Research On Aeroengine Performance Parameter Prediction Model Based On Chaotic Time Series Theory

With rapid development of China's civil aviation industry, how to protect aircraft flight safety has become an increasingly important issue. Facing the serious marketing competition, it is vitally important for the airlines to achieve any possible profit and minimize the service cost with the prerequisite of safety. Aeroengine is the most important part of plane. Its structure is most complex, and it requires the highest precision and reliability. At the same time, the requirement on performance is the strictest because of being influenced by all kinds of disturb factors under the worst work condition. According to the statistics data, the maintenance cost of aeroengine accounts for more than 30 percent of the total operating cost of airlines, which reveals much controllable margin. So the maintenance engineers of airline companies must detect the aeroengine anomalies in time and take necessary measures to prevent the aeroengine from failure through monitoring the aeroengine performance and forecasting the aeroengine performance parameters. Preventing the aeroengine from failure can save the maintenance costs for the airline companies, and it is of significant importance in ensuring the civil aviation safety and security.Taking into account the nonlinear characteristics of aeroengine and the chaotic characteristics of aeroengine performance parameters, the thesis will do exploratory study for forecasting the aeroengine performance parameters based on chaotic time series theory.In the first Chapter, the background and significance of the research project as well as the forecasting method based on the chaotic time series theory have been introduced. The current research progress and status of chaotic time series forecasting theory and aeroengine prediction have also been summarized both in domestic and abroad. In the second Chapter, the chaotic time series forecasting theory has been introduced and analyzed, including chaotic identification of the time series, phase space reconstruction theory of the chaotic time series, the selection methods and algorithms of the phase space reconstruction parameters (both optimal delay time and optimal embedding dimension included). In the third Chapter, the phase space reconstruction parameters (both the optimal delay time and optimal embedding dimension) are calculated with actual time series data of the aeroengine performance parameters and the phase spaces are reconstructed. In order to determine whether the time series have chaotic characteristics or not, the maximum Lyapunov exponents of all the time series are calculated. This Chapter is the foundation of forecasting aeroengine performance parameters. In the fourth Chapter, the forecasting methods of chaotic time series theory are discussed, including the adding-weight one-rank local-region method and the method based on maximum Lyapunov exponents. The application of chaotic time series theory on forecasting the aeroengine performance parameters is studied based on the above two forecasting methods and the forecasting results are analyzed and evaluated. In the fifth Chapter, a comprehensive forecasting model with the combination of both chaos theory and multi-regression is proposed based on the chaotic time series forecasting theory. The proposed forecasting model is verified through actual aeroengine time series data and the forecasting result is satisfying. At the end of the thesis, computer simulations of the above three forecasting algorithm models are conducted with the programs. The forecasting programs are then encapsulated into single module packages respectively and integrated into the aircraft remote fault diagnosis system. The integrated diagnosis platform is promising in the aeroengine fault forecasting.