缺陷的检出概率是零件损伤容限设计和寿命预测的重要基础数据。针对发动机中的关键转动件——粉末涡轮盘中的夹杂缺陷，进行了超声检出概率研究。制作了不同埋深不同孔径的平底孔人工缺陷试样，模拟粉末盘中的非金属夹杂缺陷。采用6组探头的超声水浸分区聚焦检测工艺对上述人工缺陷试样进行检测，获得了不同缺陷的超声检出概率数据，并绘制出缺陷检出概率（POD）曲线。结果表明，采用特定的检测工艺，在95%的置信水平下，以90%概率检出的缺陷当量尺寸为0.167 mm。

The probability of detection (POD) of defects are fundamental data for the damage tolerance design and serving life prediction. In this work, the POD of ultrasonic inspection was investigated for inclusions within powder metallurgical turbine disks, which are vital rotating parts of aeroengines. Artificial FBH samples were fabricated with different diameters and buried depths, as mock-up for non-metallic inclusions in disks. The zoned ultrasonic immersion inspection procedures by 6 sets of ultrasonic probes were used for the inspection of the above-mentioned samples, and the statistical data of POD for different artificial defects were acquired, based on which a POD curve of probability versus defects sizes were drawn. The results tell that for this specific inspection procedure, the equivalent detectable defect size at a probability of 90% is 0.167mm, giving 95% confident level.