Configuration of the model based predictive controller MPC10x10 for Tennessee Eastman Benchmark Process in SIMATIC PCS 7
Objective of the Application
In addition to the model based predictive controller ModPreCon for up to 4x4 interacting manipulated and controlled variables the new large predictive controller MPC10x10 for up to 10x10 manipulated and controlled variables is introduced in the context of PCS 7 V8.1. The MPC10x10 not only offers larger variable numbers but also some principally new functions which are introduced in this application note.
This document expands the applications notes "Model based predictive multi-variable control at the example of a distillation column" (Entry-ID 37361208) and " Fluidized bed dryer - design of a predictive controller with operating point optimization" (Entry-ID 61926069) in which the functions of the ModPreCon are already described and is focused on the new features of the MPC10x10.
Using the example of a processing plant with several units (Tennessee Eastman benchmark process with reactor, separator, stripper and compressor) the design of the MPC10x10 as a supervisory master controller is explained, dealing with six manipulated variables, five controlled variables and three measurable disturbance variables.
The following issues are treated in this application:
- Identification of process models with numerous manipulated and controlled variables
- Dynamic online optimization considering constraints
- Setting of target values for manipulated variables
- Simulation studies inside MPC configurator
- Commissioning and test of the controller
SIMATIC PCS 7 V8.1
|Content of downloads||
Configuration of the MPC10x10 for Tennessee Eastman Benchmark Process