加权动态SVDD在非线性过程监测中的应用研究

Application of Weighted Dynamic SVDD in Nonlinear Process Monitoring

由于化工过程的复杂性,数据往往存在动态以及序列之间具有相关性特点,传统的支持向量数据描述(Support Vector Data Description,SVDD)方法,很难保证故障监测的准确性和实时性,提出一种基于加权的动态SVDD(WDSVDD)在线实时故障监测方法,引入动态方法,考虑了数据之间的序列相关性,利用加权的方法把有用的信息突出显示,利用SVDD方法建立模型,实现了在线实时故障监测。该方法不仅克服了过程数据非高斯、非线性特性对故障监测带来的影响,并且考虑了数据的动态特性和序列之间的关系,通过在数值仿真和TE过程实例中的应用验证了方法的有效性。

Due to the complexity of the chemical process, the data are often characterized by dynamics and correlation between sequences. Traditional support vector data description（SVDD） methods are difficult to guarantee real-time monitoring online. A Weighted-Dynamic-SVDD（WDSVDD） method was proposed to monitor fault in real time online. The dynamic method was introduced, and the correlation between the data was considered. The weighted information was used to highlight the useful information. The model was established by using SVDD method, and the online real-time fault monitoring was realized. The method not only overcomes the adverse effect of non-Gaussian and nonlinearity, but also considers dynamic characteristics and correlation between sequences of the processing data. Applications in the numerical simulation and TE process instance verify the effectiveness of the proposed method.