Research Protoype Industry

The paint of modern cars consist of many coats applied in a highly complex and automated process. An optimal paint quality is not only needed for decoration purposes but also protects the car from damage by environmental influences. Paint shop managers need to handle complex chemical processes and optimise many parameters simultaneously to achieve the desired quality. To support the paint shop managers in this complex task, a quality prediction model was created which enables the mangers to test different parameter settings before they are applied on the shop floor. The managers can interact with the ‚Least Absolute Shrinkage and Selection Operator‘ (LASSO) prediction model via a simple Graphical User Interface (GUI) shown in the video. What happens in the video:

  • The first step when working with the model is the selection of the underlying data. By setting the beginning and the end point in time, the corresponding model data are selected.
  • Similarly, the α parameter can be set by the user. This parameter determines how many of the most influential process variables should be considered. The variables are sorted in descending order of their impact on the outcome.
  • The variables can then be changed by simply dragging and dropping the bubble. The green bars indicate which parameter setting would lead to the largest quality improvement.
  • After each change of the parameters, the forecasted quality becomes immediately visible.
  • The parameter configuration can be printed for distribution purposes.