Performance Monitoring Technology For Turbo-Shaft Engines

This paper does research on modeling and performance monitoring of turbo shaft engine.Firstly, the aero-dynamic model of Turbo shaft engine has been set up, based on this model, we get the four-dimensional nonlinear equations and three-dimensional nonlinear equations, the steady-state model and dynamic model of this engine have been accomplished by finding the solutions of the four-dimensional nonlinear equations and three-dimensional nonlinear equations, using interpolation law to establish a simplified start-up model , using NI technical to design the friendly simulation interface.Secondly, performance monitoring technology has been researched. on basis of the Turbo shaft nonlinear model, the state-variable model (SVM) has been established. Proposed a new self-optimizing for the establishment of turbo shaft state variable model, this method overcomes the shortcomings of the disturbance and fitting. Combining the Steady point in the state variable models and steady-state basis points, the adaptive model has been set up by interpolating algorithm or fitting method. On the same amount of fuel, the simulation results of adaptive engine and non-linear model show that the adaptive model has high precision.Thirdly, for the performance degradation of engine during service, the efficiency of the compressor and gas turbine has been considered as health parameters. ASVM has been set up including the health parameters. The coefficient matrixes of the ASVM are got by self-optimizing law. In a single Steady ASVM points a Kalman filter has been set up, the health parameters of the engine have been estimated with high veracity. According to the test data of engine, the R matrix has been confirmed, the impact of Q matrix has been analyzed. By interpolating algorithm or fitting method, the adaptive Kalman filters have been designed, on different work points, the health parameters have been estimated with high precision.Finally, in allusion to the error between the adaptive model and real engine, a neural network has been designed to raise the adaptive engine accuracy, the Turbo shaft performance monitoring lay on this work.