This report aims to analyze the variables that participate in the melting process and to validate the arc stability relationships with the other variables in order to optimize the model currently used to calculate the temperature and melt rate during the process. For this we will use the data provided by Gerdau of the identification furnace EAF 75137 and process of ID 3987.
Firstly, to evaluate possible relations between variables, we select only the variables that are measured, leaving the calculated ones from outside, that is, only the variables whose real values we know will be analyzed, avoiding the undesired inclusion of some calculation error. The chart below shows the behavior of all these variables. The data were standardized so we can compare the behavior of two or more variables with different scales.
Analyzing only the variable Arc Stability we noticed the presence of peaks, which seems to be an indication of shutdown. The first peak occurs around 13 minutes, indicating a first shutdown. However, this shutdown is believed to occur around 7 minutes, which is when Arc Stability exhibits stability behavior. The first shutdown seems to have a different purpose from the others leading to believe that it is the moment in which more scrap is thrown in the furnace for melting. Another variable that reinforces the idea of shutdown is the Current variable. When we analyze these two variables together, we notice that the positive peaks of Arc Stability coincide with the Current negative peaks. For Arc Stability we also noticed a decrease in values between the 14th minute and the 29th minute, which leads us to suspect that by this time much of the material released has already been melted. Another important analysis that leads to believe in the shutdown theory is the relation between the Arc Stability variable and the Electrical Energy supplied to the process, which also remains constant shortly before the Arc Stability peaks. Another evidence that strengthens the idea that most of the scrap has already been melted down to the first 30 minutes is the behavior of the variable that corresponds to Methane injection, Natural Gas, which has two significant increases and then tends to have a more steady behavior after half an hour of process.
Besides graphical analysis, it is statistically interesting to analyze the correlation between variables that have a suspected link between them. For this, the Pearson correlation index for cited variables was calculated and we test for significance by analyzing the p-value of the test. If the p-value is less than the 5%, then the null hypothesis that the correlation is not significant is rejected. Otherwise (p-value > 0.05), we conclude that a correlation between variables is not statistically significant. The correlation between the Arc Stability and Natural Gas and the correlation between the Arc Stability and Electric Energy were analyzed. Pearson’s values and correlation indices are shown in the table below.
## Natural Gas Electrical Energy
## P-valor 0.1894084 2.093900e-12
## Índice de Perason 0.0759751 -3.911212e-01As we can see, the tests show that the correlation between the Arc Stability variable and the Electrical Energy is significant while the correlation between the Arc Stability and the Natural Gas isn’t. However the correlation between the Arc Stability and some other variables such as Current, Eletric Power and Arc Voltage is elevated, as we can see in the table below.
## Current Arc Voltage Eletric Power
## Arc Stability -0.7742157 -0.8181938 -0.6338763