Data Acquisition and Communication in CNC Machine Tools for PdM

Resumo

In addition to the traditional requirements of quality and productivity, new production paradigms also demand fault prognosis and diagnosis capabilities that classical maintenance methods do not provide. In this scenario, the application of predictive maintenance emerges as a viable alternative. Predictive maintenance is a technique that uses monitoring systems to analyze the condition of assets, based on data collected from sensors, to prevent or avoid failures. By estimating the future condition of machine components, maintenance costs can be reduced, operating time can be increased, and maintenance tasks can be optimized. Despite these advantages, the technique is not yet widely adopted within the manufacturing sector because it requires sensing and data processing resources that make its implementation costly. To facilitate the implementation of this technique in the manufacturing sector, a low-cost hardware for data acquisition was proposed. Based on the system's components, an architecture for communicating the collected data and an application model were proposed, which allow for the visualization of this data and serve as a basis for implementing predictive algorithms for fault detection and anomaly prediction. After implementation on a demonstration machine, the system is expected to be capable of collecting machine data in real-time, enabling the identification of faults and anomalies, thereby reducing downtime and costs. This adds benefits to various organizational levels, such as at the operational level, through the reduction of unplanned stops and increased safety. In the maintenance and process sector, it facilitates a transition from preventive to proactive methods. In terms of strategic results, it leads to the reduction of operational costs and an increase in the reliability and predictability of production.