With the continuous development of the space industry, the role of satellite become more and more important in China’s national economic construction, disaster prevention and mitigation. Power system is one of the important sub-systems that directly impact the in-orbit safe and affection of satellites. Satellite solar array determine the current output of whole satellite. The paper shows the solar array current prediction method based on unsupervised learning which can solve the low-orbit satellite solar array current prediction problem. This method introduces the competition elements that establish the mapping relation between the historical data and the competition element, obtains the best sample through the competition between the competition elements in the prediction processes, the relation functions take the best sample data as the benchmark which can realize the prediction of solar cell array output. Through competition the information of temperature, earth reflection, conversion efficiency and attenuation factors in the sample data are introduced effectively, and the description of such factors in the prediction process is avoided. Through the actual data analysis, we realize the extrapolation of the one-year current mean error is not more than 0.4 a, and the maximum error is not more than 0.5 a. The prediction algorithm for the solar cell array of low orbit satellites without the mathematical description of temperature, earth reflection, conversion efficiency and attenuation factors can predict the reasonable introduction of the above factors.
Low-Orbit Satellite Solar Array Current Prediction Method Based on Unsupervised Learning
Lect.Notes Social.Inform.
International Conference on Wireless and Satellite Systems ; 2020 ; Nanjing, China September 17, 2020 - September 18, 2020
28.02.2021
13 pages
Aufsatz/Kapitel (Buch)
Elektronische Ressource
Englisch
Fault probability reduction method for low-orbit satellite solar cell array diverter
Europäisches Patentamt | 2023
|Solar array driving swinging method for low-dip-angle orbit radar satellite
Europäisches Patentamt | 2015
|Cluster and Integral In-Orbit Solar Array Performance Prediction
British Library Conference Proceedings | 2014
|