Research On Integrated Management And Analysis Technology Of Spacecraft Test Data

In recent years,with the continuous development of China's space industry,aerospace enterprises have accumulated a large number of multi-source heterogeneous scattered spacecraft test data.It is increasingly urgent to manage and analyze these data comprehensively.And it is the integrated management of spacecraft test data and then the effectively understanding of these data that is the basis of application and analysis of spacecraft data.In this paper,considering the increasingly striking demand of spacecraft testing and that of the comprehensive management and analysis of the mining feature of the spacecraft test data,these two kinds of research are carried out.Firstly,considering that the spacecraft multi-source heterogeneous spacecraft test data is enormous,scattered,various resources and not uniform format,the integrated management technology of spacecraft test data based on Hadoop is studied.By analyzing the characteristics and storage status of structured test data and multimedia test data in spacecraft test data,Hbase and HDFS are used to realize the comprehensive integration and quick query of spacecraft test data.Meanwhile,the multimedia test data of spacecraft batch import and quick search are also realized.The function is verified by using the real satellite operation data and the multimedia data of simulating satellite test process.The above research lays the foundation for the analysis of the mining feature of the spacecraft test data.Secondly,for the spacecraft test data is similar to the typical industrial large data lying in the lower valuable density and the higher requirements of the testers in analyzing application deeply,and in the face of the demand that the tester needs to quickly understand the data and determine whether the data has the analytical value,the analytical techniques of the mining feature of spacecraft test data is studied.By analyzing the characteristics of spacecraft test data and the common spacecraft data mining methods and technologies in requiring data quality and characteristics,the analytical method of the data mining feature,which fuses several analyses like association analysis,principal component analysis,envelope analysis,data fitting and classification-clustering method,is presented.The algorithm function is verified by using the simulation data set and the public data set.The significant degree of the relevance,trend and classification characteristics in the various parameters of the target data set is displayed by combining two-dimensional visualization method with threedimensional one.This result can help testers to determine the test data mining value.Finally,the demonstration software,which can realize the integrated management and the analysis of data mining feature of spacecraft test data,is designed.This software also has the functions of importing,querying and downloading real satellite test data,and analyzing the mining feature of satellite test data like the correlation,trend,and classification characteristics.By using Python and its associated library,as well as combining the way of thrift framework,the constitution,data processing,algorithm implementation and interface design of the demonstration software are accomplished.The thrift framework is used to realize the remote access function of Hbase.The usability and performance of the software are verified and evaluated by using the real satellite operation data provided by the partner.The preliminary experimental results show that the integrated management and analysis technology of spacecraft test data proposed in this paper has good adaptability to the real satellite test data.This technology can realize the integrated management,rapid retrieval and bulk download of satellite heterogeneous test data.And based on this,the data mining feature is analyzed and displayed to help testers to effectively understand the mining value of satellite test data,which will support the subsequent more complex data analysis and mining work.