Research On Key Techniques Of Real-time Monitoring For Aircraft Flight Safety

Flight safety is the assurance of the Airlines operation. With the development of ACARS (Aircraft Communication Addressing and Reporting System) in civil aviation industry, the idea of"real time"monitoring recognizing and diagnosing of the aircraft safety condition is introduced to civil aircraft. Therefore, to meet the tide of real time flight safety monitoring, the application of the ACARS in civil aircraft is researched in this thesis, the structure of the DFDR/QAR(Digital Flight Data Recorder/Quick Access Recorder)data frame and ACARS report data frame are analyzed, and the structure of real time flight safety monitoring is proposed. Based on real time flight safety monitoring architecture, deep researches on aeroengine gas path condition monitoring, flight control system trend monitoring and flight path reconstruction problem are carried out. The application of route plane through ACARS link layer display that the real time flight safety monitoring system based on the ACARS transceiver will play an important role in aircraft safety monitoring. The main content and innovation in this paper may be summarized as follows:1. The historical data source provided by QAR is used to solve the insufficient data for flight safety monitoring system, so the fast flight data decoding system is researched to ensure the real time application. A new decoding algorithm based on dynamic list is proposed which can fulfill the universal decoding according to the user's requirements; the FDDAS (Flight Data Decode and Analyse System) is developed. With the contrast to the traditional method of QAR decode method; the algorithm designed in this thesis can be more efficient. Decoding system guarantee the training and validation data set for real time safety monitor model.2. ACARS report data set is used for real time safety monitor to identify the condition of the aircraft. Researches based on PATS-EG platform are carried out to enhance the real time ability of ACARS report application. A report decoding method based on ARINC-618 protocol and ARINC-618 protocol is put forward. With the interface of Winsock, the messages distributing function in LAN is achieved. And the design of TPlus index accelerates the access speed of ACARS report data set. The experiment for route plane ACARS date link validates the efficiency of ACARS report decoding algorithm.3. The simulation of VHF communication between ground station and ACMS of route plane is built under ACARS-EG. Based on it , the architecture of real time flight safety monitor is proposed, then the range of research is established include aeroengine gas path condition, trend of flight control system and flight path reconstruction. In the process of flight safety monitor, the data sets from different flight data acquisition systems are of multi-sources. In order to keep the consistency in selecting and describing multi-sources, the data pre-processing as well as the integration of multi-sources is given in this thesis.4. The theory and the algorithm of the Support Vector Machine (SVM) are applied in aeroengine gas path condition monitor. The deviations of engine cruise data are analyzed, and the data set for modeling is obtained. The architecture of condition monitor model is established. Model parameters selection is conducted using Pattern Search Method; a remote diagnosis model based on SVM is built. By decoding and storing the ACARS report, real time cruise data sets are acquired, and the monitor model is adopted in processing real time data set for condition recognition. Model output of route application shows that SVM algorithm is reliable and accurate, and condition monitor model based on SVM is suitable for real time monitor of gas turbine engine.5. In order to monitor the condition and trend of flight control system, the application of support vector regression machine (LS-SVR) to predict the trend of flap system. By analyzing the fault of flap system, a model based on time-consuming of flap action is presented to predict the trend of flap. Then, LS-SVR algorithm is trained to establish the time-consuming model. The final prediction error (FPE) principle is used to determine the embedding dimension, and an adaptive grid search (AGS) method is brought forward to optimize the model parameters of LS-SVR. Simulation result shows that LS-SVR model is feasible and accurate for flap system trend prediction, and can extend its application to other flight control system.6. Real time flight path reconstruction problem for flight safety monitor is investigated under interpolation algorithm; a novel method based on cubic spline is presented to solve the sparseness of ARARS report. Cubic spline interpolation is compared to linear interpolation for the reconstruction of flight path; the reproduction of track is performed with finite address report. The validity and rationality of the interpolation method is verified through a simulation base on PATS platform.