مقاله انگلیسی رایگان در مورد طبقه بندی ریزساختار پیشرفته با روش داده کاوی – الزویر 2018

The microstructure of advanced steels is usually controlled by sophisticated thermo-mechanical processing or heat treatments post hot rolling [1]. Depending on chemical composition and process control, the microstructure of such steels may consist of a range of different phases. If the microstructure consists of more than one phase, the properties of the material strongly depend on the type and distribution of the respective phases [2]. Therefore, it is crucial to determine the type and amount of the different phases in order to assess the underlying structure-property relationship. Traditionally, microstructures of steels are characterized by using standard metallographic procedures based on chemical etching and light optical microscopy (LOM) and they are classified by comparing the microscopy images with reference series. Especially for steel and its complex microstructures the comparison with reference series is strongly dependent on the expert’s subjective opinion. Nonetheless, steel is still one of the most widely used materials because of its excellent mechanical properties and the huge variety of applications [3]. Therefore, there is significant interest in the devolvement of objective quantification techniques for steels. In order to characterize steel, the microstructures can be etched for example with a structure etching such as Nital [4] or color etching techniques like Beraha‘s etchant [5]. Due to different contrasts obtained by etching the ferritic matrix can be distinguished from a pearlitic, bainitic or martensitic second phase. However, these etchings are limited to empirical approaches and quickly reach their limits, especially for the discrimination of different phase constituents in steels that exhibit more than two phases. Furthermore, the microstructures of complex multi-phase steels are usually too fine to be resolved by light optical microscopy. A proper characterization requires modern metallographic techniques such as high resolution scanning electron microscopy (SEM) or electron back-scatter diffraction (EBSD) [6,7]. Therefore, any approach aiming at identifying the phase constituents of multi-phase steels has to rely on morphological or crystallographic parameters accessible by these techniques [8–13]. Recently, several studies have focused on EBSD for the microstructural characterization of steels, as this technique can provide direct information on the phase composition [6,7,8,14]. For example, in Ref. [14] a multitude of steel grades from different manufactures has been studied and an EBSD-based classification model was proposed. It was shown that the kernel average misorientation (KAM) deduced from EBSD measurements can be used to distinguish between ferrite, bainitic ferrite and martensite. Although those EBSD-based approaches have proven to work out for some steels, the phase separation by means of EBSD is very subjective as it strongly depends on a proper selection of  the preparation, measurement and evaluation parameters [14].