摘要
提出了一种基于深度学习的工业自动化控制DCS系统维护方法,以提升系统维护的智能化水平与效率。采用长短期记忆网络进行故障预测,利用卷积神经网络结合注意力机制进行故障诊断,并通过深度强化学习优化维护策略。结果表明,该方法在设备故障预测准确率、诊断精度及维护成本优化方面均有显著提升。
This paper proposes a deep learning based maintenance method for industrial automation control DCS systems to enhance the intelligence and efficiency of system maintenance.Using long short-term memory networks for fault prediction,combining convolutional neural networks with attention mechanisms for fault diagnosis,and optimizing maintenance strategies through deep reinforcement learning.The results show that this method significantly improves the accuracy of equipment fault prediction,diagnostic precision,and maintenance cost optimization.
作者
章海峰
高瞳
ZHANG Haifeng;GAO Tong(Zhejiang Zhongkong Science and Education Equipment Co.,Ltd.,Hangzhou,Zhejiang 310051,China;Beijing Hollysys Intelligent Technology Co.,Ltd.,Beijing 100000,China)
出处
《自动化应用》
2025年第6期12-14,共3页
Automation Application
关键词
深度学习
分布式控制系统
故障预测
deep learning
distributed control systems
fault prediction
