摘要:Condition monitoring has traditionally been deployed for the monitoring of a single component, such as a rotating machine. The application of condition monitoring techniques to entire petrochemical process plants, however, remains lacking. This can be ascribed to the significant complexity involved in applying existing techniques to such plants. In this work a novel energy-based fault detection scheme is applied to an autothermal reformer of a gas-to-liquids process. The performance of the technique is evaluated, against a set of representative faults, and it is shown to provide adequate fault detection performance. The use of the energy-based detection scheme shows promise in terms of reduced modelling efforts and increased computational efficiency.