An acoustic dataset for surface roughness estimation in milling processMendeley Data

Abstract

Machining process involves numerous variables that can influence the desired outcomes, with surface roughness being a critical quality index for machined products. Surface roughness is often a technical requirement for mechanical products as it can lead to chatter and impact the functional performance of parts, especially those in contact with other materials. Therefore, predicting surface roughness is essential. This dataset comprises 7444 audio files containing acoustic signal samples recorded using a 44.1 kHz microphone during the milling of mild steel with a tungsten carbide tool on a BFW YF1 vertical milling machine. Various combinations of speed, feed and depth of cut were used, and surface roughness values measured using a Carl Zeiss E-35B profile-meter are provided for each combination. Additionally, an example workflow indicating the possible use of the data to estimate the surface roughness from the acoustic signals is presented. This dataset is the first publicly available resource for surface roughness measurement using sound signals in milling, offering significant potential for reuse in related research and applications.

Keywords

• Condition monitoring
• Machine learning
• Machining
• Acoustic
• Milling