The Research On Aeroengine Adaptive Modeling And Fault Diagnosis

The fault diagnosis of aeroengine components and that of areoengine control systems are closely related. They must be taken into account comprehensively. The integrated fault diagnosis including control system sensors and actuators failure detection, components performance monitoring and fault diagnosis of aeroengine is studied based on the platform with high performance microprocessors, which becomes necessary and possible gradually. The three key technologies of the integrated fault diagnosis were researched in this paper. Adaptive modeling based on the measured data, diagnosing aeroengine gas path components fault based on neural network multisensor data fusion and the development and performance testing of a multi-function high performance dual-DSP embedded fault diagnosis platform are the main contents of the paper.Firstly, the neural network identification modeling method is studied and validated by the measured data of a given micro turbine engine. The theories of the improving Back Propagation network, Radial Basis Function network and Wavelet Neural Network were analyzed respectively. The methods of identification modeling and generalization ability testing were put forward. And then simulation analyses were made based on the measured data of the turbine engine. The antinoise properties of the models were researched. And the effect of these models based on the three neural networks was compared.The aeroengine gas path components fault diagnosis method based on the technology of neural network multisensor data fusion and fuzzy logical theory was also researched in the paper. And the method was used in a given turbo-shaft engine. Good results were shown by the experimental results.Finally, in order to offer a running-platform and a simulant environment of airborne computer for the research on aeroengine integrated fault diagnosis, a multi-function high performance embedded fault diagnosis platform based on a fixed-point DSP and a floating-point DSP was developed. And a series of performance testing experiments were operated on the platform. The aeroengine sensor failure detection system was designed based on an in-flight adaptive model. And a hardware-in-the-loop simulation experiment based on the fault diagnosis platform was developed. Experimental results show that the basic function needs of aeroengine integrated fault diagnosis could be met by the platform.