Fault diagnosis of a simulated industrial gas turbine via identification approach

In this paper, a model‐based procedure exploiting the analytical redundancy principle for the detection and isolation of faults on a simulated process is presented. The main point of the work consists of using an identification scheme in connection with dynamic observer and Kalman filter designs for diagnostic purpose. The errors‐in‐variables identification technique and output estimation approach for residual generation are in particular advantageous in terms of solution complexity and performance achievement. The proposed tools are analysed and tested on a single‐shaft industrial gas turbine MATLAB/SIMULINK ® simulator in the presence of disturbances, i.e. measurement errors and modelling mismatch. Selected performance criteria are used together with Monte‐Carlo simulations for robustness and performance evaluation. The suggested technique can constitute the design methodology realising a reliable approach for real application of industrial process FDI.