摘要
运载火箭开展试验评估需要利用小子样的实测的飞行试验数据和数字试验数据进行融合评估。为避免小子样的飞行试验数据的特征被大样本数字试验数据覆盖,提出利用Bayes融合评估方法进行运载火箭试验数据的数实融合评估。基于数字试验数据和飞行试验数据的一致性校验结果可计算数字试验数据的可信度,并从大样本的数字试验数据中利用聚类分析方法选择代表点,最后利用数字试验数据的可信度,对飞行试验数据和代表点进行Bayes融合评估。应用该方法对运载火箭的入轨精度指标4次飞行试验数据和1500次数字试验数据进行Bayes融合评估,评估结果融合了数字试验数据和飞行试验数据的特征,与该运载火箭的设计值更为接近,验证了该方法的可行性和正确性。
Launch vehicle test and evaluation requires the fusion and assessment of small-sample flight-test data with large-scale digital-test data.To prevent the characteristics of small-sample flight-test data from being overwhelmed by large-sample digital-test data,a Bayesian fusion-evaluation method is proposed for integrating digital and flight-test data.Based on consistency-check results between digital and flight-test data,the credibility of the digital-test data is calculated.Representative points are then selected from the digital-test data via clustering analysis,and finally the credibility weight is used to perform Bayesian fusion of the flight-test data with these representative points.The method is applied to fuse four flight-test datasets and 1,500 digital-test datasets for the launch vehicle orbit-insertion accuracy index.Results retain features from both datasets and are closer to the design specification,thereby verifying the method's feasibility and correctness.
作者
陈浩
黄彭奇子
姚惠生
张银辉
蔡文炳
CHEN Hao;HUANG Pengqizi;YAO Huisheng;ZHANG Yinhui;CAI Wenbing(Beijing Institute of Tracking and Telecommunication technology,Beijing 100094,China;College of Arts and Science,National University of Defense Technology,Changsha 410072,China)
出处
《航天技术与工程学报》
2025年第4期79-85,共7页
Journal of Space Technology and Engineering
