Automated procedure for coke microstructural characterization in imagej software aiming industrial application

• Published in: Fuel (Guildford) 2021-11, Vol.304, p.121374, Article 121374
• Main Authors: Agra, Anderson Azevedo, Nicolodi, Anderson, Flores, Bruno Deves, Flores, Ismael Vemdrame, da Silva, Guilherme Lizeiro Ruggio, Vilela, Antônio Cezar Faria, Osório, Eduardo
• Format: Article
• Language: English
• Subjects: Automation; Coke; Coke microstructure; Coke quality; Coke strength; Computer programs; Heterogeneity; Image analysis; Image processing; Industrial applications; Light microscopy; Metallurgical coke; Microstructure; Optical microscopy; Parameters; Porosity; Software; Spatial distribution; Temperature gradients
• ISSN: 0016-2361; 1873-7153
• DOI: 10.1016/j.fuel.2021.121374

[Display omitted] •22.4–19 mm particles preserved microstructure with minimal damage to critical pores.•At least 56.8 mm2 area to reduce edge pores negative effect for coke microstructure.•Keeping other settings constant, five coke pieces provide a representative analysis.•The automated image analysis employed allows obtaining all microstructural parameters.•The defined set of conditions aligned representativeness and operational practicality. Coke strength is closely related to its microstructure, which can be defined as the spatial distribution of the coke matrix and its porosity. Even though the importance of coke microstructure is recognized and many studies were conducted on the subject, up to date there is no standard methodology established to evaluate such material, which makes it difficult to compare and correlate coke microstructure and quality parameters. Among the techniques available to obtain microstructural parameters that may be related to coke strength, the association of optical microscopy and image analysis shows greater potential. However, representatively characterizing coke microstructure has proven to be a complex activity due to coke high heterogeneity, a result of the different behavior of coal components during carbonization and the thermal gradient differences in the cokemaking process. For this reason, it is essential to establish experimental conditions such as the number of samples analyzed, coke particle size and analyzed area for each sample, to circumvent the heterogeneity problem. In this work, it was shown that the analysis of 56.8 mm2 of each of five polished blocks for coke sample, made with coke particles of 22.4–19.0 mm, was the best condition to align representativeness and operational practicality. Besides, a macro was developed to automate the image analysis routine using the ImageJ software, where each step of the procedure was optimized especially for metallurgical coke. Combining the best experimental parameters determined with the fast automated image processing generated a reliable and representative way to characterize the coke porous microstructure.