INCA MPC

How the MPC operates is explained in detail in the course. The implementation of the control strategy and the tuning of the controller will be discussed in the course.

At the end of the course the participant should have insight in how the controller trades-off the different requirements posed in the control strategy and how the process behaviour influences the control actions.

The performance of a model predictive controller is highly dependent on the quality of the model. The more accurate the model, the better the controller will perform. Modeling is therefore a crucial phase in the whole control project and hence also in the course. In multi-variable process control identification is used to obtain the process models. The identification procedure and the identification tool Modeler are discussed extensively. The quality of the model is highly dependent on the quality of the data, obtained from dedicated on-site tests. Testing in general, as well as more sophisticated testing based on PRBNS (Pseudo Random Binary Noise Signal) or PRMNS (Pseudo Random Multi level Noise Signal) are a topic of discussion.

Hands-on exercises form an important part of the course. For identification and model predictive control each approximately ¾ of a day is spent on hands-on training. A rigorous model will be used to simulate the process. In the exercises the participant will be guided through the modeling and the control design phase for this "process". At the end of the course the controllers designed will be commissioned on this "process".

Schedule:

Day 1

• Introduction:
  • The INCA MPC product suite
  • The overall project approach
• Identification of process dynamics:
  • The identification project approach
  • Identification of FIR models
  • Step and PRBNS-test signals
• Hands-on exercise Modeler

Day 2

• Recapitulation of the first day
• Identification of process dynamics:
  • Signal preprocessing of data
  • Design of random test signal
  • Identification of parametric models
• Hands-on exercise Modeler

Day 3

• Recapitulation of the second day
• Model Predictive Control:
  • Definition of the control problem
  • The basic components of MPC: Prediction, steady state optimisation and dynamic optimisation
  • Least squares control
• Hands-on exercise INCA MPC Engine

Day 4

• Recapitulation of the third day
• Model Predictive Control:
  • Absolute priority control
  • Setpoint, constraints and zone specification
  • Tuning of INCA MPC Engine
  • How process dynamics limit the controlled process behaviour (Delays, inverse response, directionality and their influence on the tuning)
• Hands-on exercise INCA MPC Engine
• Discussion and closure of the course

Training dates and location:

• Date: to be announced soon
  
• Location: Boxtel, The Netherlands

Contact: Christiaan Moons, tel. +32 16 39 30 83, e-mail christiaan.moons@ipcos.com 

Training Registration:

Download registration document