Digital Twin for Machining Tool Condition Prediction

Abstract Digital twin introduces new opportunities for predictive maintenance of manufacturing machines which can consider the influence of working condition on cutting tool and contribute to the understanding and application of the predicted results. This paper presents a data-driven model for digital twin, together with a hybrid model prediction method based on deep learning that creates a prediction technique for enhanced machining tool condition prediction. First, a five-dimensional digital twin model is introduced that highlights the performance of the data analytics in model construction. Next, a deep learning technique, termed Deep Stacked GRU (DSGRU), is demonstrated that enables system identification and prediction. Experimental studies using vibration data measured on milling machine tool have shown the effectiveness of the presented digital twin model for tool wear prediction.