Contents lists available at ScienceDirect Journal of Manufacturing Processes journal homepage: www.elsevier.com/locate/manpro Monitoring tip-based nanomachining process by time series analysis using support vector machine Fei Cheng a , Jingyan Dong b, ⁎ a Management School, Hangzhou Dianzi University, Hangzhou, China b Edward P Fitts Department of Industrial and System Engineering, North Carolina State University, Raleigh, NC, USA ARTICLE INFO Keywords: AFM tip-based nanomachining Process monitoring Tip wear detection Time series data Support vector machine ABSTRACT In this paper, time-series data analysis and pattern recognition using a multi-class support vector machine (SVM) were studied to monitor the state changes of the AFM tip-based nanomachining process with respect to the machining performance and tip wear. Time series data (i.e. machining force from the process), which has transient, nonlinear, and non-stationary characteristics, was collected by a data acquisition system. Three status detection features including the maximum force, peak-to-peak force value, and the variance of the collected lateral machining force, were extracted to classify the state of the nanomachining process. Directed Acyclic Graph Support Vector Machines (DAGSVM) with a Gaussian Radial Basis Kernel Function (RBF Kernel) was constructed to identify the different process states. Using this multi-class SVM, the machining process and the tip wear can be classified into three regions, which are effective machining with a sharp tip, transition region and bad/no machining with severe tip wear. The experimental data showed that the accuracy of the SVM was over 94.73% in both binary and ternary classifications, which confirmed that the SVM-based pattern recognition technology via time series data could successfully monitor the tip wear and process performance for tip-based nanomachining process. 1. Introduction The tip-based vibration-assisted nanomachining process is a pow- erful tool for the fabrication of nanometer scale patterns and structures. In the nanomachining process, the AFM probe is used to not only image the sample, but also acts as the tool to remove materials mechanically to produce nanoscale structures. The state of the AFM tip (such as its sharpness) directly affects the quality of the machined nano-structures and the scanned surface images. For example, a worn tip will result in blurred images with reduced resolution when used for imaging. On the other hand, a blunt tip with severe wear cannot effectively remove materials during machining process, due to the inadequate stress from the larger top radius. Therefore, the real-time detection and monitoring of the nanomachining process and the tip states, including tip wear diagnosis and tip life prediction, has become one of the critical issues of the nanomachining processes. Wear diagnosis, which relies on the analysis of feature factors, processing parameters, and pattern quality, is valuable to predict the tip's remaining use time or distance based on the level of tool wear. One method of tip wear detections is to measure the tip geometry directly by using the scanning electron microscope (SEM). Skarman et al. applied a high-resolution SEM to find the geometry of tips used in tapping mode [1]. The image deterioration from the AFM was corre- lated with factors like tip cracking, tip contamination and tip supply vendors. Ho et al. reported that if an AFM probe contacts the surface by oscillation in each cycle, severe probe wear or damage was observed. If the probes were oscillated in a “near-contact” tapping mode, image quality can be well preserved for longer scanning distances [2]. The direct SEM observation can visually reveal the degree of tool wear, but is incapable for continuous online identification. Therefore, a few model analysis methods such as molecular dynamics (MD) and finite element method (FEM) had been developed. Gotsmann et al. applied a method to continuously monitor the tip radius during machining pro- cess based on the pull-off force between the AFM tip and sample surface [3]. Bernal et al. performed in-situ transmission electron microscopy (TEM) and molecular dynamics (MD) studies of nanoscale single-aspe- rities made of tetrahedral amorphous carbon contacting single-crystal diamond, to correlate the asperity's geometry and the adhesion mea- surements, and analyzed the scatter characteristics of pull-off force [4]. In addition, a few researchers also applied MD method to study tool https://doi.org/10.1016/j.jmapro.2019.01.011 Received 28 September 2018; Received in revised form 22 December 2018; Accepted 5 January 2019 ⁎ Corresponding author at: Department of Industrial and Systems Engineering, North Carolina State University, 414-C Daniels Hall, Campus box 7906, Raleigh, North Carolina, 27695-7906, USA. E-mail address: jdong@ncsu.edu (J. Dong). Journal of Manufacturing Processes 38 (2019) 158–166 1526-6125/ © 2019 Published by Elsevier Ltd on behalf of The Society of Manufacturing Engineers. T