Diagnosis and Isolation for Multiple Gas Path Performance Degradations of Turbofan Based on a Bank of Unknown Input Observers

摘要

Observers are usually used in on-line FDI system for aero gas turbine engines, especially for their control system sensors. However, when a bank of traditional Kalman Filters is used to monitor gas path performance, it seems difficult to locate or isolate their degradations exactly. This problem will be more difficult when more than one kinds of performance degradations happen. As a result, a bank of Unknown Input Observers (UIOs) is proposed in this paper instead to generate robust residuals for gas path performance FDI. The linear model of a big turbofan considering health parameters is built at first. Then different disturbance matrices are used for modeling to make the residuals of UIOs indicate different characters for different performance declines. Some simulation results are given to verify the multi-FDI performance of UIOs designed in this paper.