Tool wear classification using time series imaging and deep learning

Abstract

Tool condition monitoring (TCM) has become essential to achieve high-quality machining as well as cost-effective production. Identification of the cutting tool state during machining before it reaches its failure stage is critical. This paper presents a novel big data approach for tool wear classification based on signal imaging and deep learning. By combining these two techniques, the approach is able to work with the raw data directly, avoiding the use of statistical pre-processing or filter methods. This aspect is fundamental when dealing with large amounts of data that hold complex evolving features. The imaging process serves as an encoding procedure of the sensor data, meaning that the original time series can be re-created from the image without loss of information. By using an off-the-shelf deep learning implementation, the manual selection of features is avoided, thus making this novel approach more general and suitable when dealing with large datasets. The experimental results have revealed that deep learning is able to identify intrinsic features of sensory raw data, achieving in some cases a classification accuracy above 90%.

Lagged encoding for image‐based time series classification using convolutional neural networks

“…Time series classification is a thriving area of study. Existing algorithms find applications in computer‐aided decision‐making systems, online monitoring in areas such as human activity recognition , automation and control , remote sensing , manufacturing , astronomy , and many other areas of science and engineering.…”

Tool Life Prediction of Ti [C,N] Mixed Alumina Ceramic Cutting Tool Using Gradient Descent Algorithm on Machining Martensitic Stainless Steel

“…In addition to machine learning algorithms, signal and image processing were also used to predict toll wear. Giovanna Mart ínez et.al used signal imaging to encode the images of the tool at specified timesteps and fed to a pre-made deep learning package for classifying the tool wear as break-in wear, steady wear, severe wear and failure region [12]. Bovic Kilundua et.al measured vibration signals on the toolholder and pseudo-local singular spectrum analysis is done to extract the features that are essential for the quality of the tool and is monitored continuously [13].…”

A CPPS based on GBDT for predicting failure events in milling

“…The data-driven methods are particularly interesting for CPPS which use data as a non-physical medium to communicate information carrying production process status to core control computers. Numerous data-driven methods have been developed since the first introduction of the concept in 1930 [15], generally classified into three mainstreams including the two multivariate statistical approaches: principal component analysis (PCA) [16,17] and partial least square (PLS) [18,19] and the third machine learning approach using algorithms such as neural networks [17,20], [21], support vector machine [22,23] and tree-based models [24]. Both PCA and PLS are able to reserve the most significant variety of the original set of parameters, which are assumed to follow Gaussian distributions.…”

“…The two techniques can also be applied simultaneously in association with the PCA to handle systems which possess both of the common yet troublesome properties [24]. As for the machine learning algorithms, the support vector machine [22,23] and neural networks [17,20], [21] have been intensively explored solely or in combination with previously mentioned multivariate statistical models to resolve the nonlinearity and dynamics between parameters those do not follow Gaussian distributions. A tree-based model with the sliding window technique has been proposed in [32], which has been shown to provide an accurate prediction of production failures with less computational complexity compared with the above two algorithms.…”