Consensus-Based Distributed Energy Man-Agement And Optimization In Smart Grids

With the rapid advancement in distributed energy resources,demand response,energy storage,electric vehicle,and etc.,the smart grid is expected to be the integration of power flow,information flow,and service flow,which is more secure,efficient,robust,reliable,self-healing,and etc.Smart grids will upgrade the structure of power energy and thus can alleviate the problems including envi-ronment pollution,global warming,and so on.As a result,the smart grid has become a significant and hot research area for both industry and academic communities.Energy management plays a key role in the efficient,stable,and reliable operation of smart grids,which is of paramount importance.Especially,much attention has been paid to consensus-based distributed energy management algorithms in recent years.These algorithms overcome shortcomings of traditional centralized ones,such as high communication and computation cost,low scalability,and so on.However,most of existing works are faced with four challenges.First,the existing optimization models neglect the effect of transmission losses,which is only effec-tive for undirected communication networks.Secondly,how consensus-based distributed energy management algorithms perform under imperfect network conditions,for instance,communication delays,is still lack of enough theoretical analysis.Thirdly,cyberattacks modeling and the corre-sponding resilient schemes are still opening problems.Lastly,whether the privacy of demands will be disclosed and how to preserve the privacy are not investigated.In order to solve these problems,the distributed energy management with transmission losses and directed communica-tion is taken into consideration.The effect of communication delays and stealthy attacks on the consensus-based algorithm and study the related privacy concern is also anayzed.The main works and contributions are summarized as follows.1.A brief literature review on the development of smart grids,energy management,and robust,secure and privacy-preserving energy management issues,is provided.2.For the distributed energy management problem with ideal communication networks,the effect of transmission losses on the balance of generation and demand is modeled.Then a nonconvex optimization problem with the objective of maximizing social welfare is formu-lated.To solve the problem,the nonconvex optimization problem is first transformed into a convex one and the sufficient condition is then obtained to guarantee that both problems have the same optimal solution.For directed communication,a consensus-based distributed algorithm is proposed to solve the problem and its convergence and optimality of the final solution is then analyzed theoretically.3.For the distributed economic dispatch problem with communication delays,how constant delays can affect the performance of a typical consensus-based algorithm is analyzed.For the cases with unbounded generation power,the original system under delays is first augmented and the sufficient conditions are provided to guarantee its stability.Then,the upper bound of the learning gain parameter is acquired by utilizing the eigenvalues theory of matrix.The convergence analysis is generalized to bounded generation cases.To guarantee the optimality of the final solution,an iteration law for iterations when not all neighbors' information is received is proposed and the optimality of the final solution is proved.4.For the distributed economic dispatch problem under stealthy attacks,stealthy attacks is modeled for both online and offline cases.First,it is shown that stealthy attacks for online cases through injecting false data into broadcast information do not exist,while they exist for offline cases.Then,it is found that stealthy attacks by injecting constant false data into the parameters of cost functions can be realized for both online and offline cases.Meanwhile,a resilient consensus protocol and a resilient distributed optimization algorithm are proposed,respectively,where the number of adversaries is not limited by the network connectivity.These results can provide theoretical basis for designing defense strategy for distributed en-ergy management.5.For the privacy concern of consensus-based distributed energy management,it is revealed that the private information including the power consumption and the sensitivity of the pow-er consumption to the electricity price can be disclosed.A secret-function-based privacy-preserving algorithm is then proposed and its convergence and the optimality of the final solution theoretically is analyzed.In order to relax the assumption on the secret function,a privacy preserving algorithm is proposed,where each node utilizes local real information for state update and broadcasts the one with noise.The privacy degree of the proposed algo-rithms is analyzed through(?,?)-data-privacy.In the end,the thesis is concluded and some future research works are discussed.