Model-based condition monitoring of gas turbines for power generation duty

This letter has presented an overview of activities to date with respect to the development and demonstration of anomaly detection functions, which analyze key parameters from an industrial gas turbine during its starting sequence. The software modules have been shown to be able to successfully detect anomalies in the start sequences. The software has been developed using a combination of simulated and actual plant performance data and is currently in the process of being implemented in the field. In the longer term, as the level of operational experience and knowledge grows, the anomaly detection functions can be augmented to include prognostic and diagnostic functionality. This will permit not only the detection of anomalous behavior but also the provision of information to plant operators as to the suspected cause of the anomal