5TH International Congress on Technology - Engineering & Science - Kuala Lumpur - Malaysia (2018-02-01)

Anfis Fault Diagnostic Model For Turbo-jet Engine

Due to the extreme operating condition of gas turbines, they are very prone to failure. Consequently there is a tremendous demand for fault diagnostics methods for Gas Turbines in order to identify symptoms and get failure warnings prior to the actual failure. While assorted diagnostic models for gas turbine have been introduced, this research focuses on the development of a reference model for diagnostics of a turbo-jet engine using Adaptive Neuro-Fuzzy Inference System (ANFIS) in real time domain. Once the model was tested and trained, set of residuals were generated. Using 3-Sigma control limits thresholds were set for healthy operating gas turbine. Using Graphic User Interface (GUI) in MATLAB, a software was developed. The result show that (i) the model has very low prediction error for different sets of data, (ii) Amplitude-modulated Pseudo Random Binary Signals (APRBS) seem to be very promising method for model training and validating, (iii) Sensitivity analysis shows that the model has very high sensitivity towards incipient and step faults, (iv) GUI software can offer a simple, cheap and reliable diagnostic software, and (v) The model is capable of detection, identification and isolation of multiple faults. The remarks are useful to further enhancement of conventional approach to gas turbine diagnostics.
Farbod Rahimi, Tamiru Alemu Lemma