An on-line soft-sensor for control and optimization of crude distillation column

Soft-sensing techniques are widely applied in such practical situations where some process variables involved in control and optimization systems can not be measured directly or reliably due to the technical limitations or economic expense. This paper proposes an on-line soft-sensor for the quality control and energy optimization of crude distillation column. Based on the system analysis, a two-stage soft-sensor has been developed for the estimation of internal liquid and vapor loads, and product quality specifications of sidestreams. The on-line correction mechanism is also developed to eliminate the estimated bias from the off-line laboratory data. The measurements of the soft-sensor are used and to derive the constraints of the energy optimization system as well as to feedback to the quality control system.