Study On Uncertainty Operation Optimization Of Three-phase Unbalanced Distribution Network

Driven by the "carbon peaking and carbon neutrality goals",renewable energy represented by photovoltaic and wind power has developed rapidly,and the proportion of connected in the distribution network is increasing.Distributed renewable energy with high proportion and distributed connection to distribution network.On the one hand,the output has obvious intermittency and uncertainty,on the other hand,single-phase random connection intensifies the three-phase unbalance characteristic of the system,which brings great challenges to the optimal operation of the distribution network.The deterministic optimal operation decision based on three-phase balanced network model is difficult to effectively guide the operation of distribution network.the correlation between uncertain factors in the distribution network and the three-phase unbalance characteristics of network topology are considered,and the uncertainty operation optimization of three-phase unbalanced distribution network is studied.The main work and achievements are as follows.(1)Unbalanced distribution network power flow analysis and uncertainty portrayalBased on the Taylor expansion of the voltage and power balance equations,a high-precision linear approximate power flow model is constructed for the three-phase unbalanced distribution network to achieve complete decoupling of voltage magnitude,active and reactive power.The parametric probability distribution model,nonparametric estimation model and fuzzy affiliation function are used to characterize the differential uncertainty of wind power,photovoltaic and load,respectively.Rank correlation coefficient and Euclidean distance are introduced to measure the spatial correlation of renewable energy output,so as to determine the optimal Copula function and form the joint marginal distribution of wind and solar output.2)Distributionally robust coordinated optimization of unbalanced distribution network based on spatio-temporal correlation scenariosAiming at the inaccurate estimation of the output and the lack of spatio-temporal correlation representation,a distributionally robust coordinated optimization method for unbalanced distribution network based on multi-source spatio-temporal correlation scenarios is proposed.Based on Copula joint marginal distribution function and improved Kmeans clustering algorithm,typical scenarios with spatiotemporal correlation are generated.Based on the constructed scenarios,1-parametric and ∞-parametric constraints are introduced to construct high confidence scenario uncertainty sets containing real probability distributions without distribution function assumptions.Based on this,the day-ahead distributionally robust coordinated optimization model of distribution network is established considering energy storage system,on-load tap changer and reactive power compensation equipment with the minimum cost within the operation period as the objective function,and the optimal decision of equipment under the worst probability distribution is obtained by iterative solution of C&CG algorithm,achieving the economically optimal objective of day-ahead optimization in the uncertainty scenario.3)Two-stage distributionally robust coordinated optimization of unbalanced distribution network under chance constraintsIn order to further improve the economics of distributionally robust optimization decision-making,a two-stage distributionally robust chance-constrainted coordinated optimization method for distribution network is proposed.The normal cloud model is used to improve the first and second order moments of the traditional moment uncertainty set and improve the problem of conservativeness.In addition,a twostage optimization strategy combining day-ahead offline optimization and day-in real-time optimization is proposed: in day-ahead stage,a distributionally robust chance-constrained coordinated optimization model is constructed based on the improved moment uncertainty set.With voltage and branch power as chance constraints,voltage violation or line overload under certain risk probability is allowed to obtain the hourly action decision of each equipment;In day-ahead stage,minute action decisions are obtained through rolling reactive power optimization to ensure that system voltage and line power meet operation constraints.The test results indicate that the proposed optimization strategy can improve the system’s risk resistance while maintaining high economical decisions.