Distributed And Coordinated Volt-VAR Control Model And Algorithm For Active Distribution Networks Considering Dispatch Requirements

The extensive penetration of distributed generation(DG)provides new means of control for both the transmission and distribution systems for the safe and economical operation.By regulating the reactive power output of inverters,DGs such as photovoltaic and energy storage can provide continuous voltage support and reactive power compensation for the distribution network.Moreover,by batch regulating of the active power of massive DGs in distribution network,the transmission network can treat the distribution network as a dispatchable unit.However,due to the large number of nodes in the distribution network,it is hard to centrally control many DGs deployed in a dispersed manner.To solve this problem,investigating control strategies suitable for limited communication and measurement conditions is necessary.Besides,traditional voltage reactive control in the distribution network is performed by substations.The coordination between slowly substation control and fast DG control need to be studied after the DGs participate in the regulation.When the distribution network is dispatched as a responsive resource in the transmission network,the total active power regulation quantity must be assigned reasonably among the DGs.Simultaneously,the voltage fluctuation caused by the active power control needs to be alleviated by reactive voltage control in DG itself.The transmission system also needs to be matched with corresponding real-time scheduling strategies to effectively utilize the rapid regulation capability of distributed power supplies in the distribution network.This paper systematically studies the distributed coordinated Volt-Var control model and algorithm for DGs in the distribution network considering the dispatch requirements,consisting of the following points:(1)A distributed reactive power control strategy based on the branch flow is proposed to minimize the total active network loss while ensuring that each node's voltage in the feeder complies with the limit value.This strategy can be used in a distribution feeder without synchronous phase measurement(PMU),while the load buses are all unmeasurable.Still,one DG can communicate with other DGs nearby in real-time.The strategy is based on a linear trend model.It takes advantage of the sparsity of interconnections between nodes to represent the optimal reactive output of each DG under unconstrained conditions as a function of the measurable voltage and power flow of the neighboring DG nodes and to achieve distributed control by processing node voltage and reactive power constraints through a dual decomposition.For unmeasurable load nodes,the voltages are estimated using the voltages and branch flows of the DG nodes at both sides.(2)For the distribution feeder equipped with micro PMU(?-PMU),a decentralized reactive power voltage control strategy for the distribution network is provided,which does not rely on real-time communication between DGs.According to the linear power flow model,the partial derivative of the DGs' reactive output to the system's active power loss can be formulated as a function of the local voltage phase and the root node's voltage angle.On this basis,a local control strategy for reducing active net loss is proposed.Moreover,an analytical solution of the optimal reactive power compensation scheme to provide voltage support for a single node is proposed,significantly improving the speed of fixing voltage problems.(3)An automatic voltage control strategy for substations coordinated with DGs is proposed for distribution networks with extensive penetration of DGs.Based on the neighboring communication network between the DGs,an average consensus protocol is used to estimate the maximum and minimum voltage in the distribution network.The maximum control capacity of the DGs over the highest/lowest voltage nodes of a feeder is predicted.The on-load tap-changing transformer settings are determined based on the predicted voltage constraint violation.The proposed strategy considers both the limitation on the transformer's action frequency and the different medium-voltage feeders' different requirements.It uses a centralized coordination method to balance each feeder's needs and determine the substation voltage level.(4)A decentralized and coordinated real-time dispatch strategy for transmission and distribution networks is proposed,taking into account the need of distribution networks to participate in transmission network dispatch.The strategy adopts an average consensus algorithm to assign the active reduction amount to each DG in distribution systems.The active power injected by the distribution network is adjusted to the value required by the transmission network.Simultaneously,a feed-forward control loop is added to the reactive power control strategy of DGs to reduce the voltage fluctuation caused by rapid active power adjustment.A data-driven centralized framework is adopted for the transmission system to identify each economic and safety indicator's sensitivity of bus power injections using the high-frequency sampling data provided by synchronous phase measurement.A local linearization model of the system is established.Then the power injection of generators and each dispatchable distribution network is optimized in real-time based on it.A dynamic-decision-based real-time dispatch strategy is proposed for a situation where too many security indicators need to be considered,making it challenging to satisfy all the constraints in a single real-time dispatch cycle.The proposed dispatch strategy first eliminates the constraints violations,minimizes the generation cost,and then expands the security margin.The thesis provides technical reference for promoting DGs' full participation in the dispatch and control of the transmission and distribution networks.