Published: 2008
Total Pages:
Get eBook
The performance diagnostics of any engine model is accomplished by estimating a set of internal engine health parameters from available sensor measurements. These sensors which comprises of a variety of gas path measurements e.g. pressures, temperatures, fuel flow and spool speeds provide information regarding the health of the engine. No physical measurement, however, elaborate or precise, or how often repeated, can ever completely eliminate the universal presence of measurement uncertainties. Instrument measurements are often distorted by noise and bias, thereby masking the true condition of the engine leads to incorrect estimation results. Measurement uncertainties encourage the inaccurate fault diagnosis, and in order to improve the reliability of diagnostic results, it is important to statistically analyse the data scattering caused by sensor noise. Leakage analysis is a key factor in determining energy losses from a gas turbine. Once the components assembly fails, air leakage through the opening increases resulting in a performance loss. Therefore, the performance efficiency of the engine cannot be reliably determined, without good estimates and analysis of leakage faults. Specifically, for energy calculations it is the air flow leaking around the components at operating conditions that is required. Consequently the implementation of a leakage fault within a gas turbine engine model is necessary for any diagnostic technique that can expand its diagnostics capabilities for more accurate predictions. The simulating methods should either, precisely measure the size of leaks or measure the air flow along gas path with sufficient accuracy. In this research, the diagnostic tool that used to deals with the statistical analysis of measurement noise and leakage fault diagnostics is a model-based method utilizing non-linear GPA. For the purpose of diagnostic, the simulation code used in this study is TURBOMATCH and the engine model Trent 500. TURBOMATCH is the name of a.
