Solutions Polymer
Cybernetica has extensive experience in developing, implementing and maintaining systems for Model Predictive Control (MPC) and optimisation of continuous, batch and semi-batch polymerisation processes worldwide.
Typical benefits
- Increased productivity
- Consistent quality
- Safe operation
- Reduced carbon footprint
- Increased utilisation of available measurements
- Higher degree of automation
Our employees have long experience with industrial implementation for a wide range of polymer processes.
Talk to one of our colleagues and challenge us on how we can contribute to better control of your polymer process (Contact us).
MPC based on adaptive mechanistic models
Polymerisation processes are often challenging to control. The reactions might be highly exothermal, and the dynamic response from manipulated to controlled variables are highly dependent on the grade being manufactured in a continuous process or on the current stage in a batch or semi-batch process. In addition, it is often impractical to measure important variables subject to control such as reaction rates and quality variables.
Model Predictive Control (MPC) based on mechanistic models is a versatile and powerful technology for control and optimisation of polymerisation processes. Models will typically include reaction kinetics, phase equilibria and thermodynamics, and quality variables, as well as cooling and other utilities.
Models of polymerisation processes are developed specifically for use in MPC applications, and uncertain model parameters are estimated online in such applications. This combination of mechanistic models with model adaptation are crucial elements in the successful application of MPC technology to polymerisation processes.
Unique NMPC technology for polymerisation processes
Cybernetica delivers systems for Nonlinear Model Predictive Control (NMPC) and optimisation of polymerisation processes based on our software suite Cybernetica CENIT. CENIT is designed for flexible implementation and our applications are easily adaptable to specific customer needs. It is more robust and has higher performance than alternative technologies. Our approach based on mechanistic models minimises impact on production during development, testing and commissioning of NMPC applications.
NMPC for semi-batch polymerisation processes
Nonlinear Model Predictive Control (NMPC) based on Cybernetica Cenit is particularly suited for control of batch and semi-batch processes due to the nonlinear process characteristics.
Typical objectives in batch and semi-batch applications are:
- Accurate and stable temperature control
- Predict and counteract rapid variations in reaction heat
- Offset free tracking of non-isothermal temperature reference trajectory
- Fast response to temperature setpoint changes
- Counteraction of known process disturbances (e.g. dosing of raw materials)
- Optimal metering/dosing of reactants and catalysts
- Minimise polymerisation time
- Control dosing with respect to product quality
- Comply with operational constraints
- Cooling capacity constraints
- Temperature constraints
- Pressure constraints
- Quality constraints
- Safety related constraints
- Control final product quality
- Comply with terminal constraints on product quality
For many polymerisation processes, the final product quality is determined by the entire batch trajectory. By predicting the entire batch, NMPC can be used to control final product quality by applying terminal constraint on quality variables.
NMPC for continuous polymerisation processes
Nonlinear Model Predictive Control based on mechanistic models shows superior performance compared to traditional linear MPC on processes with nonlinear characteristics, varying operating conditions and/or large variation in operating regions. This is particularly relevant for processes with frequent grade changes, where improved control may contribute to shorter transition times and reduced off-spec production.