, and power grid resilience, so as to develop environment-friendly and resilient operation strategies. To improve transmission-level RES integration, a robust optimization based methodology is established to accommodate uncertainties of RESs. Robust economic dispatch (ED) and unit commitment (UC) models are proposed, which are solved by algorithms based on Benders decomposition and column-and-constraint generation (C&CG), respectively. The models can generate ED and UC strategies with sufficient operational flexibilities for any scenario considered in uncertainty sets. Inadequate or excessive spinning reserve, which can undermine RESs’ effects in reducing emissions, is avoided. Thus power grid environmental sustainability is also enhanced. To improve distributed RES integration in distribution systems (DSs), dynamic network reconfiguration is studied. As distribution system dynamic reconfiguration (DSDR) relies on real-time operations of remote-controlled switches (RCSs), this thesis addresses a research gap to identify critical switches that optimally enable DSDR to assist distributed RES integration. Considering the uncertainties of loads and renewable distributed generations (DGs), a robust critical switch identification model solved by the nested C&CG algorithm is constructed. Results show that, by a limited number of switching actions of only several critical switches, DSDR can significantly reduce DG curtailment. As for emission control, this thesis explores a novel scheme considering air pollutant dispersion. Using the Gaussian plume model, each load center’s ground level air pollutant concentration (GLAPC) resulting from the generations’ emissions is estimated. The GLAPC estimation is added into the proposed robust ED and UC models as a constraint and cost, respectively. A production costing model with GLAPC limits is also formulated to select the coal type of each coal unit and decide purchase quantities of coals and gas, which determine partial operation conditions for ED and UC. Case studies verify this scheme’s effectiveness in air pollution control of power grid operations. This thesis also investigates the problems of DS automation and mobile generation resource dispatch to enhance power grid resilience. In automating DSs, RCSs are allocated to enable prompt restoration of DSs. RCS allocation models for three different reliability objectives, i.e. minimizing customer interruption cost, minimizing system average interruption duration index, and maximizing the amount of loads that can be restored by the allocated RCSs, are proposed. Illustrative cases show that a small number of RCSs can substantially enhance restoration capacity of DSs. As for mobile generation resource dispatch, mobile emergency generators (MEGs) are concentrated on. A two-stage framework consisting of pre-positioning and real-time allocation is constructed to dispatch MEGs as DGs into DSs to restore critical loads by forming multiple microgrids. A scenario decomposition algorithm is developed to solve the stochastic mixed-integer linear programming pre-positioning problem. Case studies indicate that the proposed dispatch method provides resilient response strategies of MEGs to greatly reduce both outage scale and duration.