A new Smith predictor and controller for control of processes with long dead time

Area of Science:

• Control Engineering
• Process Control Systems

Background:

• Smith predictor configurations are standard for controlling processes with significant dead time.
• Proportional-Integral (PI) or Proportional-Integral-Derivative (PID) controllers are typically employed in these systems.

Purpose of the Study:

• To investigate the performance benefits of using a PI-PD controller within a Smith predictor configuration.
• To demonstrate improved set point and disturbance responses compared to conventional PI/PID controllers.

Main Methods:

• Analysis of PI-PD controller parameter selection methods.
• Application of standard forms for controller design when the plant transfer function lacks zeros.
• Comparative analysis of PI-PD versus PID controller performance.

Main Results:

• The PI-PD controller yields superior set point and disturbance responses in specific scenarios.
• The standard forms provide a straightforward algebraic method for PI-PD parameter tuning.
• The study identifies inherent difficulties in using PID controllers for certain plant dynamics.

Conclusions:

• A PI-PD controller, when integrated into a Smith predictor, offers enhanced control performance for processes with long dead times.
• The proposed algebraic approach for parameter selection is effective, particularly for zero-less plant transfer functions.