Fully distributed bad data processing for wide area state estimation

This paper exploits the possibility of fully distributed bad data processing for wide area state estimation in large interconnected power systems. Built upon our recent work on fully distributed state estimation, in which no central coordinator is needed for provable convergence of state estimation with the centralized results, this paper presents a new method to detect and identify bad measurement data in such a distributed setting. Based on the concept of error residual spread decomposition, each administrative control area performs its own Chi-squares test and largest normalized residual test. Only the largest normalized residual for each control area is exchanged with a predetermined set of control areas, all of which belong to the same error residual spread area. The communication burden for the control areas is moderate, therefore, the proposed method is applicable as part of real-time wide area monitoring tools. Numerical examples using the IEEE 14-bus system show that the proposed distributed bad data processing algorithm leads to the same results as centralized methods.