Resource-efficient approaches for predicting cutting forces under tool wear in turning nickel-based alloys for industrial applications: comparison of a White-Box and Black-Box model

Cutting forces are a limiting aspect for the performance of machining operations. Therefore, their prediction is crucial to enable an optimisation of machining parameters. An efficient method for the prediction of cutting forces for entire machining programs are engagement simulations combined with adequate force models. Especially when turning nickel-based alloys, where rapid tool wear is present causing an increase of cutting forces, the prediction of tool wear and cutting forces for a wide range of process parameter values remains challenging. Therefore, this study focused on the development and comparison of industry-friendly approaches: a classical empiric-mathematical model and a Feedforward Neural Network (FNN) model. Based on a comprehensive study of face turning experiments, the models were developed considering an ideal balance between model simplicity and accuracy. Both models were validated for new unseen cutting conditions and achieved good agreement with the corresponding measurements. The presented comparative study highlights the strengths of each approach concerning their application within an industrial environment.