Decentralized Robust Bad-Data Pre-Cleaning with Subsequent Hierarchical Fast Recursive Estimation

Abstract

In this paper, a new hierarchical static state estimation method for large-scale power system is introduced. For each subsystem and its tie-lines, a robust local bad-data precleaning (BDPC) (detection, identification and correction) is achieved before applying a new fast recursive state estimator (RSSE). The BDPC technique is based on non-linear transformation by which four bad data indicators (RV. Rθ'. RQ. RQI) are defined and used for precleaning bad data in ail available types of measurements. The fast RSSE is based on the direct use of the measurement redundancy together with the network flow equations to achieve a recursive estimate that verifies the network conditions at stationary opertation. A reasonable data exchange between the interconnected subsystems is exploited to handle tie-lines bad-data at the subsystem level. A comparison with a well known algorithm (Kurzyn. 1983. 1987) is given as well. The proposed algorithm have been applied on the AEP 30-bus test system. The results gained on this network demonstrate that the presented technique is a powerful tool for on-line monitoring and control in large-scale power systems.