Workflow-transfer benchmark of optical surface-roughness measurements against an internal tactile baseline across engineering materials and manufacturing processes

Abstract This study benchmarks exported optical roughness workflows against an internal tactile profilometry baseline across six engineering materials and multiple surface-generation processes. Rather than testing formal optical–tactile equivalence, the analysis examines which optical system–illumination workflows warrant prospective validation for specific material–process–parameter groups, and where tactile confirmation remains necessary. The main contribution is a reproducible workflow-level benchmark that integrates retrospective lowest-discrepancy, fixed-configuration, and leave-one-surface transfer analyses, thereby distinguishing local retrospective optimisation from more transferable performance. Eight profile roughness parameters were analysed. Optical–tactile discrepancy depended strongly on workflow choice. Among the lowest-discrepancy system–illumination combinations available in the archive, the smallest median discrepancies were observed for amplitude parameters including R t , R v , R z , and R z 1 max . Exported R s m values required separate treatment because of scale sensitivity: multiplying retained optical R s m values by 1000 reduced the retrospective lowest-discrepancy median from 99.9% to 17.3%, although 25% of material–process groups still exceeded 30%. Across strictly positive parameters, tactile random variability was generally smaller than the observed between-workflow discrepancies; the relative Type A standard uncertainty of the tactile mean had a median of 1.45%. A fixed-configuration sensitivity analysis showed that the complete-coverage fixed workflow with the lowest median discrepancy achieved 24.8%. In a leave-one-surface workflow-transfer check, selecting workflows from other material–process groups produced a held-out median discrepancy of 32.8%, with 52% of held-out groups still exceeding 30%. Overall, the resulting decision summaries establish a workflow-level benchmark for selecting optical candidates for prospective validation, while clearly distinguishing such candidates from claims of formal optical–tactile equivalence.