Enhancing Efficiency in Coal-Fired Boilers Using a New Predictive Control Method for Key Parameters.

Area of Science:

• Boiler engineering
• Combustion control
• Renewable energy integration

Background:

• Renewable energy integration causes frequent load changes in coal-fired boilers.
• Load fluctuations lead to deviations in key operational parameters, reducing efficiency and stability.
• Existing control methods struggle with complex variable-load conditions.

Purpose of the Study:

• To develop a novel predictive control method for precise parameter control in coal-fired boilers under variable loads.
• To enhance combustion stability and operational efficiency in the context of carbon neutrality goals.
• To reduce energy consumption and CO2 emissions.

Main Methods:

• A coupled Transformer-gated recurrent unit (GRU) predictive model was developed for high-accuracy prediction of oxygen content, bed pressure, and main steam temperature.
• A Whale Optimization Algorithm (WOA)-based real-time optimization strategy was employed for coordinated control adjustments.
• The method was validated on a 300 t/h boiler over a 24-hour load cycle.

Main Results:

• The coupled Transformer-GRU model achieved high prediction accuracy (e.g., MAE of 0.095% for oxygen content).
• The predictive control method significantly reduced fluctuations in oxygen content (62.07%), bed pressure (50.95%), and main steam temperature (40.43%).
• Field validation demonstrated substantial improvements over traditional control methods.

Conclusions:

• The novel predictive control method effectively enhances boiler thermal efficiency and steam quality under variable load conditions.
• The method offers an estimated annual efficiency gain of ~1.77% and significant CO2 emission reductions.
• This approach supports the stable integration of renewable energy sources while maintaining boiler performance.