SUPCON Achieves Breakthrough in Industrial AI Application for Circulating Fluidized Bed Boilers in Thermal Power Plants

Recently, SUPCON and Jiande Thermal Power Plant of Wynca Group successfully passed the acceptance of their boiler combustion optimization project based on industrial AI. In this project, SUPCON leveraged various industrial AI technologies, including the Transformer architecture, to develop a combustion optimization control solution. The implementation marks the successful application of industrial AI in cogeneration, ensuring the safe and intelligent operation of three circulating fluidized bed boilers at Jiande Thermal Power Plant. The project has also significantly enhanced the overall efficiency of the plant, setting a benchmark for digital and intelligent transformation in the industry.

01 Pain Points?in Boiler Operation within the Cogeneration Industry

• Traditional control methods based on conventional algorithms face limitations when dealing with scenarios with strong coupling, high inertia, big latency, and multi-variable control in boiler combustion, leading to delayed load response.
• The complex and variable nature of combustion often results in incomplete combustion or high flue gas losses, causing low thermal efficiency in traditional control modes.
• Under the traditional main steam header control mode, the boiler load is manually allocated, preventing multi-mode distribution based on boiler efficiency, load conditions, and combustion status, thereby affecting overall boiler combustion efficiency.
• Conventional boiler NOx regulation suffers from significant inertia and delay. Increased ammonia injection is often taken to prevent NOx exceedances, resulting in high ammonia slip issues.

These factors directly or indirectly lead to low automatic control rates, large fluctuations in operating parameters, delayed adjustments, incomplete combustion, high flue gas losses, and low thermal efficiency.

02 Construct AI Combustion Optimization Control System based on?Transformer Architecture

SUPCON's?boiler combustion optimization control solution aims to achieve big?data?optimization based on?industrial AI, utilizing?the Transformer architecture?and integrating time series forecasting models and stacked recommendation algorithms.?By combining feedforward and feedback control methods, key parameters can be?optimized and controlled, leading to more stable boiler combustion. Multi-objective collaborative optimization ensures that the combustion process avoids overheating and slagging, maintaining a stable combustion state and optimal bed temperature, thereby enhancing operational efficiency and improving boiler combustion efficiency.

03 Enhance?Overall Efficiency, Stability, and Profitability at Jiande Thermal Power Plant

SUPCON's industrial-AI-based intelligent boiler combustion optimization control system helps Jiande Thermal Power Plant achieve an automatic control rate of over 95%, reduce the average fluctuation amplitude of key operating parameters by more than 30%, and lower coal consumption per ton of steam by over 1% compared to manual operation. By using key combustion parameters as feedforward features, ammonia slip was reduced by over 20% compared to manual control while keeping NOx levels within standards. This effectively reduces the workload of operators, enhances boiler combustion efficiency, and significantly contributes to the goals of improving personnel efficiency, stabilizing operations, and increasing profitability at Jiande Thermal Power Plant.

Industrial intelligence is a crucial driver for the future development of the industry. SUPCON is committed to continually providing intelligent solutions with "AI+Safety" "AI+Quality" "AI+Green" "AI+Profit" and leading the revolutionary transformation of the industrial.