Main content area

An adaptive neuro-fuzzy approach to bulk tobacco flue-curing control process

Wu, Juan, Yang, Simon X., Tian, Fengchun
Drying technology 2017 v.35 no.4 pp. 465-477
barns, color, data collection, databases, leaves, models, odors, process control, tobacco, tobacco industry
Bulk tobacco flue-curing process significantly affects the quality and fragrance of cured tobacco leaves. The control of bulk tobacco flue-curing process is therefore quite important for tobacco industry. In this work, a neuro-fuzzy-based method for controlling bulk tobacco flue-curing process was proposed. In particular, an adaptive network-based fuzzy inference system (ANFIS) was developed to predict the set point changing time. To illustrate the applicability and capability of the ANFIS model, the proposed approach was tested with a bulk tobacco flue-curing barn database, which included totally 574 data sets obtained in the four curing cycles. The results demonstrated that the proposed approach could be applied successfully and provide high accuracy and reliability for bulk curing barns. Furthermore, to analyze how input factors affect the bulk tobacco flue-curing control process, the selection of input linguistic factors was also discussed. The factors of color and curing phase were found to have the most substantial influence on curing control process. A comparative study among the proposed neuro-fuzzy approach and other related methods was also performed. Both the statistical measures and visual assessment illustrated that the proposed ANFIS method outperformed the other methods in this study, which further showed the effectiveness and reliability of the neuro-fuzzy approach to bulk tobacco flue-curing control process.