High-order Sliding Mode and High-gain Observers for State Estimation and Fault Reconstruction for a Nonlinear CSTR

The aim of this paper is to design a high-gain with a higher order sliding mode (super-twisting) observer-based outputs and faults estimations for a continuous stirred tank reactor (CSTR). The high-gain observer is used in nominal plant for observing the system states. Whereas the adaptive-gain super-twisting is proposed to tackle system faults and therefore can be reconstructed from the equivalent output error injection signal. The designed observer thanks to the higher order sliding mode allows reconstructing faults without chattering effect caused by traditional sliding mode. The global observer convergence is proved using the Lyapunov theory. Simulations results show the effeteness of the observer adopted in this paper.