基于鲁棒估计的动态数据校正方法

针对目前动态数据校正方法存在的缺陷，本文在以前的工作基础上把基于鲁棒估计原理构造的鲁棒估计函数用于含异常点类型的过失误差的数据校正，Monte Carlo模拟结果及对TE问题的校正计算结果表明，这种基于鲁棒估计的过换误差侦破和数据协调同步方法可以在得到协调数据的同时很准确的侦破和识别出测量数据中所含的过失误差，具有较强的优越性。

An approach of dynamic data rectification based on the robust estimator principle

Because industrial process is often under changes, it is necessary to rectify the measurements of the dynamic process. The detection and identification of gross error in dynamic process data is the bottleneck of dynamic data rectification. Present dynamic data rectification meth-ods have limitations. Based on previous research work, the constructed robust estimator function is used for dynamic rectification with abnormal measurements. By using robust method, the object function of rectification problem is insensitive to abnormal data in measurements, so the abnormal data almost have no effect on the data rectification. A CSTR example and TE challenge problem are used for testing the approach. The calculation results and Monte Carlo (MC) simulation results show that this robust estimator method can detect the gross error exactly and obtain good reconciled results simultaneously. The comparison with non-linear programming method shows that this approach is more efficient.