数据挖掘技术在飞行试验数据分析和气动参数辨识中的应用研究

为了检验飞行试验辨识结果的可信度，首次将数据挖掘技术用于飞行试验数据的分析和气动参数辨识，初步解决了试验数据有限、数据信息含量差别较大给聚类、分类、回归等分析处理带来的问题；提出了利用不同时间段、不同飞行批次的飞行数据在划分区间的气动特性分布来检验辨识结果可信度的方法；以某飞行器为对象，利用数据挖掘技术，建立了基于飞行试验数据的气动力数学模型，检验了辨识结果的一致性和可信度，并与地面试验结果进行了比较分析，给出了地面试验预测误差。多批次飞行试验数据的整体辨识结果表明，所发展的方法是可行和有效的。该项研究为验证辨识结果的可信度、建立基于飞行试验数据的气动模型提供了新的途径，并可应用于CFD 和风洞试验的验证与确认。

Application research of data-mining method to flight test data analysis and aerodynamic parameter identification

In order to effectively verify the reliability of identification result of flight data, data-mining method is successfully applied to flight test data analysis and aerodynamic parameter identification in this paper for the first time. Problems existing in clustering, assorting, regression because of insufficient number of flight test data and big differences of information among flight test have preliminarily been resolved.A new way of reliability test of identification result is presented according to aerodynamic characteristic distribution in some certain range of different time and different flights.Through dealing with and analyzing groups of flight data of a certain flying vehicle using this proposed data mining technique,aerodynamic model based on flight data is constructed,coherence of the identification results is checked,it’s reliability is evaluated,and the prediction error is given by comparing identification result with ground test data.The whole identification results of groups of flight data of a certain flying vehicle indicate that the method developed in this paper is feasible and effective.This research gives a new way to test reliability of identification result and construct aerodynamic model based on flight data,and it is also useful for verification and validation of CFD and wind tunnel test.