定比值系统传感器单一故障诊断与容错控制  

Single Sensor Fault Diagnosis and Fault-Tolerant Control for Constant Ratio System

在线阅读下载全文

作  者:那文波 王伽豪 高雁凤 刘志威 昝琪 王铮 NA Wenbo;WANG Jiahao;GAO Yanfeng;LIU Zhiwei;ZAN Qi;WANG Zheng(College of Mechanical and Electrical Engineering,China Jiliang University,Hangzhou Zhejiang 310018,China)

机构地区:[1]中国计量大学机电工程学院,浙江杭州310018

出  处:《传感技术学报》2025年第2期256-262,共7页Chinese Journal of Sensors and Actuators

基  金:浙江省自然科学基金委公益技术应用分析测试项目(LGC21F030001)。

摘  要:针对定比值系统传感器故障信息难以提取和故障重构实时性问题,提出了基于时空特征的传感器故障诊断与容错控制方法。该方法首先通过滑动窗口思想预处理离线数据,使其在固定窗口内呈现阶段性特征;其次通过融合自注意力机制和长短期记忆网络(LSTM)分别构建故障定性与定位网络、故障估计网络,以此获取传感器的故障类型、强度、位置和时间信息;最后基于故障补偿思想搭建容错控制模型,实现定比值系统传感器故障信号在线重构。实验结果表明,所提方法的诊断准确率达到97.53%,相较于传统的CNN-LSTM方法准确率提高了6.65%,验证了所提方法的有效性和准确性。Targeting at the problems of difficult extraction of sensor fault information and real-time fault reconstruction in constant ratio systems,a sensor fault diagnosis and fault-tolerant control method based on spatiotemporal characteristics is proposed.Firstly,this meth-od preprocesses offline data by using the sliding window idea to present periodic features within a fixed window.Secondly,a fault quali-tative and location network and a fault estimation network are constructed by fusing the self-attention mechanism and the long and short term memory network(LSTM),respectively,to obtain the sensor fault type,intensity,location,and time information.Finally,a fault-tol-erant control model is built based on the idea of fault compensation to achieve online reconstruction of sensor fault signals in a constant ratio system.The experimental results show that the diagnostic accuracy of the proposed method reaches 97.53%,which is 6.65%higher than the traditional CNN-LSTM method,verifying the effectiveness and accuracy of the proposed method.

关 键 词:故障诊断与容错控制 时空特征 自注意力机制 长短期记忆网络 

分 类 号:TP277[自动化与计算机技术—检测技术与自动化装置] TP302.8[自动化与计算机技术—控制科学与工程]

 

参考文献:

正在载入数据...

 

二级参考文献:

正在载入数据...

 

耦合文献:

正在载入数据...

 

引证文献:

正在载入数据...

 

二级引证文献:

正在载入数据...

 

同被引文献:

正在载入数据...

 

相关期刊文献:

正在载入数据...

相关的主题
相关的作者对象
相关的机构对象