Coordinated Optimal Operation Of District Integrated Energy Systems Considering Multiple Uncertainties

High penetration of renewable energy imposes significant challenges on power system operation and control.One fundamental issue in the power industry is to facilitate economically efficient and environmentally friendly system operation and resource management against inherent uncertainties from renewable power sources and loads while maintaining system reliability and security.Increasing deployment of distributed energy resources in power distribution systems and microgrids,such as distributed combined heat and power plants,electric heat pumps and air conditioners,dramatically intensifies energy interactions and information exchange among electric power systems and district heating/cooling systems,leading to the formation of district integrated energy systems.District integrated energy systems have been widely recognized for enlarging power system controllability,enhancing the overall energy utilization efficiency at a local level,and contributing to scalable renewable energy integration with cost reduction.However,many problems remain unexplored with respect to theoretical study and industrial practice in the field of integrated energy systems.Due to the nonlinearity in electric power and heating flow models and complex interactions among different energy systems,further research is needed to improve modeling accuracy and scalability of devices and networks.On the other hand,high penetration of renewable energy sources imposes significant uncertainties on secure and efficient operation of integrated energy systems.Traditional separate operation of different energy systems fails to guarantee global cost reduction and renewable energy integration.To this end,this dissertation covers topics including modeling of district integrated energy systems considering nonlinear network flows,flexibility evaluation for integrated energy systems with storage capability in district heating networks,and coordinated operation optimization of district integrated energy systems under multiple uncertainties.The overall goal to the dissertation is to provide essential tools and analytics for reliable and efficient operation of integrated energy systems.(1)This dissertation proposes a comprehensive model for district integrated energy systems considering nonlinear and nonconvex network flows.A linearized power distribution flow model is developed based on simplified Z-bus sensitivities.A novel convex relaxation model of combined heat and power dispatch is proposed using simplified thermal dynamic models and constraint relaxation.An adaptive solution algorithm based on dynamic bivariate partitioning is devised to improve relaxation quality with satisfactory computational performance.The proposed modeling scheme for district integrated energy systems can produce enhanced solutions with good accuracy and desirable computing efficiency,contributing to better modeling scalability for system operation optimization.(2)This dissertation proposes a operational flexibility evaluation method of district heating networks in combined heat and power dispatch based on a generalized thermal storage model.A direct quantification method is developed to obtain flexibility metrics assessing the ability of heating networks in providing balancing service for integrated energy systems.Four different control modes of district heating systems,such as variable mass flow rates and variable supply temperature,are modeled separately to provide a comprehensive view into the operational principles of heating systems.A simplified solution algorithm for combined heat and power dispatch is developed based on sequential linear programming to significantly improve the computational efficiency with desirable modeling accuracy,avoiding extensive simulation in traditional methods.The proposed flexibility evaluation method is verified with comprehensive case studies to demonstrate the effectiveness and scalability in evaluating the operational flexibility of slow-dynamic heating networks.(3)This dissertation proposes a stochastic receding-horizon control method for active power distribution systems considering probabilistic uncertainties of renewable energy and electricity loads.Multiple distributed controllable resources are jointly optimized in a multiperiod online stochastic scheduling framework.The voltage limitations are reformulated as chance constraints to indicate the probabilistic reliability index of voltage qualification rate,and achieve trade-offs between cost reduction and voltage regulation.The affine-disturbance parameterization is utilized to render the stochastic chance-constrained problem computationally solvable via mixed-integer second-order cone programming.The simplified Z-bus sensitivity,combined with sequential linear programming,is developed for computationally efficient estimation of system nonlinearity while guaranteeing modeling accuracy.The proposed method can facilitate reliable and efficient online operation of active power distribution systems,and improve voltage regulation capability,control efficiency and renewable energy integration.(4)This dissertation proposes an efficient robust dispatch method of combined heat and power systems based on extended disturbance invariant sets.The efficient robust dispatch model is developed based on set analysis,computation geometry and mathematical programming to characterize uncertainty sets of renewable power and multi-energy loads and their impacts.The robustness and computational efficiency are enhanced by solving a nominal uncertainty-free problem without introducing auxiliary variables and constraints.A direct constraint tightening algorithm based on the dual norm is developed to efficiently derive multiperiod tightened constraints.The budget uncertainty set is newly combined with constraint tightening to flexibly adjust the level of conservativeness of the robust solutions.The proposed efficient robust dispatch method can provide robust scheduling strategies for combined heating and power systems with significantly improved computational and economic performance.(5)This dissertation proposes a hybrid stochastic-interval operation strategy for multienergy microgrids.A novel formulation of uncertain operation regions is proposed to describe performance and efficiency uncertainties of distributed energy resources using interval-based convex combination.Probability density functions are utilized to capture injection uncertainties from renewable power sources and multi-energy loads.A scenario-based two-stage algorithm is devised to solve the hybrid optimization problem,which can preserve heterogeneous information of interval and probabilistic uncertainties in the entire optimization process.The proposed hybrid stochastic-interval operation scheme for multi-energy microgrids comprehensively characterizes both internal and external uncertainties to generate scheduling strategies presented as a set of intervals with probability attributes,which provide more credible guidelines for decision making.