模拟电路在线故障检测与诊断的FPGA实现

FPGA Implementation of Online Fault Detection and Diagnosis for Analog Circuits

选择节点电压、幅频特性和相频特性构成故障特征向量集,并根据多通道宽带数据采集接口和驱动程序,给出全数字式的正交幅相特性实时检测方法。利用网络分析与信息融合技术,设计一种模拟电路在线检测和诊断方案,并在现场可编程门阵列硬件系统中进行实现,以解决诸多模拟电路故障诊断算法缺乏测试平台而停留在仿真阶段的问题。实例分析结果表明,该系统可有效识别电路软故障,在存在元件容差的情况下,其识别准确率高达96%,时间消耗低于60 ms。

This paper chooses node voltage, characteristics of amplitude-frequency and phase-frequency as fault feature vector set, which can be captured by the multichannel wide-band data acquisition interface equipped with corresponding driver program. Taking advantages of network analysis and information fusion, it proposes the full-digital and orthogonal approaches for detecting magnitude-phase characteristics in real time. The paper develops and implements a modular analog circuit fault diagnostic hardware architecture based on Field Programmable Gate Array（FPGA）, solving the problem that most diagnosis algorithms of analog circuits are staying in the simulation stage as lacking of related platforms. Case analysis results show that the system can deal with the tolerance fault efficiently, and its recognizing accuracy achieves high up to 96%, while the time consumption is within 60 ms, although there is some overlapping data when tolerance is considered.