Multiresolution GPC-Structured Control of a Single-Loop Cold-Flow Chemical Looping Testbed.

Area of Science:

• Chemical Engineering
• Control Systems Engineering
• Energy Systems

Background:

• Chemical looping is a promising near-zero emission power generation technology from coal.
• The process exhibits complex nonlinear, multi-scale dynamics and significant model uncertainty, challenging traditional control methods.
• Existing robust control techniques are often inapplicable due to the inherent process complexity.

Purpose of the Study:

• To develop advanced multiresolution modeling and model-based control strategies for chemical looping processes.
• To address the limitations of traditional control methods in handling nonlinear, multi-scale dynamics and model uncertainty.
• To design self-tuning controllers capable of real-time adaptation and effective process management.

Main Methods:

• Utilized nonlinear autoregressive with exogenous input (NARX) model structure, nonlinear in wavelet basis but linear in parameters, for temporal dynamics identification.
• Employed gradient descent optimization for calculating control inputs and wavelet model parameters by minimizing a quadratic cost function.
• Implemented Lyapunov stability theorem to ascertain identification and tracking error convergence for self-tuning schemes.
• Augmented temporal control with a spatiotemporal multiresolution self-tuning deadbeat controller for enhanced performance.

Main Results:

• Developed a self-tuning controller design methodology integrating tunable multiresolution system representations into a generalized predictive control structure.
• Augmented the temporal control loop with a spatiotemporal multiresolution self-tuning deadbeat control loop.
• Demonstrated the effectiveness of the proposed methodology in generating fast, recursive real-time algorithms for controlling highly uncertain nonlinear multiscale processes.
• Validated the approach through data from implemented temporal and spatiotemporal solutions for a chemical looping cold flow tracking problem.

Conclusions:

• The proposed multiresolution modeling and self-tuning control approach effectively manages the complex dynamics of chemical looping processes.
• The methodology provides a robust solution for controlling highly uncertain, nonlinear, and multi-scale systems in real-time.
• This work advances the control strategies for near-zero emission power generation technologies like chemical looping.