基于不确定度预估的靶场实测数据录取方法

靶场实测数据具有数据样本少、获取代价高等特点,从中剔除异常值,在军事、经济及时间上有重要意义。为降低数据录取给试验鉴定结论带来的误判风险,提出不确定度法,具体由预先评定出的测量结果不确定度和在实际测量过程中获得的最佳估计值来共同确定剔除异常值的实测数据录取方法,结合某型坦克炮身管内径鉴定试验,说明了不确定度法的内容及步骤。通过人为设定样本集并附加异常值进行MATLAB仿真分析,结果表明无论是样本量少,还是样本量多,运用不确定度法剔除异常值具有较高的可靠性。不确定度法充分考虑测量系统及测量过程的实际情况,运用计量结果等现有信息一次性评定出测量不确定度,即可在实际测量中进行重复应用,有利于靶场试验从策划到实施的质量控制。同时,在靶场采用不确定度法录取实测数据,具有简单快捷的特点,不仅满足一线的快速使用,而且还能显化出靶场计量在试验一线的贡献度。

A Method of Measurement Data Admission Based on Uncertainty Estimation

In a weapons testing agency,measured data samples are few and costly to obtain.It is of great military,economic and time significance to remove outliers from them.In order to reduce the risk of erroneous judgment for test conclusions caused by data admission,an uncertainty method is proposed,and specifically the measured data admission method for excluding outliers is jointly determined by the pre-assessed measurement result uncertainty and the best estimate obtained in the actual measu-rement process.Taking the qualification test of a tank gun barrel,the content and steps of the uncertainty method are described.A MATLAB simulation analysis of artificial sample sets and additional outliers shows that the uncertainty method is highly reliable for removing outliers irregardless of the quantity of test samples.It is concluded that the method fully considers the actual situation of the measurement system and the measurement process,and uses existing information such as measurement calibration results to evaluate measurement uncertainty only once,which can then be reused in actual measurement.It is beneficial to the quality control of range tests from planning to implementation.At the same time,the application of this method in the range for measured data admission is simple and quick,which can not only be readily used in the field,but also show the contribution of range metrology to testing work.