Multi-agent Identification And Estimation Of Grid Branch Parameters Based On Improved Clustering

Grid branch is the carrier of electricity transmission, which plays an extremely important role in power system, and the mathematical model of power system also needs accurate circuit parameters. In practice, however, the parameters of input error, the actual measurement error and the causes of branch aging, are likely to cause the branch parameter errors, which have long-term adverse effects on the state estimation, power flow calculation, fault analysis of power system scheduling, control, security evaluation and so on, and even lead to the issue of harmful control action and affect the normal and stable operation of power system.Error parameters and bad results can directly lead to the results of parameter estimation not exist or unreasonable, for which this paper makes a deep analysis and research, and the main contents are as follows:Firstly, the theory of clustering analysis was studied, and the clustering of definition and basic process was understood. Based on this, a clustering method was selected, whose characteristics were analyzed and summarized, and then the disadvantages of this method were improved; As an important basis of grid branch parameter identification and estimation, data was analyzed by the improved clustering method.Secondly, a method of identifying suspicious parameters and positing bad data by clustering multi-section calculation of parameters was proposed. Based on this, a method of correcting suspicious parameters and bad data was proposed to obtain approximations with less deviation. By utilizing these data, both performance of convergence and results reliability of augmented state estimation were improved.Finally, the comprehensive relevance was got by means of analyzing three aspects of physical connection, electrical connection and power transmission and adopting game theory. It can become the basis of avoiding residual contamination and residual flooded, as a result to further improve the determination accuracy. To enhance efficiency, multi-agent theory was introduced to achieve the performance required above, and BAG and CAG scheme and amend to guarantee reliability of final results.