# Multi-Sensor monitoring dataset for milling process with varied parameters and materials

## TL;DR

This paper introduces a dataset for milling processes using multiple sensors to monitor parameters and materials, aiming to improve machining efficiency and quality.

## Contribution

The novel contribution is the creation of a high-quality multi-sensor dataset for milling with varied parameters and materials.

## Key findings

- The dataset includes 115 samples with variables like cutting depth, feed rate, and workpiece materials.
- 15 slot milling samples help calibrate mechanical milling force coefficients for physics-informed machine learning.

## Abstract

Real-time monitoring of milling parameters is essential to improve machining efficiency and quality, especially for the workpieces with complex geometry. Its main task is to build the relationship between the parameters and the monitoring data. As the relationship is challenging to be established solely through mechanism-driven or data-driven methods, the physics informed method, based on prior physical laws between physical signals and milling parameters, becomes the optimal method. However, this method is limited due to the lack of a high-quality dataset. Therefore, a multi-sensor monitoring dataset for the milling process with various milling parameters and milling materials is built. The variables include cutting depth, cutting width, feed rate, spindle speed and workpiece materials (aluminium alloy 7030 and CK45 steel). The multi-sensor includes force, vibration, noise, and current. A dataset comprising 115 samples is built, including 100 samples collected using the 'all factors' method, and 15 slot milling samples using two different workpiece materials. The 15 slot milling samples are used to calibrate mechanical milling force coefficients, which is beneficial for developing a physics-informed machine learning algorithm.