应用电磁无损检测仪检测钢铁件渗碳层深度时必须标定该仪器的灵敏度。然而由于渗碳工艺和标准渗碳层深度检测方法的限制，无法利用标准试样标定的方法评价电磁无损检测仪对薄渗碳层深度的检测灵敏度。为此，应用灰色预测理论，研究了一种基于有限渗碳层深度集的灵敏度预测方法。在现有有限渗碳层深度集的基础上，通过数据预测，扩展出虚拟深度，从而预测仪器的灵敏度。试验表明，利用GM(1.1)模型对渗碳层深度检测数据建模时 ,其模型平均精度高于95%；利用GM(1.1)模型进行渗碳层深度的预测时 ,预测的最大误差在5%以内。

Sensitivity must be calibrated when nondestructive electromagnetic testing instrument is used to detect the depth of carburized layer of steel piece. However, due to the limitations of carburization technology and the method of testing the depth of standard carburized layer, it is rather difficult to evaluate the sensitivity testing concerning the depth of thin carburized layer by standard sample calibration. Hence, grey prediction theory is applied to make a study on the sensitivity prediction method based on a limit set of carburized layer depth. It is followed by the prediction of instrument sensitivity after the expansion of virtual depth via data. Experiments indicate that the average precision is above 95% when GM(1.1) Model is applied in testing data modeling of the depth of carburized layer, while the maximum error of prediction is within 5% when GM (1.1) is used in the prediction of the depth of carburized layer.