Introduction to Feed-Forward Control

Overview

• Purpose: Enhance performance of control systems, especially for disturbance rejection.
• Method: Identify and measure the disturbance, then provide a feed-forward input to the controller.

Cascade Control

• Use: Control the source of disturbance locally before it affects the main process.
• Comparison: Preferable if the same final control element can control both disturbance and process.

Examples

1. Shower: Prevent water temperature spike when toilet flushes (cold water pressure drops).
2. Car Approaching a Hill: Adjust gas pedal preemptively to maintain speed going uphill.
3. Chemical System: Measure in a feed stream and adjust heat to reactor to maintain conditions.

Feedback vs. Feed-Forward

• Feedback: Reacts after disturbance (e.g., car slowing down on a hill prompts more gas).
• Feed-Forward: Anticipates disturbance for proactive adjustment.

Feed-Forward Control System Example

Components

• Hot Water: Comes into the heat exchanger (shell side).
• Cold Water: Enters and leaves heat exchanger (tube side).
• Goal: Cool down hot water without mixing it with cold water.

Variables

• Controlled Variable: Temperature of cold water exiting the system.
• Manipulated Variable: Flow rate of the hot water stream.

Disturbances

• Example: Changing cold water flow rate affects the system.

Design of Feed-Forward Controller

• Objective: Ensure output (Y) is unaffected by disturbance.
• Method:
  1. Write algebraic equations for block diagram.
  2. Set output to zero to find input signal (U).
  3. U = Disturbance * Feed-Forward Controller.*

Process and Disturbance Transfer Functions

Transfer Functions

• Feed-Forward Controller: Disturbance transfer function / Process transfer function.
• Implementation: Use proportional gain when process and disturbance delays are similar.

Dynamic vs. Static Feed-Forward Controllers

• Dynamic: Requires accurate modeling of delays.
• Static: Easier to implement, assumes delays and gains are approximately equal.
• Constraint: Not feasible if process dead time > disturbance dead time.

Modified Block Diagram with Feed-Forward Controller

• Setup: Measures disturbance, adjusts controller output to proactively counteract disturbance.
• Sensor Dynamics: Possible inclusion of transfer function for sensor dynamics.

Performance: Disturbance Rejection

• Example: Feed-Forward controller reduces effect of disturbances, maintaining setpoint.
• Practice Problem: Develop static and dynamic feed-forward controllers, recognizing feasibility constraints.

Summary

Cascade Control

• Components: Two sensors, two controllers, one valve.
• Attributes: No model required, small settling time for inner loop.

Feed-Forward Control

• Components: One controller, includes disturbance model.
• Attributes: Dead time constraints, model-based approach.

Recommendations

• Cascade Control: Use when two sensors and same control element can manage disturbance and process.
• Feed-Forward Control: Use when cascade is not applicable or inner loop is missing.

Additional Resources

• Course Website: furman.edu/PDC for more examples and detailed problems.