AS_2

CONTROLLO

tecnica

Marzo 2017

Automazione e Strumentazione

On-Off control, cascade control, ratio control and override logic

are successively implemented in some model loops and checked

with students. Loops configuration, performance impact and

process added value are as well analysed. Feed-Forward con-

cept is explained and implemented in reboiler duty control in

order to anticipate column load changes, like training material

Figure 3

details.

The validity of the implemented control strategy to reject distur-

bances is checked in front of the changes in process variables cre-

ated by a transfer function block. Noise is also generated with the

help of a transfer function. Students also learn how to incorporate

simple equipment and instrumentation malfunctions like valve

stiction or heat losses into the simulation model.

How to incorporate to the model complex automated control

sequences that can be part of a control philosophy is by using

the Event Scheduler tool, like a sequence of automatic steps for

MVs-FFs.

As three-day course closure, attendees learn how to extract data

generated by the simulation model in order to be used by third-

party off-line analysis tools, as well as how to import historical

process data and how to use them as boundary condition or con-

troller set-point in a dynamic simulation. Students do also com-

pare the performance of two alternative control layouts by dupli-

cating the exiting simulation case and just modifying the control

settings in the copied model.

Simulation for Multivariable Control

Multivariable model predictive control (MPC) top-

ics are not covered in the initial three day courses.

If necessary, course content can be extended to

see how MPC controllers can provide a superior

control, either using the basic MPC implementa-

tion offered by the dynamic process simulator or

by using a commercially available MPC. For this

last case, the controller is configured by step testing

the column dynamic model and exporting the gen-

erated results to the MPC identification package.

Once configured, the MPC is used to control the

column, with the same user interfaces than the ones

in real control rooms.

Other uses of Simulation for Process Control

In the extended course coverage, additional use of

process Simulation for Process Control is shown in

exercises where

Relative Gain Array

(RGA) tech-

niques are used; where anti-surge control of centrif-

ugal compressors is configured; where a

Smith-Pre-

dictor

or (SISO MPC) controller is implemented for

large dead-time processes; where

OLE for Process

Control

(OPC) is used as communication protocol

to control the dynamic model with an external con-

trol algorithm developed in MatLab; and where key

control variables are optimized with a SQP algo-

rithm in the steady state simulation model.

References

[1]

Svrcek, William, Mahoney, Donald, Young,

Brent.

A Real-Time Approach to Process Control

John Wiley & Sons, Ltd., 2006. ISBN: 978-0-470-

02533-8.

[2]

Dissinger, Glenn. “Studying Simulation”.

Hydrocarbon Engineering

, May 2008.

[3]

McMahon, Terry. “Process Simulation and Pro-

cess Control”.

Chemical Engineering Progress

American Institute of Chemical Engineers, p. 19,

September 2013.

[4]

Courses and Events

. Inprocess. 30 November

2016.

http://www.inprocessgroup.com/en/news/.

Figure 3 - Example of training material showing how to take advantage of a Feed-Forward controller

performance

Figure 2 - Example of the Training Material describing the necessary steps to determine the SS Gains