Title :
Designing model-based fault estimator for a separation column
Author :
Gorinevsky, Dimitry ; Mylaraswamy, Dinkar
Author_Institution :
Honeywell Labs., Cupertino, CA, USA
Abstract :
Considers a practical approach to model-based fault diagnostics. The fault estimation is performed by a delayed FIR filter that is designed as a noncausal Wiener filter. The approach is demonstrated using a case study of a process upset in a separation column in a petrochemical plant. The designed fault estimator demonstrates early and reliable detection of the fault. It presents a conclusive evidence of the fault more than 250 minutes earlier than the action the operator took during the upset based on direct sensory data
Keywords :
FIR filters; Wiener filters; fault diagnosis; nonlinear filters; petroleum industry; process control; state estimation; delayed FIR filter; model-based fault diagnostics; model-based fault estimator; noncausal Wiener filter; petrochemical plant; process upset; separation column; Delay estimation; Fault detection; Feeds; Finite impulse response filter; Laboratories; Manufacturing; Petrochemicals; Process control; Temperature control; Wiener filter;
Conference_Titel :
American Control Conference, 2001. Proceedings of the 2001
Conference_Location :
Arlington, VA
Print_ISBN :
0-7803-6495-3
DOI :
10.1109/ACC.2001.945987
