Predicting and managing device status based on time series analysis

Abstract

Modern high-performance computing systems require high reliability and stability, as equipment failures can lead to significant losses of resources and time. To address this issue, this paper proposes a method for intelligent anomaly detection and equipment condition prediction as part of an automated supercomputer control system. The approach is based on the analysis of time series of operating parameters and the application of machine learning algorithms to identify hidden signs of degradation and deviations from the optimal operating mode. The system architecture provides for the integration of modules for collecting data from sensors, intelligent signal processing, and real-time control decisions. Implementation of this method will enable a transition from reactive maintenance to proactive management, increase the reliability of computing complexes, and ensure more efficient use of their resources. The developed approach can be adapted for use in other critical technical systems that require predictive maintenance of equipment. doi 10.54708/19926502_2026_3021123