New Technology Enhances Model Predictive Control
- Type: Conference Presentation
- Conference Type:
AIChE Spring Meeting and Global Congress on Process Safety
- Presentation Date:
March 21, 2010
- Skill Level:
The standard way of testing a process causes the configuration and tuning of the PID control system for the process to be embedded in the model developed for the Model Predictive Controller (MPC). New technology has become available that can be used to interchange the independent variables with the dependent variables in a controller model. This interchange of variables permits any MPC model to be corrected mathematically, when the PID configuration and/or the tuning of the PID controllers are changed. The interchange of variables is a matrix pivot operation that deals with vectors rather than whole numbers. Many of the apparent model problems with MPC controllers are caused by retuning and configuration changes in the PID controllers that are in the scope of the MPC. Until the development of the new technology, significant changes in the PID configuration or tuning required retesting the process. The pivot operation effectively eliminates the need to retest the process for MPC controllers for changes in the PID control system. The maintainability of the MPC controller is improved with new technolgy, which permits the reconfiguration of the individual PID loops in the controller. For example, a control valve in the MPC may become inoperative during a run and cannot be fixed until the process is shutdown for maintenance. The pivot operation can switch the valve to manual in the MPC model, which will maintain the integrtiy of the controller's model. Many times the process is moved to a new region of operation, which causes one or more of the PID controllers to be unstable. Addition of software added to the pivot operation permits the PID controllers in the MPC model to be retuned and a new MPC model created with new PID tuning.