Image classification of the furnace bed height in recovery boilers
TAPPI Journal · 31.07.2026 · 25 (7) · 440
DOI: 10.32964/tj25.7.440 ↗Kurz gefasst
Fachartikel in der Zeitschrift TAPPI Journal, online erschienen am 31.07.2026. Verfasst von Tortsen Meyer und 1 weiteren Autor:innen. Die Arbeit nennt 13 Literaturangaben.
Abstract (Originalsprache)
The goal of this study is to develop an approach for classifying char bed height in a recovery boiler during operation. The resulting classified images form a dataset that can be used by the mill to support operational improvements. The study integrates data from the digester, evaporation plant, recovery boiler, and recausticizing plant, along with the classified image data. A convolutional neural network (CNN) was trained on bed camera images that were manually annotated to classify furnace bed height. The model was then used to generate a binary bedheight parameter over a 16-month period, and an extreme gradient boosting (XGBoost) model was applied to identify relationships with operating parameters across the recovery cycle. The outcomes of this research may offer deeper insights into the processes underlying bed dynamics. Ultimately, the developed system may serve as a tool for optimizing bed conditions and supporting stable combustion.
Zitieren
APA
Meyer, T., & DeMartini, N. (2026). Image classification of the furnace bed height in recovery boilers. TAPPI Journal, 25(7), 440. https://doi.org/10.32964/tj25.7.440
Vancouver
Meyer T, DeMartini N. Image classification of the furnace bed height in recovery boilers. TAPPI Journal. 2026;25(7):440. doi:10.32964/tj25.7.440
BibTeX
@article{Meyer2026,
author = {Meyer, Tortsen and DeMartini, Nikolai},
title = {Image classification of the furnace bed height in recovery boilers},
journal = {TAPPI Journal},
year = {2026},
volume = {25},
number = {7},
pages = {440},
doi = {10.32964/tj25.7.440}
}
EUVIMED