254 Prakt. Met. Sonderband 52 (2018) 
Zum Gahr et al. [3] proposed a new microstructural parameter relating all these quantities Mn 
in white cast irons. Dogan et al. [4] performed different abrasion tests and analysed the #”" 
microstructure of TiC MMCs, finally concluding that the mean free path in the matrix had the © “ 
best correlation to volume loss throughout all the tests. Polak et al. [5] employed linear oe 
regression to detect which parameter has more influence in the abrasion behaviour in Ni- for 
based systems with WC carbides. Zhang et al [6] published a relation between the effective i 
groove cross-sectional area and the ceramic-particle reinforcement volume content and its ~~ 
mean size in aluminium matrix composites. Nilsson and Olsson [7] investigated the relation 
between size and morphology of carbides in high-speed steels (HSS) regarding their 
tribological behaviour. Péhl et al. [8] determined an approximated mathematical expression oo 
for the removed volume and carbide content in multi-phase metallic materials. oes 
Despite the importance of these parameters in ongoing research in the field of tribology, the 
most employed method to characterize 3D microstructures based on 2D images is still the 
linear intercept (LI) method [9]. The LI method consists in drawing lines on a micrograph 
and counting the number of intercepts of the carbide-carbide and carbide-matrix interface 
as well as measuring the length of each segment [9]. Because of such a repetitive and 
subjective work, the results obtained strongly depend on the image quality and user skill. 
Even though that this method can be partially implemented in some image analysis software 
[5]. the information that can be extracted out of it is limited: there is no information about 
size dispersion or shape of the hard phases. Therefore, a new method was developed with 
the following three questions in mind [10]: 
(i) How big is the carbide size dispersion? 
(ii) How regular is the shape of the carbides? 
(ii) How do these attributes affect the relevant parameters to wear resistance? 
With this shortage and a future computational application in mind, the authors developed a This COM 
new stereological method that (i) is based on features that are measurable with any image / 
analysis software, such as areas and perimeters and (ii) relates the carbide size distribution 
and its shape with the mean free path in the matrix and the carbide intercept size. 
In particular, the application in the field of PM steels is promising, where the manufacturing 
processes allow more control over microstructural parameters [11], and where the Tse: = 
discussion about the relation between microstructure and abrasive wear behaviour usually ais 
takes into account the content or carbide size [4-8]. This approach is integral, in the sense I. 
that it relates the most important abrasion resistance parameters to the microstructure- al 
descriptive ones, enabling the reproducibility and comparability of the hard phases. 1 Re 
As studie 
2. Characterization methods ho 
2.1. Obtaining the relevant microstructural parameters 
The proposed scheme can be divided in 4 steps for more clarity: 
1- Obtaining an image (or series of images) of the desired material in the mesoscopic ora 
level. N 
2- Converting the grayscale image (or series of images) into binary images. ne 
3- Measuring the required quantities. Moose 
4- Computing the microstructural parameters relevant to the abrasive wear behaviour — 
of the alloy. 
Since the detailed explanation of the mathematical derivation of the method is out of the a be 
scope of this document, only the main features and their implementation will be discussed an 
here. A complete description of each of these steps can be found in [10].