广义帕累托分布在轴承钢夹杂物评价方面的应用

通过ASPEX试验检测技术与广义帕累托分布(GPD)模型,分析了中国和日本4批超洁净GCr15轴承钢中的TiN夹杂、氧化物夹杂与硫氧复合夹杂3种类型夹杂物的数据。发现所有钢中TiN夹杂物均可通过GPD模型计算出最大夹杂物极限尺寸,但氧化物夹杂、硫氧复合夹杂则在某些情况下由于数据中存在少量远大于平均值的极端值而无法得到极限尺寸;这是由于氧化物夹杂在产生后上浮的过程中可能由于偶然的碰撞、聚集长大至很大尺寸,硫化物容易包裹氧化物形成较大的硫氧复合夹杂物,而微米级的TiN多是在钢水凝固过程中析出的,聚集长大可能性小。另外,GPD模型计算结果表明,转炉工艺对于氧化物夹杂的尺寸控制较电炉工艺为优;但硫氧复合夹杂物的尺寸与冶炼工艺的关系更为复杂,无明显规律。

Application of Generalized Pareto Distribution in evaluation of inclusions in bearing steels

The non-metallic inclusions in four batches of ultra-clean GCr15 bearing steels produced in both China and Japan were scanned and measured by ASPEX and then classified into three categories, i.e. TiN, oxides and the sulfur-oxygen composite. And the inclusions were also analyzed by the Generalized Pareto Distribution (GPD) probability model. It was found that the maximum inclusion limit sizes of TiN in all four batches can be calculated by the GPD model. However, they could not be calculated for both oxide and sulfur-oxygen composite inclusions in some cases due to a small number of extreme values far greater than the average value. This was because oxide inclusions may grow to an extremely large size due to accidental collisions and aggregation during the floating process after generation, and sulfides are easy to encapsulate oxides to form large sulfur-oxygen composite inclusions. Different from these two types of inclusions, TiN inclusions were mostly precipitated during the solidification of molten steel, and the possibility of aggregation and growth was small. In addition, the GPD calculation results indicated that the maximum oxide inclusion limit sizes in the BOF process was smaller than that in the EAF process, suggesting that the former was better for controlling oxides. However, the relationship between the size of sulfur-oxygen composite inclusions and the smelting process was more complex and had no obvious regularity.