动力装置性能试飞科目中,经常需要飞机稳定平飞,获取稳态飞行数据。但在稳态飞行过程中,由于种种原因,总会出现一些异常值。这些异常值应当被标记,有助于工程师对数据的有效性进行进一步判断。本文主要介绍了用于标记稳态飞行异常值的AEDC(Arnold Engineering Development Center)方法,将AEDC方法与莱茵达准则、肖维勒准则对比,并通过小样本的算例验证,证明AEDC方法对样本数并不敏感,能识别异常点,对稳态数据标记实用且有效。 Power unit performance test flight subjects, often require stable aircraft fly, access to steady-state flight data. However, during steady-state flight, there are always some outliers due to various reasons. These outliers should be flagged to help engineers make further judgments on the validity of the data. In this paper, we mainly introduce the AED (Arnold Engineering Development Center) method, which is used to mark the steady-state flight anomalies. The AEDC method is compared with the Rhinenda criterion and the Schwauler criterion and validated by a small sample. Number is not sensitive, can identify abnormal points, for steady-state data marking practical and effective.