Abstract: In response to the problems such as the increase of operating load, the aggravation of product quality fluctuations and high energy consumption caused by the variety of crude oil, based on the principle of distillation, material balance, energy balance and big data, the on-line twinning model of distillation columns and the real-time perception technology of crude oil feed property were developed. The characteristic of this technology was to calculate the properties of crude oil according to the data of temperature, pressure and flow of crude oil distillation column, without the need for the installation of the real-time raw material analysis,nor the input of crude oil database and tank analysis data. Using scientific verification methods and processes, the veracity of the raw material property prediction model and twin process model was verified by using the production data of atmospheric and vacuum distillation units.An intelligent optimization control system for atmospheric column was developed,and the reliability of the system was verified by production data. By optimizing the system, the impact of crude oil switching on product properties was reduced, the naphtha yield was increased by 0.9% at the top of the atmospheric column, the yield of crude oil was increased by 0.8% atthe second line, and the energy consumption was reduced by 1.43%.

Key words: intelligent perception, crude oil distillation, online twin model, intelligent control, energy consumption optimization