Identification and estimation of power system branch parameter error

This paper presents an off-line approach for network branch parameters errors detection, identification and correction. The proposed approach uses a decoupled augmented state estimator which increases the V-thetas state vector for the inclusion of suspicious series branch parameters. The normal equations technique is used to deal with the resulting augmented model and the normalized residuals test is used to the process of parameter error identification. The idea is to use several measurement snapshots, in order to estimate the suspicious series branch parameters in a sequential form, taking advantage of the fact that those parameters can be considered time-invariant for the time period of those measurement snapshots. The proposed approach enables the identification of suspicious branches through an identification-index vector IIV, where each element of IIV is the ratio between the number of measurements incident to each branch whose normalized residuals are larger than one specified threshold value and the number of measurement incident to each branch of the network. Several simulation results (with the IEEE 14, 30 and 57 bus system) have shown the proposed approach is applicable to single and multiple parameter errors in adjacent and non-adjacent branches. Due to space limitations, this paper presents only results with the IEEE 14-bus system.