DUAL-MODULAR ANOMALY DETECTION ALGORITHM FOR INDUSTRIAL PREDICTIVE MAINTENANCE: Validation on NASA IMS Bearing Dataset

This technical note presents the validation of a novel "Dual-Modular Filter" algorithm designed for the Early Warning and Anomaly Detection of mechanical faults in industrial rotating machinery. Unlike traditional amplitude-based methods (RMS, FFT, Kurtosis), this approach utilizes a "Numerical Trajectory Analysis" method based on specific modular arithmetic constraints. The algorithm was stress-tested against the NASA IMS (Intelligent Maintenance Systems) Bearing Dataset (2nd Test). The results demonstrate a 0% False Positive Rate during the healthy operation phase and successful Early Warning detection (Critical Alarm) starting approximately 24 hours prior to catastrophic failure. KEY FINDINGS:1. Noise Immunity: The algorithm maintained a 0.00% violation rate during the healthy operation period (Feb 12–17), proving high robustness against industrial noise.2. Early Detection: A critical alarm was triggered on Feb 18 at 06:22, providing a 24-hour lead time before the bearing collapse on Feb 19.3. Computational Efficiency: The method relies on bitwise operations (O(1) complexity), making it suitable for low-power Edge AI and IIoT embedded systems. LICENSE WARNING:This work is licensed under Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International (CC BY-NC-ND 4.0).You are free to share (copy and redistribute the material in any medium or format) under the following terms:- Attribution: You must give appropriate credit to the author (Ender UYGUN).- NonCommercial: You may not use the material for commercial purposes.- NoDerivatives: If you remix, transform, or build upon the material, you may not distribute the modified material. For commercial licensing, integration rights, or technology transfer inquiries, please contact the author directly.