Skip to Main Content
This paper presents a new Fault Tolerant Nonlinear Model Predictive Controller (FTNMPC). The proposed controller utilizes a Fault Detection and Identification (FDI) scheme based on a novel adaptive extended Kalman filter (EKF) technique. For this purpose a Multi Sensor Data Fusion (MSDF) methodology is incorporated to enhance the estimation accuracy and reliability. A series of illustrative sensor test scenarios has been organized in a Continuous Stirred Tank Reactor (CSTR) benchmark process, which is a typical nonlinear process case study, to comparatively assess the resulting performances of the proposed FTNMPC against typical sensor faults consisting of calibration biases and excessive-variance noises.