Robust Neuro-identification of Nonlinear Plants in Electric Power Systems with Missing Sensor Measurements


Fault tolerant measurements are an essential requirement for system identification, control and protection. Measurements can be corrupted or interrupted due to sensor failure, broken or bad connections, bad communication, or malfunction of some hardware or software. This paper proposes a novel robust artificial neural network identifier (RANNI) by combining a sensor evaluation and (missing sensor) restoration scheme (SERS) and an ANN identifier (ANNI) in a cascading structure. This RANNI is able to provide continuous on-line identification of nonlinear plants when some crucial sensor measurements are unavailable. A static synchronous series compensator (SSSC) connected to a power system is used as a test system to examine the validity of the proposed model. Simulation studies are carried out with single and multiple phase current sensors missing; results show that the proposed RANNI continuously tracks the plant dynamics with good precision during the steady state, the small disturbance, the transient state after a large disturbance and the unbalanced three-phase operations. The proposed RANNI is readily applicable to other plant models in power systems.


Electrical and Computer Engineering


Duke Power Company
National Science Foundation (U.S.)

Keywords and Phrases

Auto-Associative Network; Missing Sensor Restoration; Particle Swarm Optimization; Radial Basis Function Network; Robust Neuro-Identification; Static Synchronous Series Compensator

Document Type

Article - Journal

Document Version


File Type





© 2008 Elsevier, All rights reserved.