The Prediction Research Of Satellite Healthy Status Based On Engineering Parameters

With development of several ten years in aerospace industry,we accumulate much satellite engineer experience and data.It becomes an important thing in this field to solve how to make use of those experience and data to predict the satellite healthy status.It is possible to use satellite engineer data to predict the satellite healthy status with the development of computer science and big data technology.This paper researched the satellite engineer data application method,using it to predict satellite healthy status,with plenty of application values and some creativity.In this paper,according to reading those papers of relative to operation of engineer data,from data pre-operation,No SQL database with its store methods and data prediction those three domains to do researches,and combined to DAMPE satellite engineer data to exam the correctness.This paper gives a data pre-operation method,including decision tree,tree pruning,condition of decision tree to be strong and the scalability of decision tree,data store method on four popular No SQL database to store the engineering data,and operating method with four data analyzing methods,including polynomial fitting,ARMA,nonparametric regression and back propagation network to operate the engineering data,to give a status to the payload according to the data.In the experiment,we can get the result that adjusted decision tree and its development methods can operate the classification problems.After decision of parameters,besides polynomial fitting,ARMA,and back propagation network also can predict the progress of engineer data,can analyze and infer satellite healthy status.