Optimization of sootblower operation based on machine learning to improve efficiency and NOx reduction

Joko Santoso, Agus Setyawan,  Muchammad,
Optimization of sootblower operation based on machine learning to improve efficiency and NOx reduction,
Applied Thermal Engineering,
Volume 281, Part 1,
2025,
128598,
ISSN 1359-4311,
https://doi.org/10.1016/j.applthermaleng.2025.128598.
(https://www.sciencedirect.com/science/article/pii/S1359431125031904)
Abstract: The issue of slagging deposits on pipes inside the boiler can lead to negative impacts. The efficiency of heat transfer between the combustion gas outside the pipes and the water inside becomes ineffective due to the high thermal resistance of slagging deposits. This results in problems such as reduced boiler efficiency and increased fuel consumption. The sootblower functions to remove soot, ash, and other deposits adhering to the outer surface of the pipes. However, improper operation patterns of the sootblower can cause damage to boiler pipes. If the sootblower is operated too infrequently, slagging and fouling buildup on the pipes will increase. Conversely, operating the sootblower too frequently can lead to excessive use of steam and potential erosion of the pipe surfaces. By optimizing sootblower operation using machine learning, a more targeted operating pattern can be achieved in accordance with the cleanliness factor target. This optimization can reduce the average sootblower operating frequency by two times across all boiler areas and decrease steam consumption by 54 tons per day or 1,681.53 tons per month. It also helps achieve the heat rate target according to the 2022 baseline and reduce NOx emission. With fewer sootblower operations, pipe erosion can be minimized, and lower steam consumption makes the sootblower operation pattern more effective in reducing slagging and fouling buildup and NOx emission are lowered.
Keywords: Slagging; Sootblower; Machine Learning; Supercritical Boiler; Heat Rate