Nonlinear Control of Fouling in Polyethylene Reactors

Area of Science:

• Chemical Engineering
• Polymer Science
• Process Control

Background:

• Fouling in industrial low-density polyethylene (LDPE) reactors compromises operational safety, polymer quality, and productivity.
• Existing control methods may not adequately address the dynamic challenges posed by fouling in polymerization processes.

Purpose of the Study:

• To introduce and evaluate a neural Wiener model predictive control (NWMPC) strategy for mitigating fouling in LDPE reactors.
• To integrate a soft sensor model for online monitoring and activation of fouling-defouling (F-D) mechanisms.

Main Methods:

• Implementation of a neural Wiener model predictive control (NWMPC) system.
• Development and integration of a soft sensor model to detect fouling thresholds.
• Activation of a fouling-defouling (F-D) mechanism triggered by the soft sensor.
• Comparative analysis against a linear state-space model predictive control (SSMPC).

Main Results:

• NWMPC demonstrated fast, stable, and robust performance across various control scenarios.
• The integrated soft sensor and NWMPC system enabled effective online monitoring and control of F-D processes.
• NWMPC achieved quicker grade transitions and reduced resource consumption compared to SSMPC.

Conclusions:

• NWMPC is a highly effective strategy for addressing fouling issues in LDPE polymerization.
• The combination of NWMPC and soft sensing provides a robust solution for maintaining reactor safety and product quality.
• This approach offers significant advantages in terms of control efficiency and resource utilization over traditional methods.