DOI: 10.12187/2025.03.011
Application of an improved OSELM algorithm in online detection of moisture content in the tobacco strip redrying process
ZHANG Lei
MA Yongshuai
HONG Binbin
XIONG Kaisheng
XU Dayong
DU Jinsong
LI Yinhua
ZOU Quan
Abstract：To address the challenges in directly detecting the moisture content of tobacco strips(a key quality indicator)and the significant delays in offline moisture measurements during the redrying process,this study proposes an adaptive modeling method using an improved Online Sequential Extreme Learning Machine(OSELM)for real-time online monitoring of moisture content at the drying zone exit.First,domain-specific expert knowledge and mutual information analysis were combined to select auxiliary variables most relevant to moisture content,thereby improving model generalization while maintaining predictive accuracy and reducing computational complexity.Subsequently,an OSELM-based modeling approach with adaptive forgetting factors(AFFs)was developed to address the strong nonlinearity and time-varying dynamics.The AFF strategy dynamically adjusted according to process variations,significantly enhancing the soft sensor's online tracking performance under complex operational conditions.Finally,validation using real-world production data from an industrial redrying facility showed that the proposed method outperforms traditional soft sensing approached in both detection accuracy and response time,thereby confirming its superior effectiveness.
Keywords：tobacco strip redrying machinemoisture content of tobacco stripmutual informationsoft sensoronline sequential extreme learning machineonline detection
Publication Date：2025-06-15
Online Publishing Date：2025-08-15(First online date of this platform, not the publication date of the document)
Pages：9( 95-103 )