Analyzing Machining Cycle Time Anomalies via CNC and Operational Data

Abstract

As manufacturers are transitioning to Industry 4.0, real-time monitoring of manufacturing processes has become more prevalent. This paper presents an intuitive anomaly analysis tool designed to support data-driven decision-making for multiple Computer Numerical Control (CNC) machines in a production environment. By leveraging internal Programmable Logic Controller (PLC) signals enriched with tool information from G-code, and operational scheduling data, the proposed tool identifies anomalies in machining process durations without relying on theoretical cycle time estimates from G-code or CNC interpolation strategies. Instead, it computes actual program duration distributions after applying rigorous filtering to address operational disruptions such as server issues, maintenance, or shift changes. Results reveal a near-linear correlation between these computed statistics and theoretical machining durations predicted by a machining simulator, validating the pipeline’s reliability. The developed decision-support application integrates cycle time statistics into comprehensive tables and visualizations, allowing users to analyse data at program, tooling, and G-code levels. This approach facilitates the detection of specific anomalies or bottlenecks in the process, including significant cycle time variance in tools or G-code sections, tool sequencing issues, recurrent anomalies in subprograms or G-code ranges linked with axes and table movements, unoptimized machining parameters or other causes. By focusing on robust data preparation, this work highlights the critical role of preprocessing in achieving actionable insights for production optimization.