Energy-based visualisation of an axial-flow compressor system for the purposes of fault detection and diagnosis

Abstract

As the complexity of industrial systems increases, the necessity also increases for more reliable, robust and accurate Fault Detection and Diagnosis (FDD) techniques. This article proposes a methodology for the design of an FDD system, based on an energy visualisation of an axial-flow compressor system. This method investigates the steady state and transients of the system for residual generation. A nonlinear model is derived for this compressor system that are capable of modelling compressor instabilities such as rotating stall, surge and leaks. A power-energy analysis is done using the state variables of the system under normal and fault conditions and are visually presented. Two fault conditions are analysed in terms of power and energy. The first test case introduces a leak in the system. The second focuses on the more complex situation when the system enters an unstable operation, namely rotating stall. The results indicate that the power and energy of the system may be useful for the purposes of FDD