基于独立元分析的数据重构方法及应用

在工业系统采集数据的过程中，因为种种原因会发生数据遗失的现象。为了更好的对工业过程进行分析评估、优化及监控，往往需要重构遗失的数据。主元分析法（PCA）常用于重构遗失数据，但是由于PCA要求观测数据服从正态分布，而实际工业系统获得的数据往往很难满足条件。因此提出一种基于独立元分析（ICA）的数据重构方法。首先使用在正常运行情况下获得的原始数据建立ICA模型，然后利用相关的监控统计量规则来重构遗失的数据，最后通过在TE过程上的仿真应用，验证了该方法的可行性及与PCA相比较的优越性。

Data Reconstruction and Application Based on Independent Component Analysis

During the sampling in industrial system, there are various reasons resulting in the cases of data missing. In order to evaluate, optimize and monitor the industrial process more effectively, missing data usually needs to be reconstructed. Principal component analysis (PCA) is usually used for reconstructing missing data, but it requires that measured data must be subject to normal probability distribution, which sometimes can not be satisfied in actual industrial system. A method of data reconstruction was introduced based on independent component analysis (ICA). ICA model was built using the measured data under normal operation condition (NOC), and then missing data was reconstructed by the correlative monitoring indices, and the simulation on TE process was carried out to prove the feasibility of the method and the advantage comparing with PCA.