Model Predictive Control Introduction
The MPC Component for Proficy CSense provides Proficy CSense blueprint developers with a powerful model predictive control component for the development of advanced process control solutions.
About Model Predictive Control (MPC)
MPC is a method of process control that uses dynamic models of the process, usually linear empirical models. These models are used to predict changes in dependent variables caused by changes in independent variables, in complex dynamic systems.
-
Dependent variables: Measurements that represent either control objectives or process constraints.
-
Independent variables: Generally, set point values, such as pressure or temperature, or control elements, such as valves or dampers.
Using process models, process values and plant measurements, MPC calculates changes required to the independent variables to keep the dependent variables operating close to the defined target values. When a further change is required, MPC repeats the calculation.
System Architecture
The MPC Component for Proficy CSense consists of two components:
MPC Editor
The MPC Editor is a standalone application that is used to design and configure controllers. The application can import and export MPC models from comma-separated value (.csv) files. These model files are then imported into the Model Predictive Control block in the Architect.
Model Predictive Control block
The Model Predictive Control block executes a fully configured MPC model within the context of a CSense Architect blueprint. The block allows for input mapping and model selection. If you want to design a new model using the MPC Editor, refer to Controller Configuration Using the MPC Editor.
Related topics: