# EcoStruxure™ Control Expert - Predictive Control Library (Model Based Control - APC)

# EcoStruxure™ Control Expert - Predictive Control Library (Model Based Control - APC)

### Overview

#### EcoStruxure™ Control Expert - Predictive Control Library (Model Based Control)

Predictive Control Library is Model based Advance Process Control Library to minimize process dynamics and Delayed response

Model Predictive controllers contain a model of the process which allow the system to predict the impact of the output on the process.

The delay in reaction is modelled within the controller The actual output is tracked against the predicted output to monitor the impact of the change

A model based predictive controller is a controller that uses a model in real time for the computation of the control action to be applied.

### Benefits

#### Model

The model which is embedded in the controller is a mathematical equation that computes a 'model' output which is comparable to the process output PV (Process Variable).

The model represents the relationship linking the process input(s) to the process output.

Identify this model and estimate the parameters from the recorded plant tests. The model is used to predict the process output and to compute the control action in order to satisfy a given target specified on the PV.

#### Future Desired Trajectory

At present time (n), the process output is PV(n) and the set point value is SP(n).

The future desired trajectory (so-called reference trajectory) is the desired behavior of the process output to move from its present value PV(n) to SP(n) in the future

The reference trajectory is computed by a first order system (see above) and the response time of this trajectory is the closed loop response time: the PV will respond to a set point step change with the response time given by the user.

#### Solver

The solver is the part of the controller which computes the control action to be applied in such a way the predicted output at time (n+H) is equal to the reference trajectory at the same future instant.

The computed control action takes into account the constraints which limit the input moves (high and low limits and rate of change).

#### Self Compensation

Some non measured variables may disturb the process.

With unmeasured ramp-type disturbance, a bias between PV (Process Variable) and SP (Set Point) may appear.

The aim of the self-compensator is to reject this kind of disturbance, and to avoid such a bias.