Research On Centralized-Distributed Coordinated Control Of Active Distribution Networks Under Multivariate Coupling

The rapid development of renewable energy such as distributed photovoltaics has brought serious uncertainty and volatility to the secure,reliable and efficient operation of active distribution networks.To assure the operational security and economic benefits of the active distribution network,a variety of flexible resources including energy storages and controllable loads are increasingly accessed.Then the multivariate coupling characteristics among controllable devices are more prominent,including multi-period operation coupling,active/reactive power and voltage cooperative control coupling,multi-energy complementary coupling,etc.,which introduces new challenges to the coordinated operation of active distribution networks.With the continuous advancement of communication technology,the distributed method outperforms the centralized and the local schemes on computational efficiency and privacy protection,and can realize the rapid coordinated optimization and control of the distribution network with the participation of a large number of controllable devices and users.However,under the complex coupling of multi-space-time and multi-energy,the traditional distributed control methods still face some problems,such as complicated iterative communication,failure to fully employ the response potential of flexible resources,and limited applicability to multiple controllable devices.In addition,considering the obvious differences in equipment and network size as well as control requirements in distribution networks of different voltage levels,it is still challenging to promote the optimality and rapidity of wide-area flexible resource regulation by using a single distributed method.Therefore,considering the multivariate coupling,this dissertation conducts research started from distributed control of multiple controllable devices and multi-energy networks.Then the centralized-distributed coordinated control strategy of active distribution networks is proposed to unite the advantages of both centralized and distributed control methods.The overall goal of the dissertation is to improve both the operational safety and efficiency of the active distribution network,and the main contributions are summarized below:(1)This dissertation proposes a back-and-forth communication based distributed voltage regulation method of active distribution networks concerning multi-energy storages.To improve the operational security and economy of the network,the multi-energy storage integrates battery energy storage,thermal energy storage,and thermostatically controlled load considering the thermal inertia of buildings.Based on the decomposition of the optimal conditions of the centralized voltage regulation formulation,a distributed back-and-forth communication framework is developed,and the sensitivity that can reflect the ability of each multi-energy storage to alleviate voltage violations of the whole network can be obtained through only one iteration.Combined with the iterative local gradient descent method,a multi-period independent optimization for each agent is constructed,which ensures the global optimality of the network voltage regulation.The proposed distributed voltage regulation method effectively handles the optimization problem with coupling constraints of power network,temporal correlation and multi-energy interaction,which fully utilizes the multi-period voltage regulation capability of multi-energy storages with high computing efficiency to guarantee the network voltage security.(2)This dissertation proposes a distributed voltage control method of active distribution networks with global sensitivity.To realize distributed control with high precision,a modified Dist Flow power flow model is constructed to simplify the coupling relationship of voltage and power among nodes while ensuring the accuracy of power flow calculation.The global sensitivities of all node voltage deviation penalties to each nodal active power,reactive power and voltage are newly established,and the optimal voltage control target of the distribution network is decomposed into second-level autonomous control of diversified controllable devices responding to their own global sensitivities change.Based on the modified Dist Flow model,an improved distributed back-and-forth communication framework is developed to facilitate fast and accurate calculation of all global sensitivities with high communication and computational efficiency.The proposed distributed voltage control method avoids a large number of communication iterations and optimization calculations that may be involved in traditional distributed schemes,and significantly enhances the applicability for multiple controllable resources to participate in distributed coordinated voltage control in real time.(3)This dissertation proposes a fully distributed optimization method for the coordination between the distribution network and the district heating network considering multi-energy coupling.Energy agents are formulated in the multi-energy network.The networks connecting agents fully model the nonlinearity of power flow and the heat transfer delay.By introducing auxiliary observation variables between adjacent agents,the complex energy-flow coupling constraints between agents are replaced by linear consensus constraints.Combined with the alternating direction method of multiplier and the correlation analysis among variables,a distributed algorithm for the coordination of multi-energy networks is proposed,which establishes sub-problems that can be solved by each agent in parallel iteratively.Further,the closed-form solutions for all sub-problems are deduced to extremely accelerate the computation speed of distributed optimization.The proposed distributed optimization method not only respects energy agents’ privacy,but also realizes the optimal coordination of flexible resources in the multi-energy network with a much-reduced computational burden in each agent.(4)This dissertation proposes a multi-time-scale centralized-distributed coordinated voltage control method in multi-level active distribution networks.To ensure voltage control performance in terms of optimality and instantaneity simultaneously,centralized and distributed control methods are adopted for medium-voltage and low-voltage distribution networks,respectively.An integrated distributed optimization method is developed for constructing the communication patterns across multi-voltage levels as well as inside low-voltage networks,which promotes the optimal multi-period economic operation of multi-level networks with high convergence.To avoid voltage issues caused by second-level uncertainties in real time,the distributed voltage control with global sensitivity is applied in the low-voltage network.Meanwhile,an efficiently solved voltage control optimization model is formulated in the medium-voltage network,which ameliorates voltages of the entire network while avoiding privacy disclosure of the low-voltage users.The proposed voltage control method realizes the efficient coordination of centralized and distributed strategies on multiple time scales,and improves the control efficiency of various distributed resources in cross-level and cross-regional distribution networks.