The vertical roller mill used in the cement industry is a crucial piece of equipment that operates continuously and is essential for production. However, frequent breakdowns in its components, particularly the roller ends, have resulted in multiple production stops. To address this issue, the concept of analyzing the mill's sound vibrations to identify potential defect patterns was explored. Sound analysis was conducted during instances of terminal failures. Applying signal processing techniques such as FFT, multiresolution analysis, and SAC-DM (based on chaos theory) enabled the observation of a recurring pattern in the equipment's behavior that preceded the appearance of failures in the vertical mill's roller ends. It was noted that multiresolution analysis and SAC-DM were highly sensitive to these patterns, allowing the company's maintenance department to anticipate necessary repairs. This study employed sound analysis using a smartphone, multiresolution analysis, and density of maxima to detect failure patterns in vertical roller mills. It was observed that SAC-DM proved to be particularly sensitive to defects, enabling timely intervention and maintenance.
