Measurement noise attenuation in modified Smith predictor and automatic offset controllers for integrator plus...

Area of Science:

• Control Engineering
• Process Control
• Automation Systems

Background:

• The Åström-Smith predictor (ASP) is a modified controller for Integral Proportional Derivative Time (IPDT) models.
• ASP excels in ideal conditions but struggles with measurement noise, leading to transients and steady-state errors.

Purpose of the Study:

• To compare the performance of the modified Åström-Smith predictor (ASP) with an automatic offset controller (AOC).
• To evaluate the controllers' effectiveness in handling measurement noise and time-delayed processes.

Main Methods:

• The automatic offset controller (AOC) combines a higher-order stabilizing controller (SC) with a full disturbance observer (DOB).
• The disturbance observer (DOB) compensates for dead-time and integral modes.
• Adjusting the order of derivatives in the SC and low-pass filters in both SC and DOB enhances AOC performance.

Main Results:

• AOC demonstrates significant improvements in speed, robustness, and measurement noise attenuation for time-delayed processes.
• Even with noise filtering, ASP can exhibit permanent control errors or instability at higher noise amplitudes.
• AOC proves simpler and more reliable than ASP, especially when dealing with measurement noise.

Conclusions:

• AOC is a simpler and more reliable universal controller for processes with dominant first-order dynamics compared to ASP.
• The effectiveness of ASP as a universal controller depends heavily on the amplitude of measurement noise.
• AOC's performance requires careful selection of derivatives and tuning of low-pass filters, but offers superior robustness.