点击次数：

论文类型：期刊论文

发表时间：2012-06-01

发表刊物：INTERNATIONAL JOURNAL OF INNOVATIVE COMPUTING INFORMATION AND CONTROL

收录刊物：SCIE、EI、Scopus

卷号：8

期号：6

页面范围：4111-4122

ISSN号：1349-4198

关键字：Data reconciliation; Support vector regression; Parameter estimation; Gross error detection

摘要：Process data measurements are important for process monitoring, control and optimization. However, process data may be deteriorated by gross errors in measurements. Therefore, it is significant to detect and estimate gross errors with data reconciliation. Meanwhile, in any modern petrochemical plant, the plant-wide mass data derived from process data rendering the real conditions of manufacturing are the key to the operation managements such as production planning, production scheduling and performance analysis. In this paper, an extended support vector regression approach for data reconciliation and gross error detection is proposed and applied to deal with the plant-wide mass balance problem. The proposed approach could simultaneously detect and estimate gross errors like measurement bias and process leaks. Then the proposed approach is applied to address the plant-wide mass balance problem with measurement bias and mass movement information lost, because of its superior characteristic for the issue. Both simulation and application results in this paper demonstrate that the proposed approach is accurate and effective to address plant-wide mass balance.