Análise exploratória e correlação das variáveis de lingotamento contínuo de uma siderúrgica.

Abstract

Following the industry’s trend of collecting and guiding the decisions based on data, this essay utilizes data analysis tools, a subdivision of data science, to treat and analyze a group of data, taken from a steelmaking plant’s databank. The objective of this work was to execute an exploratory data analysis on the information collected on the process of continuous casting, detect trends, and check the correlation between variables, and between variables and quality defects. H istogram plo ts , boxplots, and s catterplots were u sed to analyze the data using a statistical methodology, and as a last step, a correlation between variables, and variables and quality defects was done. T rends in relation to conti nuous casting speed and hydrogen content in the steel were found, besides weak correlations that indicates some dependence on the relationship between time spent by the ladle on the continuous casting tower and hydrogen content, and hydrogen content with the presence of longitudinal cracks on the slabs produc ed in the steelmaking plant. Finally, h ypotheses and future works proposals were presented.