|Abstract:|| In many application domains, process monitoring and process optimization have to deal with functional responses, also known as profile data.
In these scenarios, a relevant industrial problem consists in discovering which specific parts of the functional response is mostly affected by the process changes. As a matter of fact, knowledge of the specific locations where the curve is more sensitive to process changes can bring several advantages. It can be exploited to design specific monitoring devices directly focusing on the functional data pertaining to the selected intervals. Secondly, the dimensional reduction can eventually bring to an increase of the power to detect process changes.
This paper proposes a methodology to inferentially select the parts of the output functions that are more informative in terms of the underlying factors. The procedure is based on a non-parametric domain-selective ANOVA for functional data, which results in the selection of the intervals of the domain presenting statistically significant effects of each factor. To illustrate its potential in industrial applications, the proposed procedure is applied to a case study on remote laser welding, where the main aim is monitoring the gap between the welded plates through the observation of the emission spectra of the welded material.|