Design And Implementation Of The Fault Prediction System For Spacecraft Based On Telemetry Data

Recently years China’s space industry is developing at rapid speed.with more and more space probe has been put into use, how to ensure the spacecraft Work at a steady state is become a crucial problem. Spacecraft run in the unknown environment, environmental change, performance loss and other reasons may make the spacecraft fault. So to carry out the management of the spacecraft’s orbit is very difficult. It need a lot of manpower and material resources to diagnose and solve when the spacecraft fault occurs. If we can found the abnormal situation in advance through the spacecraft various aspects performance and data inference, we will take preventive measures to reduce the probability of failure in the minimum range, so as to ensure the reliability of spacecraft.Therefore, fault prediction in aerospace measurement and control area has the very strong practical significance.The forecasting accuracy of existing data driven predictive algorithm is not the same for the different changes of the telemetry data. If we can use the most suitable for the variation in the model to predict the telemetry data, then the prediction accuracy will be greatly increased. Based on the above background, this paper makes an in-depth study on the overall requirements, system design and key technology of the Data Driven Prediction System. The specific research work is as follows:1. Analysis of the telemetry parameter variation characteristics, combined with the current mainstream prediction method for telemetry parameters variation of applicability, proposed a coding scheme based on data driven decision making of the telemetry parameters: first treats the prediction data of three major forecasting methods of modeling and the model evaluation model for the prediction of the data to get fit. Finally, according to the forecasting model of the data of prediction of future values. 2. Forecast the overall design of the system, through analysis to the system demand prediction system is divided into data receiving module, data pre processing module, parameter prediction interface module, display module, database management module and so on. And the integration of each module to complete the design of the whole system. 3. Design and implementation of the key algorithms of the system, system function mainly depends on the realization of the prediction algorithm, this thesis according to the main idea of ARMA model, grey theory, curve fitting algorithm, and combined with the design principles of object oriented, design and implementation of the class prediction algorithm, for calling the various functional modules of the system. 4. To verify this system,with the collected telemetry data as an example, the applicability of the system is determined by comparing the real future value and the value of the system.; and through multiple sets of comparative experiments on the system performance is evaluated, the verification system of high reliability...