Data analytics plays a significant role in the realization of Industry 4.0. By generating context-related persistent datasets, every manufacturing process in real production becomes an experiment. The vision of Internet of Production (IoP) is to enable real-time diagnosis and prediction in smart productions by acquiring datasets seamlessly from different data silos. This requires interdisciplinary collaboration and domain-specific expertise. In this paper, we present a novel tool wear monitoring system for milling process developed in the context of IoP. This system is based on high-frequency data from the numerical control of the production machine without additional sensors. The novelty of this paper lies in the introduction of virtual workpiece quality and fusion of multiple build-in sensor signals and a force model as decision support. This bridges the time gap between quality inspection and production at the shop floor level, establishes an automated statistical process control system, and provides a more plausible prediction of tool lifetime. The monitoring of two different milling processes in a real production environment is exemplary demonstrated in this paper. The first case is a face roughing process with the aim of rapidly removing large amounts of material. The second case is a face finishing operation that follows roughing and aims to achieve the desired surface quality.
A practical digital twin for machine tools is proposed in this study. The proposed digital twin is capable of time-domain simulation of machine tools and consists of a controller model, machining process model, and machine dynamic model. To predict the quality of the machined surface after the finishing processes, a precise dynamic model is required. The developed dynamic model consists of an interaction force model, vibration model, and friction force model. A linear auto regressive with exogenous inputs (ARX) model is adopted for the interaction and vibration models. Based on a systematic analysis of the disturbance forces of the machine tool, the friction characteristics are extracted to a displacement-dependent friction model and velocity-dependent friction model. A nonlinear Hammerstein model is adopted to identify the friction. Online identification systems based on the recursive least-squares (RLS) method are developed and tested for each model.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.