MPC Grade Change Control With Forced Input Ramping in Paper Machines
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The challenge in papermaking is to perform grade changes rapidly and minimize off-specification production while optimally coordinating process input movements, which is difficult due to complexities in control configuration and operation.
Innovation Solution
A model predictive control (MPC) system is developed with a controller reference trajectory design technique that automatically determines process output reference trajectory delays, provides the entire planning process output reference trajectory at the start of the grade change, and uses process input forced ramping for optimal coordination of process inputs to maintain desired trajectories without significant deviations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If process input variables are ramped in open-loop to achieve rapid grade changes, then productivity is improved, but manufacturing precision deteriorates due to deviations from specification
Solution Approach 1:
The system pre-calculates and provides the entire planning process output reference trajectory at the start of the grade change, rather than incrementally. This allows the MPC controller to have full knowledge of the desired path ahead of time, enabling it to coordinate input movements optimally to achieve both rapid transitions and precise tracking without significant deviations from specification.
Solution Approach 2:
The system implements forced ramping of process inputs with dynamically adjusted trajectories over a prediction horizon. The MPC controller continuously optimizes the ramping profiles based on current process state and predicted future behavior, allowing the system to adapt the speed and coordination of input changes to maintain precision while achieving rapid grade changes.
2Manufacturing precision
If multiple process inputs are coordinated to move along optimal trajectories, then manufacturing precision is improved, but device complexity increases due to control configuration requirements
Solution Approach 1:
The MPC controller serves multiple functions simultaneously: it predicts future process behavior, optimizes coordinated input trajectories, enforces constraints, and handles forced ramping requirements. This multi-functionality is achieved through a single unified control algorithm rather than multiple separate control systems, reducing overall configuration complexity while maintaining precise trajectory tracking for multiple process inputs.
Solution Approach 2:
The system changes the parameter representation of control trajectories by working with reference trajectories defined over a prediction horizon rather than simple setpoint sequences. This parameter transformation enables the controller to optimize entire trajectories at once and coordinate multiple inputs more effectively, improving precision without proportionally increasing complexity.
3Productivity
If the entire planning process output reference trajectory is provided at the start of grade change, then productivity is improved through faster transitions, but ease of operation worsens due to increased setup complexity
Solution Approach 1:
The MPC controller performs self-service by automatically generating and optimizing the coordinated input trajectories based on the provided reference output trajectory. The controller independently determines the optimal timing and coordination of each input variable without requiring manual configuration of individual trajectory parameters, reducing setup complexity while enabling fast grade changes.
Solution Approach 2:
The system uses feedback from the prediction horizon to continuously refine the coordinated trajectories. The MPC controller monitors actual process behavior against predicted behavior and adjusts the optimization accordingly, allowing the system to achieve fast transitions with robust performance without requiring extremely precise manual setup of trajectory parameters.
Data Source
Figure 1~2B
Figure 2A
Figure 3
AI summary
A controller reference trajectory design technique to enable high-performing automatic grade change performed by a model predictive control (MPC). Techniques include: (1) automatic determination of appropriate process output reference trajectory delays to enable optimum coordination of process input movements; (2) providing the entire planning process output reference trajectory ramp at the start of the grade change instead of just incrementally as the grade change progresses, again enabling movement of the process inputs to drive process outputs along the planned future path instead of just towards the current target; and (3) use of the process input forced ramping to allow linear ramping of process inputs with optimal coordination of other process input movements to keep all process outputs following the desired trajectories. The technical benefits are faster and higher performing grade changes. In addition, the use of this technology allows easier setup and maintenance of the automatic grade change package.