Acoutic emission of CNC cutting process

Description

This dataset contains acoustic emission (AE) signals which were acquired during CNC-based cutting processes by AE sensors. These sensors were integrated into the "OptiMill F80" CNC machine for the purpose of studying the process acoustics and experimenting with data-based approaches for AE-based condition- and process-monitoring. 

Process

A CNC machine was programmed to repeatedly cut along the edge of a steel (1.4305) workpiece. After each cut, the tool was shifted by 4 mm into the workpiece to cut asnother edge. Throughout the process, the tool was alternated between one in good condition and one in poor condition after a certain number of cuts. Additionally, holes were drilled into the workpieces to simulate defects.

Measurement

The acoustic emission were acquired, using two "Fujicera 1045S" acoustic emissions sensors with the "linWave" data acquisition system by Vallen GmbH. One sensor was placed at the housing of the spindle (channel 1), while the other was placed at thebottom side of movable table (channel 2). The linWave was set to sample the acoustics at a sample rate of 1 MHz with sensor 1 set for an input range of 5000 mV and sensor 2 for an input range of 5 V. Digitial filters were not set during the acquisition.

File description

Each measurement consists of two binary files:

• The binary files consist of int16 values, which form an array that represents the raw acoustic noise of the CNC cutting process
• Each binary file belongs to one sensor
• The name of each file includes the date and time when the AE signal was acquired, as well as the channel number, indicating the sensor position
• Each measurement contains 7 or 9 cutting processes
  • First 2 or 3 processes describe cutting steel, using a tool of healthy condition
  • Next 2 or 3 processes describe cutting steel, using a tool of worn-out condition
  • Last 3 process describe cutting steel with holes (as simulated defects), using a tool of healthy condition