Moving Horizon Estimation For a Nonlinear Polyethylene Reactor

Abstract In a closed-loop control system, measurements from sensors are used as feedback to improve the system’s performance. However, these measurements can be affected by unknown disturbances or noise. Therefore, a state estimator plays a crucial role in the closed-loop system. Borealis, a polyolefins production company, currently uses a Single-Input Single-Output (SISO) estimation algorithm that updates the actual state, x, and the correction factor, e, which is an extended state included in the system model, separately. However, since some measurements are dependent on each other, it is more appropriate to use a multivariable estimation technique, such as Moving Horizon Estimation (MHE). Thus, the MHE algorithm was implemented to evaluate whether it can perform better than SISO. Two systems were used to assess the performance of the estimators. The first is a simple system called the Double Tank System, and the second is a real polyethylene production system, the Polyethylene Reactor System, which is much more complex. The performance of MHE is clearly better than SISO for the Double Tank System. In the Polyethylene Reactor System, MHE performs at least as well as SISO but does not show a clearly better performance. Therefore, it can be concluded that MHE can match the performance of SISO, but it is not recommended to replace the SISO algorithm due to the significantly longer computation time required by MHE.

Department dc:contributor.department

Chalmers tekniska högskola / Institutionen för elektroteknik

Year dc:date.issued

2025

Authors dc:creator

• Thomasson, Oskar
• Khamkhachorn, Pummarin

Advisor dc:contributor.supervisor

• Skålén, Staffan

• Keywords: Model Predictive Control, Moving Horizon Estimation, Sequential Quadratic Programming, Single-Input-Single-Output.

Language dc:language.iso

eng

Handle dc:identifier.uri

https://hdl.handle.net/20.500.12380/309663

OAI record oai:identifier

https://odr.chalmers.se/oai/request?verb=GetRecord&metadataPrefix=dim&identifier=oai%3Aodr.chalmers.se%3A20.500.12380%2F309663