Plant-wide mass balance using extended support vector regression based data reconciliation and gross error detection

In any modern petrochemical plant, the plant-wide mass data rendering the real conditions of manufacturing is the key to the operation managements such as production planning, production scheduling and performance analysis. Because of the characteristic of data reconciliation and gross error detection, it is quite suitable to address plant-wide mass balance problem using data reconciliation and gross error detection techniques. In this paper, an extended support vector regression approach for data reconciliation and gross error detection is proposed to achieve plant-wide mass balance, which can simultaneously detect and estimate measurement errors and missing mass movement information. The simulation results demonstrate that the proposed approach is effective and accurate.