Decentralized Optimization Method For Optimal Power Flow Of AC/DC Interconnected Power Systems With Wind Generation Integration

With the integration of large-scale wind power into power grids and the interconnection among areas through AC/DC tie-lines have been accelerated,the optimal power flow(OPF)in AC/DC interconnected power systems with stochastic wind power has attracted widespread attention and become a research hotspot.At present,the centralized optimization method is widely used to solve this problem.However,it faces enormous challenges.First,this method needs to collect the data of each region of the interconnected system to establish a unified model in the dispatching center,which is very difficult to maintenance these data.Second,the regional power grids require independent operation and high data confidentiality.Therefore,it is necessary to study the decentralized optimization method for solving the OPF problem of multi-area interconnected power system with large-scale stochastic wind power access.To address the above problems,the decentralized OPF modeling and the decentralized optimization method are studied in this paper.In the aspect of decentralized OPF modeling,according to actual hierarchical scheduling architecture in China's AC/DC interconnected power systems,the centralized OPF is decomposed into the sub-problems of the provincial networks which are performed in lower-level provincial control center and the sub-problem of the AC/DC tie-lines which is performed in the upper-level regional control center.The sub-problems in the provincial control centers minimize the coal consumption and consider stochastic wind power using extreme scenario based method,while the sub-problems in the AC/DC tie-line dispatch center regards the tie-line power balance corresponding to all scenarios as the goal.By means of the extreme scenarios,i.e.,the largest positive and negative error scenarios,two extreme cases,which may appear in power grids with large-scale wind power integration,can be simulated.Thus,the complex multiple scenarios can be simplified and the computational effort led by the model scale is reduced.Additionally,this decentralized OPF model and the centralized OPF model are completely equivalent,which can be verified further by the comparison of the calculation results.In the aspect of decentralized optimization,according to the original asynchronous-iteration alternating direction multiplier method(ADMM),this paper proposes a kind of synchronous ADMM and increase the relaxation factor to improve the convergence.The improved synchronous ADMM is applied to solve the decentralized OPF model of AC/DC interconnected power system with stochastic wind power access.By testing on a standard system and a real 6056-bus system,it is found that the proposed model and algorithm are reasonable and effective.Besides,by comparing the results of centralized optimization and decentralized optimization,it is found that the model and algorithm are correct.