State and Fault Parameter Estimation Applied To Three-Tank Bench Mark Relying On Augmented State Kalman Filter

S.Abraham Lincon¹ D.Sivakumar¹ J.Prakash²    

^1. Department of Electronics and Instrumentation Engg, Annamalai University, Annamalai Nagar,,INDIA

^2. Department of Instrumentation Engg, MIT Campus, AnnaUniversity, Chennai, INDIA.

Abstract

Fault detection and diagnosis (FDD) can be described as early determination (detection) and localization (diagnosis) of faulty elements in a dynamic system. In this paper, a model based approach to detect and diagnose abrupt and slowly varying faults in a three-tank benchmark system is developed. The Fault detection and diagnosis scheme is formulated as a state estimation problem by considering the fault parameter as an additional state. It is then solved as a simultaneous state and fault parameter estimation using Augmented State Kalman Filter (ASKF) and Two Stage Kalman Filter (TSKF). Extensive simulation studies performed on three tank bench mark system reveal that the FDD scheme is capable of generating reasonably accurate state and fault parameter estimates in the presence of process and measurement noises. This holds good for different types of faults. Further the performance of ASKF is compared with that of TSKF.

 Keywords: Augmented State Kalman Filter(ASKF),Two Stage Kalman Filter (TSKF),Fault Detection and Diagnosis (FDD),Three-Tank System

(P1110708002, 902 KB).

Biographies:

	S. Abraham Lincon obtained his B.E degree is Electronics and Instrumentation Engineering in the year 1984 and received M.E. degree in Power System Engineering in 1987 and another M.E degree in Process Control and Instrumentation Engineering in 2000 from the Annamalai university, Chidambaram. Presently he is working as a professor of the Department of Instrumentation Engineering in Annamalai University. His areas of research are Process Control, Fault Detection and Diagnosis and Multivariable Control.
	Dr. D. Sivakumar obtained his B.E degree is Electronics and Instrumentation Engineering in the year 1984 and further the M.E. degree in Power Systems in the year 1990 from the Annamalai University, Chidambaram. Presently he is working as a Professor in the Department of Instrumentation Engineering in Annamalai University. He is presently guiding many Ph.D research scholars in areas like Fault Detection and Diagnosis, Neural Networks and Fuzzy Logic applied to Process Control, Signal and Image Processing.
	Dr. J. Prakash obtained  his  B.E degree    in    Electronics    and    Instrumentation  Engineering    in   the   year 1993   from  Annamalai    University,   Chidambaram and he got his M.E degree in    Control System Engineering in the year 1995 from PSG college of Technology, Coimbatore. Subsequently he received his Ph.D degree in Chemical Process Control Engineering from the prestigious institute IIT Madras, Chennai. Presently he is working as an Assistant Professor in the Department of Instrumentation Engineering, Anna University, MIT Campus, Chennai. Currently he is guiding many Ph.D and M.S scholars in the areas like Instrumentation, Process Control, Fault Talent Control, Fault detection diagnosis, Multivariable Control, Digital Signal Processing

BibTex: