加权k最近邻重构分析的工业过程故障诊断

k--最近邻(k--nearest neighbor, k--NN)是一种有效的基于数据驱动的故障检测方法, 该方法在工业过程监视方面已经得到了广泛的应用. 但在过程中存在故障时, 精确地寻找故障根源和识别故障变量是故障诊断的重要目标, 也是保证工业过程安全生产的重要任务. 本文在k--NN故障检测技术的基础上, 提出了一种加权的k--NN重构方法, 对使控制指标减小最大(maximize reduce index, MRI)的过程变量依次进行重构, 进而确定发生故障的传感器. 根据理论分析并结合数值仿真对提出的方法进行了验证, 数值仿真先从精度方面验证了该方法能够有效地对故障传感器数值进行重构, 然后验证了该方法不仅适用于单一传感器 故障诊断, 对于同时发生或者因变量相关性而传播的传感器故障也具有很好的效果. 最后, 该方法被成功应用于TE(Tennessee Eastman)化工过程.

Fault diagnosis of industrial processes based on weighted k-nearest neighbor reconstruction analysis

The k-nearest-neighbor (k--NN) is an effective fault detection method based on data driven, which has been widely used in industrial process monitoring. However, investigating the root causes of abnormal events is a crucial task when the process faults have been detected, and isolating the faulty variables provides additional information for investigating the root causes of the faults. In this paper, a novel fault diagnosis method is derived using weighted k--NN reconstruction on maximize reduce index (MRI) variables, it can effectively identify the faulty variables and locate the faulty sensors. A numerical simulation is provided to validate the performance of weighted k--NN in the aspect of data reconstruction; it also show that this method is suitable not only for a single sensor fault, but also with good results for multiple sensor faults which are existing simultaneously or in propagation through variable correlation. Finally, this method is applied to TE chemical process successfully.