基于增强液态神经网络的励磁系统故障检测方法OA
Fault Detection Method for Excitation Systems Based on Enhanced Liquid Neural Networks
励磁系统故障严重威胁电力系统安全稳定运行,而工程上常用的离散时间数据驱动模型在多桥并列励磁系统的早期故障检测方面存在明显局限.此类模型因依赖固定采样序列,难以有效刻画连续电磁过程中的细微演化规律与多测量点间的空间耦合关系.同时,其对噪声干扰和数据缺失较为敏感,因而难以及时识别微弱故障特征及多时间尺度的退化过程.为此,本文首次将液态神经网络引入励磁系统故障检测领域,提出面向多桥并列励磁装置的增强型液态神经网络(E-LNN)模型.该模型通过常微分方程在连续时间上描述神经元状态演化,并引入空间-时间耦合注意力、多尺度可变时间常数和故障敏感稀疏连接三类增强机制,以同时刻画桥间电磁耦合关系、从微秒级瞬态冲击到小时级缓慢劣化的动态特征以及异常模式的差异性.实验结果表明,E-LNN 在涵盖 12 种故障模式的数据集上检测准确率达 94.83%,较次优方法提升 2.69%,在强噪声和严重数据缺失条件下仍保持稳定性能,有效弥补传统离散时间模型在鲁棒性和早期故障捕捉方面的不足,体现出良好的工程应用价值.
Excitation system faults pose a serious threat to the secure and stable operation of power systems,while commonly used discrete-time data-driven models exhibit clear limitations for early fault detection in multi-bridge parallel excitation systems.Such models,relying on fixed samp-ling sequences,struggle to accurately capture the subtle evolution of continuous electromagnetic processes and the spatial coupling among multiple measurement points.Moreover,they are sensitive to noise interference and data gaps,which makes it difficult to promptly identify weak fault features and multi-timescale degradation processes.To address these issues,we,for the first time,introduce liquid neural networks(LNNs)into excitation system fault detection and proposes an enhanced liquid neural network(E-LNN)model tailored for multi-bridge paralleled excitation devices.The model describes neuron state evolution in continuous time via ordinary differential equations and incorporates three enhancement mechanisms spatio-temporal coupling attention,a multi-scale variable time-constant scheme,and fault-sensitive sparse connectivity,to jointly characterize electromagnetic coupling among bridges,dynamic features ranging from microsecond-level transients to hour-level slow degradation,and the distinctiveness of abnormal patterns.Experimental results show that E-LNN achieves a detection accuracy of 94.83%on a dataset covering 12 fault modes,outperforming the second-best method by 2.69%.Under heavy noise and severe missing data,it still maintains stable performance,effectively overcoming the shortcomings of traditional discrete-time models in robustness and early fault capture and demonstrating strong potential for engineering applications.
马明叶;谢燕军;张舸;程建;张敬;赵学健
中国长江电力股份有限公司白鹤滩电厂,四川 宁南 615400国电南瑞科技股份有限公司,江苏 南京 210000中国长江电力股份有限公司白鹤滩电厂,四川 宁南 615400中国长江电力股份有限公司白鹤滩电厂,四川 宁南 615400中国长江电力股份有限公司白鹤滩电厂,四川 宁南 615400南京邮电大学宽带无线通信与传感网技术教育部重点实验室,江苏 南京 210003
信息技术与安全科学
励磁系统故障检测液态神经网络空间-时间注意力多尺度建模
excitation systemfault detectionliquid state neural networkspatio-temporal attentionmultiscale modeling
《信息与控制》 2026 (4)
687-700,14
国家自然科学基金项目(62303239)三峡金沙江云川水电开发有限公司宁南白鹤滩电厂科技项目(Z532302044)
评论