Application Case of the Intelligent Control System for the Atmospheric–Vacuum Distillation Unit at Zhenhai Refining & Chemical Co., Ltd.
Category:
Typical Cases
Figure 1: VOA Online Predictive Control Process The advanced control system for Unit 1 of the atmospheric–vacuum distillation unit, jointly developed by Zhejiang Bangye Technology Co., Ltd. and Zhenhai Refining & Chemical, incorporates Virtual Online Analyzers (VOAs) to predict the quality of products from the fractionation tower. VOAs provide product‑quality estimates at a much higher frequency than laboratory analyses and can be integrated into a Model Predictive Controller (MPC) for feedback control. The VOAs can be calibrated using trusted laboratory test samples to mitigate model bias and the effects of unmeasurable disturbances. Figure 1 illustrates the process architecture for implementing online predictive control with VOAs.
Figure 2: Prediction of Dry Point Quality in the Atmospheric–Vacuum Distillation Unit. The quality‑soft‑sensor system for the Zhenhai Refining & Chemical atmospheric–vacuum distillation unit includes soft sensors for the overhead dry point, the first‑cut dry point, the third‑cut 95% boiling point, and the vacuum‑first‑cut 95% boiling point. Based on the predicted trends of these quality soft sensors during oil‑change operations, the controller initiates automated adjustments in advance, enabling automatic crude‑oil switching while maintaining product quality. The application of this technology has played a crucial role in achieving precise quality control, ensuring stable unit operation, and reducing operator workload. Below, Figure 2 compares the laboratory‑determined overhead dry‑point values with the online VOA‑corrected dry‑point predictions.
It can be observed that the predicted values prior to VOA correction closely follow the trend of the actual analytical samples. When significant changes occur—such as during oil product switching—the system is able to provide timely predictions, which is a prerequisite for MPC to enable online control of product quality.
Figure 3 shows the control performance of the dry‑point index switching for the atmospheric–vacuum distillation unit, as well as the performance of the advanced control system. After the plant adjusted the dry‑point specification—raising the overhead product quality from approximately 170°C to 182°C—the dry‑point controller automatically reconfigured its manipulated variables in response to the changed control range. As observed, the controller’s manipulated variables were progressively tuned, and the predicted dry‑point quality returned to the newly set range. Laboratory analysis confirmed that the actual dry‑point quality also fell within the specified limits, demonstrating that the controller can achieve automatic adjustment of the dry‑point index.
Figure 4 shows the control performance of the advanced control system for the overhead dry point VOA during crude oil switching. By leveraging feedback from the disturbance variable (product‑type change), the predicted value of the overhead dry point is adjusted. When the product‑type transition is significant, the predicted dry point may temporarily exceed the control limits; at this point, the controller responds promptly and automatically retunes the manipulated variables. As a result, the predicted VOA quickly returns within the control range, demonstrating that, throughout the crude oil switching process, the quality of the overhead product remains essentially within the setpoint, enabling automated switching.
Address：19F, Building S3, HeRui Science and Technology Park, No.475, Changhe Road, BInjiang District, Hangzhou, Zhejiang, China