Most galvanizing plants try to reduce their by-products because significant quantities of the raw material for this sector are consumed: zinc. In particular, many of them try to reduce the humidity of the steel pieces by heating the flux or incorporating a drying oven between the flux and the molten zinc. And while it is true that they are practices that help reduce land (somewhere I have explained why water increases dross and ashes), they do not always succeed in eliminating the variation in the process.

The image above corresponds to an analysis of ash production in a plant that operates without a drying oven and under a process apparently under control. The graph shows the regression adjustment for this by-product and it can be seen that it is not normal (Gaussian) and that the process (from the point of view of ash production) is not under control, despite the fact that the plant operates according to the best practices available. Note that the model fit residuals (top left) are not on the normality line (the hypothesis of normality does not hold). It was verified that ash production responded to a linear growth of the type:

[Kg/t]Ash= 8,1 + 0,1 [Kg/t]_Zn

This is the situation we found and that we solved by a correct statistical control of the process, achieving the result reflected in the graph below.

After improvement, the residuals of the controlled model were better aligned on the normality line and, furthermore, ash production behavior could be modeled according to:

[Kg/t]_Ash = 2,8 + 0,14 [Kg/t]_Zn

Therefore, the process was controlled and – in addition – there was a notable decrease in ash production without the need to introduce a drying oven. This improvement generated enough annual savings to acquire a drying oven in the process and further decrease the generation of this by-product.