The Industry 4.0 framework needs new intelligent approaches. Thus, the manufacturing industries more and more pay close attention to artificial intelligence (AI). For example, smart monitoring and diagnosis, real time evaluation and optimization of the whole production and raw materials management can be improved by using machine learning and big data tools. An accurate milling process implies a high quality of the obtained material surface (roughness, flatness). With the involvement of AI-based algorithms, milling process is expected to be more accurate during complex operations.

 In this work, a smart milling diagnosis has been developed for composite sandwich structures based on honey-comb core. The use of such material has grown considerably in recent years, especially in the aeronautic, aerospace, sporting and automotive industries. But the precise milling of such material presents many difficulties.

 The objective of this work is to develop a data-driven industrial surface quality diagnosis for the milling of honey-comb material, by using supervised machine learning methods. Therefore, cutting forces are online measured in order to predict the resulting surface flatness.

 The workpiece material studied in this investigation is Nomex® honeycomb cores with thin cell walls. The Nomex® honeycomb machining presents several defects related to its composite nature (uncut fiber, tearing of the walls), the cutting conditions and to the alveolar geometry of the structure which causes vibration on the different components of the cutting effort.

 All experimental milling tests illustrated in this paper were carried out on a three-axis vertical machining center Realmeca® RV-8. For assessing the performance of the machining process of Nomex® honeycomb core we monitored and measured the cutting forces generated during cutting, by using Kistler dynamometer model 9129AA. Given the low level of cutting forces, the quality of the obtained machined surface allows to establish criteria for determining the machinability of the honeycomb structures.

A "surface response" (by using Neural networks) could have been built in order to predict the milling surface quality. But close milling parameters (such as splindle speed, feed rate, depth of cut) can lead to different results (depending on the material, the quality of the machining tool, etc). Therefore the construction of features associated with machine learning is useful.

Nearly 40 features are calculated in time domain and frequency domain from the raw signal in steady state behavior (transient zones are not taken into account). The features are then normalized. The input parameters for each experiment are: the tool rotation speed, the cutting speed and the depth of cut. It is then necessary to make a dimensional reduction of that feature table in order to avoid overfitting and to reduce the computing time of the learning algorithm.

 In this work, several classification algorithms have been implemented such as : k-nearest neighbor (kNN), Decision trees (DT), Support Vector Machine (SVM), ... .The different supervised learning algorithms have been implemented and compared. Each AI-based model has been applied to a set of features. From the prediction results, SVM algorithm seems to be the most efficient algorithm in this application.

Ps : we presented the preliminary results at the IEEE 23rd International Conference on Emerging Technologies and Factory Automation (ETFA), Torino, september 2018