Grinding of glass : optimization of process parameters to improve surface roughness

International audienceService life of machined brittle material and particularly glass workpieces, used in technology fields such as civil engineering, transports or optics, could be shortened by local irregularities. Indeed, grooves or sub-surface damage (SSD) can lead to cracks and fracture in operation [1,2]. Therefore, manufacturers tend to smoothen the surface finish by achieving a long and costly polishing phase. In a materal removal context, reducing damage during rough machining could also be considered since it could limit the polishing operation and eventually limit SSD. This study focuses on this proposal.Since roughness depends on the manufacturing process, material removal process must be carefully defined to obtain a surface finish that could match a service life criterion. Grinding process involves numerous parameters that can be classified in three categories : tool parameters (wheel shape, dimensions, bond material, abrasive nature and size...), machining parameters (depth of cut feed rate, wheel speed, grinding direction...) and contact parameters (temperature, lubricant type and flow, ...). As for roughness, main normalized parameters used remains Ra, the average roughness [3,4] and Rq, the root mean square roughness [5]. However, these parameters provide information on the global surface finish only and especially not on local damage which could lead to crack propagation and fracture [6]. As this study aims at optimizing the rough grinding process on a local damage point of view, the roughness parameters that would characterize the most pertinently the rough ground surface of a brittle material workpiece need to be defined.Using appropriate roughness parameters, the grinding process can be improved by the understanding and optimization of the relationship between the surface finish and machining parameters. For this purpose, depth of cut, feed rate, wheel speed and grinding direction influence on the surface finish of ground window glass samples will be investigated through a central composite design of experiments. Based on the response surface methodology (RSM) [7], it allows to evaluate every parameter influence simultaneously. Even if some of these parameters influence has already been studied [8,9], it has always been partially (one parameter alone) and mainly in a polishing context