Machine Tool Wear Classifier

1. This is a Machine Learning Program to classify Machine Tool Wear using Time series Imaging by Gramian Angular Summation Fields (GASF) and Deep Learning.
2. Vibration signals are converted to GASF images.
3. A Deep Learning Model is trained to predict tool wear states using GASF images.

Algorithm

State Description

A tool degrades(wears out) in its lifetime as per given graph. The graph can be divied into 4 zones or states. They are:

1. Initial Wear
2. Steady Wear
3. Severe Wear
4. Failure Wear

Vibration signals can be used to classify these states.

Reference

1. Tool wear classification using time series imagingand deep learning.by Giovanna Martínez-Arellano, German Terrazas &Svetan Ratchev,The International Journal of Advanced ManufacturingTechnology.

2. Dataset : PHM 2010 Data Challenge