无参考值情况下自诊断传感器设计方法研究

Research on design method of self-diagnosis sensor without reference value

中文关键词  传递熵  自诊断传感器  冗余度  贝叶斯推理

英文关键词  transfer entropy  self-diagnosis sensor  redundancy  bayesian inference

中文摘要

      对于工业现场中传感器进行故障诊断时难以获取参考值的问题, 本文提出了一种基于贝叶斯推理的无参考值情况下自诊断传感器设计方法. 首先, 运用传递熵方法对传感器之间存在的冗余关系进行量化评价, 并借助β–二项分布建立传感器的故障统计模型; 其次, 利用贝叶斯推理对有参考和无参考时的传感器测量系统设计故障自诊断算法, 对于无参考值情况下的传感器测量系统, 通过引入观测值, 进行自学习获取传感器的健康状况; 最后, 以镍闪速炉中的温度测量系统为例, 验证了文中方法的可行性与有效性.

英文摘要

      Aiming at the problem that it is difficult to obtain the reference value of sensors in fault diagnosis in industrial field, a design method of self-diagnosis sensors without reference value based on the Bayesian inference is proposed. Firstly, the transfer entropy method is used to quantitatively evaluate the redundant relationship between sensors, and the sensor fault statistical model is established with the help of β-binomial distribution. Secondly, the Bayesian inference is used to design the fault self-diagnosis algorithm for the sensor measurement system with and without reference value. For the sensor measurement system without reference value, the output classification index is introduced for self-learning to obtain the health status of the sensor. Finally, taking the temperature measurement system in the nickel flash furnace as an example, the feasibility and effectiveness of the proposed method are verified.