Fault detection and isolation is one of the major challenges every chemical industry is concerned with, because it deals with the performance of the machinery. Fault detection methods play an essential role in reducing costs, increasing safety and minimizing the effect on the surroundings. To study and compare various Fault detection and isolation techniques, the mathematical model of a continuously stirred tank reactor (CSTR) is implemented in MATLAB. The output, which is obtained from this model, is fed to Kalman filter, which in turn gives a residue that is used to detect the faults present in the system. The difference between the measured process variables and their estimate is called a residue. The same method is used for detecting faults for non-linear state equations with Extended Kalman filter and the residue is obtained from the same. From the residues obtained from Kalman and Extended Kalman filter a conclusion is made that the Extended Kalman filter gives more accurate results. © Research India Publications.
P. V. Sunil Nag, Nair, S. R. P., Gowtham, M., Sibichakravarthy, V., Shivendran, T. D., and Manjunath, S., “Kalman filtering techniques for fault detection and diagnosis in continuously stirred tank reactor (Cstr)”, International Journal of Applied Engineering Research, vol. 10, pp. 37346-37350, 2015.