A Simulation before Test (SBT) Approach for Sensor Fault Diagnosis in
State Feedback Systems with Integral Error Compensation
V.Manikandan¹, K.Ramakrishnan ^{2 ,} N.Devarajan ^3
1Faculty of Electrical Engineering, Coimbatore Institute of Technology, Coimbatore -641014,India.
²Faculty of Electrical Engineering, Pondichery Engineering College, Pondicherry, India.
³Faculty of Electrical Engineering, Government College of Technology,
Coimbatore -641 013, India.

Abstract:

The aim of this write-up is to present a Simulation Before Test Approach (SBT) towards sensor fault (deterioration) diagnosis in Linear Time Invariant (LTI) state feedback systems. The novelty of the approach lies in associating with each state feedback gain factor a scalar α, which is defined as the Sensor healthiness factor. This scalar is made to vary from 1 (no fault condition) to 0 (full fault condition) in predetermined steps. The intermediate values of α portray the deterioration modes of the sensor. To have a realistic approach towards actual working condition, system and control parameters are varied within their permissible tolerance values. The Integral Absolute Error (IAE) criterion and Integral Square Error (ISE) criterion are employed for extracting the signature of the fault and the classification is done using Artificial Neural Network (ANN) classifier. Though in general the SBT approach is a popular technique for fault diagnosis in analog electronic circuits, it is employed for the first time in this paper for the fault diagnosis in dynamical system using suitable performance indices for signature extraction. The proposed diagnosis approach is applied to a dc motor system to validate the effectiveness of the technique.

Keywords: Diagnosis, State feedback, ANN.

(Full Paper, 991 KB)

BibTex:

(Full Paper, 991 KB)