Centralized PI controller design method for MIMO processes based on frequency response approximation.

Area of Science:

• Control Systems Engineering
• Process Control

Background:

• Designing controllers for Multi-Input Multi-Output (MIMO) systems often involves complex calculations, particularly matrix inversions.
• Existing methods may struggle with processes dominated by time delays or high dimensionality.

Purpose of the Study:

• To propose a simplified frequency domain controller design method for MIMO processes.
• To reduce computational complexity in controller design by simplifying matrix inversion.
• To extend the method for non-square MIMO systems.

Main Methods:

• Evaluating the process transfer function matrix at a low frequency point to simplify inverse calculations.
• Deriving desired closed-loop transfer functions using a single tuning parameter per diagonal element.
• Employing a model matching technique for centralized Proportional-Integral (PI) controller design.
• Extending the method for non-square MIMO systems using matrix squaring or pseudo-inverse evaluation.

Main Results:

• The proposed method significantly simplifies the calculation of the process transfer function matrix inverse.
• Centralized PI controllers designed using this method demonstrate acceptable performance for lag-dominated and time-delay dominated processes.
• The technique is applicable to high-dimensional and non-square MIMO processes.

Conclusions:

• A computationally efficient and effective controller design method for MIMO processes has been presented.
• The method offers a practical approach for designing PI controllers across a range of process dynamics and dimensions.
• The extension to non-square systems enhances the method's versatility.