DIAGNOSTICS OF AERO GAS TURBINE USING OBSERVABILITY ANALYSIS
Abstract
The diagnostic analysis of a gas turbine engine is essential from a maintenance
point view in order to be able detect degradation in any component of the engine.
The early identification of engine component degradation allows the specialist to
take any action necessary to retrieve optimal engine performance. In order to
increase the accuracy of detecting and quantifying the level of degradation that
occurs in any component of the engine, an appropriate set of measurements
needs to be available. Therefore, prior to the diagnostic examination of the
engine, having appropriate knowledge of the measurement parameters that are
capable of producing a high accuracy level in the results is required.
The methodology used in this study is based on observability analysis conducted
on an aero gas turbine engine, the GE90-90B. The observability analysis follows
a procedure used to select the best measurements that are capable of detecting
degradation for both single and multiple components in the engine. The engine
was modelled using Pythia software, which is used for design point and offperformance, design point adaptation and diagnostics. The procedure used for
the selection of the best set of measurements consists of the classification of the
measurement and the analysis of sensitivity, correlation, and component subsets.
Three sets of measurement were used for the process of the diagnosis which
were selected measurements, standard measurements and the component
subset measurements for several component fault cases. Furthermore, in order
to compare the accuracy of the detection of the fault for either single or multiple
components, both non-linear GPA and root mean square (RMS) were used. The
results of the diagnostic prove that the accuracy of detection of the fault using the
standard set is low compared to the other two sets where the accuracy was high.
The suggested solution is to consider the observability analysis prior to the
selection of the standard set of sensors in order to satisfy the purpose of both
control and the diagnosis of the engine.