Model-Based Control Algorithms to Handle Complex Control Problems

The Modular Multivariable Controller (MMC) provides an effective means of controlling one or more process variables to their respective setpoints using two or more control outputs. If there are more control outputs than process variables, the MMC can drive selected control outputs to user-specified target values.

We developed the Modular Multivariable Control technology to take into account the interactions among multiple process inputs and outputs, with the controller design and tuning remaining as simple as that for PI/PID systems.

This multivariable control technology has the potential to operate processes at their most cost-effective conditions. The software is modular and, therefore, flexible enough to be adapted to different control objectives and system configurations.

MMC allows decoupling between process variables.

Changing the setpoint of one variable will not affect other process variables unless there are modeling errors, disturbances, or noise.

Top

MMC is easy to design and tune.

MMC is a model-based controller. Each PV-CO relationship is modeled as a first order lag plus deadtime. One single tuning parameter (fast or medium or slow speed) is required for each PV-CO loop. Process variables are caused to track desired trajectories.

MMC is structurally robust.

Any control loop can be put on manual without loss of performance of the rest of the control system. This feature is difficult to achieve using centralized multivariable control systems.

MMC is flexible.

Coordinated Controller and Modular Multivariable Controller blocks can be assembled in various configurations to meet specific control objectives.

The control outputs are coordinated.

The order in which the control outputs are used to compensate for a process variable deviation from its setpoint can be specified.

MMC control of a heavy oil fractionator.

Top draw rate (CO1) is used to control top end concentration (PV1).

Side draw rate (CO2) controls side end concentration (PV2).

Bottoms reflux duty (CO3) has its target value.

CO1 hits the low constraint (Time~100) due to the change in setpoint of PV1 (SP1).

CO3 is used to compensate for PV1 deviation from the setpoint during the transient period.

After the transient period and change in setpoint of PV2 (SP2), the value of CO1 increases and CO3 can be driven to its target value.

Heavy Oil Fractionator

MMC maintains a selected control output at its target value.

Two control outputs maintain two process variables at their respective setpoints while the third control output is driven to its user-specified target value. Control output target values can be changed without disturbing process variables. The control output target value can result from a process optimization criterion or an economic criterion.

MMC accounts for the relative importance of process variables.

One process variable is considered more important than the other, and is maintained at its setpoint with highest priority. With all three control outputs available, there are three goals you can accomplish in the following order of importance:

1. Maintain one process variable at its setpoint with higher priority
2. Maintain second process variable at its setpoint with lower priority
3. Maintain a selected control output at its target value

MMC handles control output and process variable constraints.

In order to improve performance of the MMC, you can enter constraints on all control outputs and process variables, and can also feed back to the controller the control outputs that were actually applied to the process.

MMC software is easy to maintain.

The MMC code is relatively short. The algorithm uses our Coordinated Controller (CC) 1xN code according to the controller configuration. The MMC procedure is easy to understand and maintain.

The Modular Multivariable Controller (MMC)

• Can be used to control one or more process variables achieving multiple control objectives.
• Is easy to design, install, tune, and maintain.
• Allows for total decoupling between process variables.
• Is structurally robust.
• Handles constraints on control outputs.
• Handles constraints on process variables.
• Can drive selected control outputs to their target values.
• Ranks the control objectives according to importance.
• Software is modular and, therefore, flexible so that it can be easily adapted to different control objectives.

Economic Benefits of the MMC

• Performance of the Modular Multivariable Controller equals or exceeds that of far more expensive multivariable controllers.
• No additional hardware is needed to install the controller.
• The MMC technology has the potential to operate processes at the most cost-effective conditions saving energy and materials and maintaining uniform quality of products.

We provide the Modular Multivariable Controller with comprehensive technical description and user manual. We describe step-by-step procedures for design and tuning. The controllers can be downloaded to your existing platform. Consulting is available.

Our model predictive controller toolkit offers solutions to basic multivariable 1x2, 1x3, 2x2, 2x3 and other controller design configurations according to your needs. The controllers can be used for split range control problems, boiler control, distillation column control, constraint control, and many other control applications. The controllers are available in FORTRAN, C-language, Bailey User Defined Functions (UDF), Honeywell Control Language and other platforms.

The Modular Multivariable Control (MMC) and the Coordinated Control (CC) are patented technologies by ControlSoft, Inc., Cleveland, Ohio.

The Modular Multivariable Control technology was developed by process control experts Doctors Coleman Brosilow, Irving Lefkowitz and Tien-Li Chia.

Top

For more information about MANTRA, use our contact form.