A Framework of Big Data Driven Remaining Useful Lifetime Prediction of On-Orbit Satellite

This paper presents a Bayesian framework to predict the remaining useful lifetime of a satellite based on on-orbit telemetry data for the essential performance parameters of critical equipment. Starting from an introduction to data mining and analysis techniques to deal with big data from heritage satellites’ telemetry, this paper describes detailed procedures on how to develop degradation path models, develop prior knowledge of distribution parameters of random coefficients, and establish Bayesian posterior distribution. A general form of the cumulative probability density function of the remaining useful lifetime is created. A literature search indicates that this paper is the first of its kind in proposing a systematic process to predict the remaining useful lifetime of a satellite based on current performance degradation measurements taking into account prior knowledge from heritage satellites. An example is presented to verify the feasibility of the proposed approach and demonstrate how to use it to predict the remaining useful lifetime of the satellite based on telemetry data for Li-Ion battery capacity losses and solar array wing power losses.