Energy based visualisation of a Brayton cycle power conversion unit for the purpose of condition monitoring

Abstract

The efficient operation of industrial plants has been the subject of many studies, due to an increased awareness of greenhouse gas emissions and dwindling natural resources. Condition monitoring is key to the efficient operation of a plant. Part of condition monitoring is fault detection and isolation (FDI). Various methods have been developed and implemented for the purpose of FDI in industrial plants. These methods are mostly based on process monitoring, requiring large amounts of data. Based on the fact that an industrial plant can be viewed as an energy transformation process, energy has been proposed as a non-domain specific monitoring parameter for the purpose of FDI. An energy-based representation of the industrial plant will, therefore, be required. Energy-based representations are commonly used during the design phase to optimise the plant or to monitor the plant's efficiency during its lifespan, but not for FDI. In this study, a basic Brayton cycle power conversion unit (PCU) is used as a case study to determine the suitability of an energy-based approach to FDI. The PCU is completely characterised by an attributed graph matrix in terms of energy and exergy. Exergy was identified as being part of the energy characterisation of a thermodynamic power cycle, such as a Brayton cycle. The primary contribution of this study, is the compilation of a unique energy-based signature, from the attributed graph matrix, visualising changes in the operating conditions of the PCU. Two methods are used in this study to create energy-based signatures, namely a residual-based method and an eigendecomposition-based method. Both methods produced unique signatures, specific to the operating conditions of the PCU, which was successfully used for FDI. A third method (enthalpy-entropy error-based method) proposed by du Rand is also presented and as a secondary contribution, the methods are compared and evaluated based on the attributes an effective FDI method should exhibit. As an additional contribution, an exergy ratio is proposed, defining the exergetic conversion ratio of the components in the PCU. The exergy ratio will serve as an indication of the effect a fault has on a component's ability to transform exergy between domains. In this study, it is concluded that an energy-based visualisation of a Brayton cycle PCU can be used for FDI.