Using Multidisciplinary Data to Improve Surface Manufacturing


Picture source:

Surface manufacturing is a process that reshapes the spatial contour of an object or creates certain patterns on a surface substrate. It has been used in numerous applications such as drilling, face milling, panel assembly, additive manufacturing and nanomanufacturing. In doing surface manufacturing, it is critical to monitor and control high-dimensional quality characteristics with high accuracy. For instance, consider a face milling process in automotive powertrain production. Surface shapes may affect the part distortion and sealing during assembly, and it is necessary to characterize the global and local variations in the surface shapes in detail. High-resolution surface measurements are expensive and slow and most quality control tools employed in industry, such as control charts, were developed for low-dimensional variation control and may generate excessive false alarms when being applied to high-dimensional surface data are two challenges that usually encounters by surface quality control.

The challenges were tackled by developing a multidisciplinary data fusion model and surface monitoring scheme. The idea is to use the multiple types of data available in a manufacturing plant or lab, such as part birth history, machine tool life and/or process variables such as material removal rate. These types of data that correlate to the surface quality characteristics were incorporated into a spatial data model to improve surface monitoring without having to increase the surface measurement resolution.