A Symbolic Dynamics Approach for Early Detection of Slowly Evolving Faults in Nonlinear Systems

V. Rajagopalan, R. Samsi, A. Ray, J. Mayer, and C. Lagoa (USA)


Symbolic Dynamics, Finite State Automata,Nonlinear Sys tems.


This paper deals with the early detection of small param eter variations in non-linear systems. The objective is to detect these variations, before they manifest themselves as a sudden change in the phase behavior. It is assumed that these parameter variations occur slowly in time. Early de tection of slow time-scale anomalous behavior in nonlinear dynamical system is achieved by observing and inferring time series data of the state or output variables, at the fast time-scale. The detection algorithm makes use of Symbolic Dynamics and Finite State Automata. As an example, we consider the Duffing equation which is a second order non linear differential equation. The methodology is validated both in simulation and experiment. This concept can be fur ther extended to dynamical systems of higher complexity.

Important Links:

Go Back