Multivariable Control / Optimization - IDCOM-HIECON

IDCOM-HIECON is a predictive multivariable controller / optimiser for industrial continuous processes.
 
 

ORIGIN
 
A first simpler version, IDCOM, had been developed in the early seventies by ADERSA and has been transferred to several industries and engineering companies.
In 2004 the technology, the software products and all CAD tools of IDCOM-HIECON were globally transferred to Sherpa Engineering company which is responsible today for the applications, the maintenance and further development.
The present version of IDCOM-HIECON is the fruit of experience and applications performed worldwide by the Process Control team at Sherpa.

 
MODEL
  
IDCOM-HIECON uses a model which is a "black box" mathematical representation of the process to be controlled.

The type of representation is here the step (or impulse) response which is well adapted to multivariable continuous processes.
 
 
 
 FONCTIONALITIES
First, IDCOM-HIECON has the capabilities of (almost) all multivariable controllers :
  • each process output moves towards its set point, following a dynamic path which is specified in terms of closed loop time response. 
  • the manipulated variables (actions) satisfy the defined min max and speed constraints 
  • the measured disturbances are taken into account as feed forward variables
  • the multivariable problem is globally solved
 
Added to these basic capabilities, IDCOM-HIECON gets its advantages from powerful specific capabilities :
  • the number of manipulated variables may be different from the number of process outputs to be controlled
  • secondary objectives, with a lower priority than the set points, can be specified to some of the process variables
  • the manipulated variables can also satisfy some secondary objectives as long as the working conditions make it possible
  • constraints can be defined on process outputs
  • all these specifications are defined in a table ("control structure") which is understandable by the user and by the control algorithm
  • it is possible to define as many "control structures" as necessary for facing different working conditions
  • the controller can work in degraded situations, with less variables (actions or process variables) than defined basically
  
 

HIERARCHICAL CONTROL

Depending on the working conditions, the objectives may be not fully reachable because of constraints.

In such situations, IDCOM-HIECON satisfies the objectives in a hierarchical order :

A hierarchy is also defined inside the category of the main objectives in such a way some of them can be satisfied a better way than others in case of conflict.

According to the hierarchical control principles, the primary objectives (set point control) are satisfied before the optimization objectives if any.

SET POINT AND ZONE CONTROL

A closed loop time response can be specified to each of the process outputs which are given a set point. Added to that, it is possible to define a range arround the set point and inside which the controller is to behave smoothly. This procedure, useful in case of noisy measurements, produces smooth actions when the measured process out put is close to the set point value without accepting any constant bias. The control strenght is given different values inside and outside the zone.


 
OPTIMIZATION FUNCTIONS

 
The local dynamic optimization, as it is performed in IDCOM-HIECON, makes adjustments of the degrees of freedom (manipulated variables) in order to satisfy a given objective (such as : maximze such flow rate) while repsecting the constraints and the primary objectives. 

The IDCOM-HIECON optimization objectives are :
  • the Ideal Resting value (IRV)
  • Maximization (or minimization)


IDEAL RESTING VALUE (IRV)

The IRV is a secondary objective whih is taken into account by IDCOM-HIECON in case there are more degrees of freedom than primary objectives (more available actions than set points).

If an IRV is defined on a manipulated variable, that variable will move towards the IRV value as long as this MV is not required for satisfying the primary objectives.

In case an IRV is defined on a process variable, it is similar to a set point objective except that it has a lower priority level.

MAXIMIZATION - MINIMIZATION

 Every variable, process output and manipulated variable (or a linear combination of variables) can be maximized or minimized.