During the lifetime of a gas turbine, its gas path components deteriorate gradually and sometimes serious problems happened. Direct physical and indirect model based methods can be used in health monitoring systems for gas turbines. The gas turbine under study is run as part of a combined cycle generation unit, sited in the BAO Steel Power Plant. The basic health monitoring system is based on vibration signal. After the vibration monitoring system failed to detect foreign object damage (FOD) fault, a health monitoring system based thermodynamic model is tried to explain quantitatively why the performance degradation happened, with the foreseeable usage as part of the online health monitoring system. The present work is based upon component level nonlinear gas turbine model, so errors caused by linearization can be avoided. The component level model of gas turbine is built as dynamic model, and the off-design performance of gas turbine is evaluated as the steady-state solution of the dynamic model. A dynamic tracking filter, which tracking field measurements with PI control loops, is incorporated into the gas turbine dynamic model. Output of the dynamic tracking filter is called correction factors, which are used as multiplicative corrective values of component performance parameters (i.e., flow or efficiency) in the gas turbine model. With dynamic tracking filter and aero-thermal dynamic model, the model based fault diagnosing of gas turbine is implemented as a three step process. As a case study, several measurement data sets are tried to detect and isolate FOD fault happened. The result demonstrates that a model based gas path analysis can detect and isolate fault even when no vibration level alarm is reported.
Mortality on grower/finisher-only swine operations in the United States
