Key Technology Research Of Prognostics And Health Management For Power System Of Hypersonic Vehicle In Near Space

For cost efficient and time reduced flights in both commercial and military areas, air-breathing engine powered hypersonic vehicle in near space(HVNS) with its revolutionary dual-use value, valued highly and investigated further by the aerospace powers. As the aeronautics and space technology’s integrated product, hypersonic vehicle features complex flight environment, wide range of flight envelop, high mach characteristics, stringent reliability requirements and so on, all of which propose many new research challenges to Prognostics and Health Management(PHM) technology of vehicle. As a focus, the study of key technology of PHM for power system of hypersonic vehicle is increasingly of great theoretical significance and application importance alone with the rapid development of hypersonic vehicle research.Guided by PHM theory and according to the four key problems of key techonolgy research of PHM for power system of hypersonic vehicle, this dissertation studies thoroughly and systematically the system design, fault diagnosis technology and fault prediction technology of PHM for power system of hypersonic vehicle using methods of information management, system integration, physics-of-failure, pattern recognition, probability theory, mathematical statistics and computer simulation. The main results and contributions of this dissertation are as follows.(1) The PHM system of power system of hypersonic vehicle is designed. Because of the PHM system of power system of hypersonic vehicle requires cover every component of hypersonic vehicle, every subsystem of hypersonic vehicle and the whole system, so the PHM system of power system of hypersonic vehicle is layered architecture. First of all, the composition of the PHM system of power system of hypersonic vehicle is designed, which divide the whole PHM system of power system into ontology PHM subsystem and ground PHM subsystem, and the structure and function of each subsystem is designed, after that, the function of the PHM system of power system is designed, which divide the whole PHM system of power system from function perspective into data acquisition system, fault diagnosis system, fault prediction system, auxiliary decision system and maintenance support system, and the function of every system is studied. At last, the hierarchical structure and physical implementation are designed.(2) The selection technology of sensor system is investigated. The sensor system as an important part of data acquisition system, the selected sensor system is in line with the requirement of the PHM system of power system of hypersonic vehicle, which will directly affect the effectiveness the PHM system of power system of hypersonic vehicle. Therefore, first of all, outlining the development trend and classification of the sensor system. After that, the selection process of sensor system is presented, and then, base on that, the selection of the sensor system is completed by studying the selection of monitoring point, the questionnaire survey of sensor system, and the choice standard and reference opinions of sensor system. At last, following the previous study, taking lithium-ion batteries as a example, the selection of temperature sensor system of lithium-ion batteries is achieved.(3) When the historical data limited or in absence of any historical data, the fault diagnosis method based on Multivariate State Estimation Technique(MSET) is proposed. Because of the historical data limited and the fault diagnosis technology requirement of power system of hypersonic vehicle, the fault diagnosis method based on MSET is proposed, which is composed of the state monitoring based on MSET and the fault detection based on Sequential Probability Ratio Test(SPRT). When the historical data limited or in absence of any historical data, through quantitative or not quantitative update the training data, the state monitoring based on MSET is achieved, and then on the basis of the study of SPRT, when the distribution of actual residual sample is unknown, the nonparametric SPRT method is introduced, and the state monitoring based on MSET + the fault detection based on Sequential Rank-Sum Probability Ratio Test(SRPRT) is proposed creatively. First of all, outlining the process of the state monitoring based on MSET, and then, base on that, the key technologies of the state monitoring based on MSET are studied. After that, according to the residual which is computed by the state monitoring based on MSET, the fault detection based on SPRT is studied. At last, following the previous study, taking lithium-ion batteries as a example, the fault diagnosis of lithium-ion batteries based on MSET is achieved, among that, when the lithium-ion batteries are fault detected, according to the first type of errors and the second type of error value is setted different, the abnormal type can be divided into "Abnormal level 3", " Abnormal level 2", " Abnormal level 1" and "Fault", and comparing the the results of the fault diagnosis based on MSET with the results of threshold testing of lithium-ion batteries, illustrating the fault diagnosis based on MSET is effectively.(4) The fault prediction method based on MSET is proposed. First of all, by identifying the relationship between the actual residual which is computed by the state monitoring based MSET and the degradation of product, the degradation model of product is established. After that, according to the historical failure data of product whose failure mechanism is the same with actual monitoring product, the abnormal standard of the product is determined. At last, following the previous study, taking lithium-ion batteries as a example, the fault prediction of lithium-ion batteries based on MSET is achieved, among that, according to the historical failure data of lithium-ion batteries whose failure mechanism is the same with actual monitoring lithium-ion batteries, the degradations of lithium-ion batteries in several different types of abnormal type are calculated, which making the abnormal standard of the lithium-ion batteries in several different types of abnormal type is determined, and acoording to the abnormal standard, predicting the state of time when the lithium-ion batteries reach to "Abnormal level 3", " Abnormal level 2", " Abnormal level 1" and "Fault".