Historical Data Analysis in Quality Improvement of Aluminum Recycling Process

This paper describes an experimental investigation into the effects of process parameters on product quality during an aluminum recycling process. In order to build a model relating recovery to the process parameter variables and then use the model for prediction, process optimisation, or process control, a historical data analysis is employed by observing patterns in the datalogged information at IMCO Recycling Ltd. In previous work, we emphasised the importance of expressing the results of an experiment quantitatively, by implementation of a Taguchi approach, to facilitate understanding, interpretation and implementation. This research presents the technical tools that are needed to achieve quality improvement in recycling of aluminum. In this study we apply the use of historical data matching through a Taguchi orthogonal analysis, in order to control and improve the quality characteristic. For this purpose, a numerical model is developed to characterise the inputs and outputs of the process. A response surface methodology (RSM) is employed to find a suitable approximation for the true functional relationship between response and the process parameter variables. The method of least squares is used to estimate the approximating polynomials. Subsequently, Taguchi's method is used to determine the relative significance of each of the parameters. Finally, process management software (written in Visual Basic) is described, and its interface with the process modelling and optimisation routines is discussed.