Economic zone data-enabled predictive control for connected open water systems.

Area of Science:

• Water Resource Management
• Control Engineering
• Artificial Intelligence

Background:

• Real-time operation of open water systems is crucial for safety, efficiency, and energy optimization.
• Existing control methods (rule-based, model-based) have limitations in complex, dynamic, and uncertain environments.
• A data-driven approach is needed to overcome the challenges of traditional control strategies.

Purpose of the Study:

• To develop a fully data-driven, zone-based control framework for connected open water systems.
• To achieve safe and energy-efficient operation through adaptive control target zone selection.
• To minimize operational energy consumption while maintaining desired water levels.

Main Methods:

• Proposed a mixed-integer economic zone data-enabled predictive control (DeePC) approach.
• Utilized lexicographic optimization to handle multiple control objectives (zone tracking, energy minimization).
• Employed Bayesian optimization for adaptive control target zone selection to balance competing objectives.

Main Results:

• The proposed method maintained water levels within the desired zone for 97.04% of the operating time.
• Achieved an average energy consumption of 33.5 kWh per 0.5 h.
• Significantly reduced zone violations and energy consumption compared to baseline and rule-based methods (e.g., 74.96% fewer violations vs. rule-based control).

Conclusions:

• The data-driven DeePC framework with adaptive zone selection effectively manages open water systems.
• The approach offers a robust solution for safe, energy-efficient operation, outperforming traditional methods.
• This method eliminates the need for explicit dynamic modeling, simplifying implementation for complex water systems.