摘要详情

ID / 提交时间

212 / 2025-06-14 14:48:03

标题

A Multidimensional Data-Driven Algorithm for Operational Failure Prediction

关键字

Fault prediction,remote operation and maintenance,joint knowledge and data-driven

主题及专题

4、System Monitoring, Fault Diagnosis and Prediction

状态

全文待审

作者

晨瑄李 / 北京

摘要

Data resource wastage, low equipment utilization, traffic overloads during bursts, memory leaks, and network device bandwidth saturation constitute major challenges hindering the intelligent development and operation of data centers worldwide. To enhance data resource utilization, detect potential equipment failures, and improve operational efficiency within aerospace flight control systems, this paper proposes a multidimensional data-driven fault prediction algorithm. A prediction model is constructed utilizing ARIMA and ConvLSTM techniques. This approach optimizes feature selection while simultaneously enhancing the algorithm's adaptability to complex data patterns and sudden anomalies. The hybrid spatiotemporal convolutional model effectively accommodates diverse monitoring metrics and demonstrates robust predictive capabilities for highly dynamic resource operation patterns. Experiments were conducted using a multidimensional dataset derived from the equipment. The results demonstrate that the proposed fault prediction method offers flexible adaptability across multiple application scenarios and achieves high overall fault prediction accuracy.