FAULT DIAGNOSIS OF A LAUNCH VEHICLE ACTUATION SYSTEM USING WAVELET NEURAL

Abstract

Small Satellite Launch Vehicle (SSLV) is a four-stage launcher that is being developed by the Indian Space Research Organization (ISRO) for launching small satellites and has the potential to enable multiple orbital drop-offs. Of the four stages, three of them are solid propellant stages that used the Flex Nozzle Control (FNC) system for controlling the deflection of the rocket nozzle. FNC system uses Electromechanical Actuators (EMAs) for closed loop position control in the pitch and yaw axis of SSLV. Fault diagnosis is a critical problem in spacecraft operations in terms of performance, safety, and reliability. Feature extraction for the generation of the dataset is done using Wavelet Transform (WT). A fault classifier based on Wavelet Neural Network (WNN) is used for the diagnosis of five system conditions. When compared with the conventional artificial neural network (ANN), WNN considers the fault type comprehensively and provides more accurate results. The modeling of launch vehicle actuation system and the fault classifier network was evaluated in the MATLAB environment