Search

The rise of low-pressure turbine outlet temperature is a characterization of three-shaft gas turbine engine degradation. This paper provides three methods to predict the remaining useful life of three-shaft gas turbine engine, which are regression analysis of standard low-pressure turbine outlet temperature, relevance analysis and multiple regression, regression analysis of heat loss index obtained based on mathematical model of gas turbine. Example shows that the three methods all can predict the life. But in the first one the linear regression can't explain the temperature fluctuations well, and the predictions of the quadratic and cubic regression are not consistent with the actual situation. The multiple regression analysis takes the multiple monitoring parameters related to low pressure turbine outlet temperature into account, increasing the adjusted R-squared, but it couldn't consider the influence of compressor washing, and the prediction interval is large, which may not meet the needs of the project. The third method can accurately know and guide the compressor washing time according to certain degradation rate, and can predict remaining useful life effectively, which is a feasible way for gas turbine health assessment and maintenance decision-making.

The application of prognostics and health management (PHM) technology in the marine gas turbine engine could improve engine reliability and availability and reduce life cycle costs effectively. This paper analyzed the problems of PHM applying to the marine gas turbine engine, such as the limits to fix sensors, the limits of monitoring means, the difficulties of weak fault diagnostics, the difficulties of extracting degeneration features and judging the fault point, and proposed several strategies to solve the problems, such as developing the high temperature wireless intelligent sensors, expanding the monitoring means and weak fault diagnostics techniques, researching combined prognostic methods, developing long-distance distributed systems and so on.