Gross error identification for dynamic system
Gross error identification for dynamic system
复制标题
DOI：
10.1016/j.compchemeng.2004.07.008
复制
发表时间：
2004-12
期刊：
影响因子：
--
通讯作者：
Mingfang Kong;Bingzhen Chen;Xiaorong He;Shanying Hu
中科院分区：
文献类型：
--
作者：
Mingfang Kong;Bingzhen Chen;Xiaorong He;Shanying Hu
数据校正可分为数据校正、粗差检测和粗差识别两部分。粗差辨识是利用测量信息和过程模型，对存在粗差的变量进行定位和估计。粗差检测与识别是数据校正的瓶颈。对于有偏型粗差，现有方法不能有效处理测量值中存在多个粗差的情况。为了解决这一问题，提出了过程粗差可辨识度的概念。推导了粗差的可识别条件。可识别条件可用于确定系统中包含的粗差是否可识别。提出了一种基于参数估计的辨识方法。识别结果表明，该方法能够准确识别测量数据中存在的多个粗差。
Data rectification can be classified into two parts: data reconciliation, gross error detection and identification. The identification of gross error is to locate variables that have gross errors and estimate their values by using measurements information and process model. Gross error detection and identification is the bottleneck of data rectification. For bias type of gross error, existing methods cannot deal with the case effectively that there are multiple gross errors in measurements. In order to address the problem, the concept of gross error identifiability of process is proposed. And the identifiable condition of gross errors is derived. The identifiable condition can be used to determine whether the gross errors contained in the system could be identified. An identification approach based on parameter estimation is proposed. The identification results show that the method can identify multiple gross errors existing in the measurements accurately.