The knowledge about a relevant process and lifetime indicative quantity, such as the hot gas temperature, is crucial for the control of a gas turbine. Since this indicative process quantity usually cannot be directly measured, it has to be estimated. The paper describes a model-based method to accurately estimate in real-time the hot gas temperature of a heavy-duty gas turbine. The method follows a well-balanced trade-off between resulting prediction accuracy and involved computational complexity. It takes advantage of the capability of a component-level dynamic model to predict the system behaviour and of the capacity of a dynamic tracking filter to adapt to the current gas turbine conditions. In a simulation study, it is shown that the proposed design can provide an accurate hot gas temperature estimation over the entire gas turbine load range, along the gas turbine lifecycle, and during fast transient manoeuvres.

This content is only available via PDF.
You do not currently have access to this content.