Abstract
With the advent of the Internet of Things (IoT) in manufacturing applications, current research is aimed at utilization of IoT data for process control. This article presents a novel approach to performing automated process feedback using a cloud-based IoT architecture. Specifically, a case study of automated spindle speed adjustment to enhance tool life in a machining operation is used to evaluate the proposed architecture. A data-driven model of tool flank wear evolution in a longitudinal turning operation is created on a cloud-based platform through measurements of a polyvinylidene fluoride thin film sensor voltage data and machine tool parameters monitored via MTConnect. The data are used to develop a Gaussian process regression (GPR) model to predict the average tool flank wear as a function of the measured quantities, which is then used to predict the remaining tool life. The performance of the GPR model is evaluated using 10-fold cross-validation and is shown to be sufficiently accurate for predicting the average flank wear with the coefficient of determination (R2) and the root mean square error values of 0.96 and 13.45 μm, respectively. A web application running the GPR model on the cloud platform is used to forecast the remaining tool life during a turning operation and when the predicted remaining tool life is less than desired, the web application commands a spindle speed override to automatically extend tool life. The architecture is demonstrated through longitudinal turning experiments on stainless steel 316L to extend tool life by 82 %. In addition, the latency of the architecture is evaluated and shown to be acceptable for the tool life enhancement application considered in this study.